Upload folder using huggingface_hub

.gitattributes

@@ -33,3 +33,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+images/logo1.png filter=lfs diff=lfs merge=lfs -text
+images/logo2.png filter=lfs diff=lfs merge=lfs -text

.gitignore

@@ -175,4 +175,5 @@ cython_debug/
 
 # outputs
 output/*
-investigators/output/*
+investigators/output/*
+search_results_2025*

app.py

@@ -15,16 +15,59 @@ def investigate(target_name, affiliations):
     
     return crew_output.raw
 
-view = gr.Interface(
-    fn=investigate,
-    inputs=[
-        gr.Textbox(label="Target name:"),
-        gr.Textbox(label="Target Affiliations (comma separated):")],
-    outputs=[gr.Markdown(label="Risk Assessment Report:")],
-    flagging_mode="never",
-    examples=[
-        ["Raz Nissim", "Ben Gurion University, General Motors"],
-    ],
-    title="OSINT Investigator",
-    description="Enter the name of your target and their affiliations (to make search easier), and get a AML Risk assessment based on their public information.",)
-view.launch(inbrowser=True)
+# view = gr.Interface(
+#     fn=investigate,
+#     inputs=[
+#         gr.Textbox(label="Target name:"),
+#         gr.Textbox(label="Target Affiliations (comma separated):")],
+#     outputs=[gr.Markdown(label="Risk Assessment Report:")],
+#     flagging_mode="never",
+#     examples=[
+#         ["Raz Nissim", "Ben Gurion University, General Motors"],
+#     ],
+#     title="OSINT Investigator",
+#     description="Enter the name of your target and their affiliations (to make search easier), and get a AML Risk assessment based on their public information.",
+#     layout="vertical")
+# 
+# view.launch(inbrowser=True)
+
+    # Clear button functionality
+def clear_inputs():
+    return "", ""
+    
+with gr.Blocks() as view:
+    gr.Markdown("# OSINT Investigator")
+    gr.Markdown("#### Enter the name of your target and their affiliations (to make search easier), and get a AML Risk assessment based on their public information.")
+    with gr.Row(equal_height=True):
+        with gr.Column(scale=3):
+            name_input = gr.Textbox(label="Target name:")
+            affiliation_input = gr.Textbox(label="Target Affiliations (comma separated):")
+            with gr.Row():
+                clear_btn = gr.Button("Clear")
+                submit_btn = gr.Button("Investigate")
+        with gr.Column(scale=1):
+            img1 = gr.Image("images/logo1.png", show_download_button=False, show_fullscreen_button=False, show_label=False, show_share_button=False)
+        
+    
+    with gr.Row():
+        output = gr.Markdown(label="Risk Assessment Report:", container=True, show_copy_button=True)
+    
+    submit_btn.click(
+        fn=investigate, 
+        inputs=[name_input, affiliation_input],
+        outputs=output
+    )
+    
+
+    
+    clear_btn.click(
+        fn=clear_inputs,
+        inputs=[],
+        outputs=[name_input, affiliation_input]
+    )
+    gr.Examples(
+        examples=[["Raz Nissim", "Ben Gurion University, General Motors"]],
+        inputs=[name_input, affiliation_input]
+    )
+
+view.launch(inbrowser=True)

investigators/src/investigators/config/agents.yaml

@@ -3,10 +3,12 @@ researcher:
     Research Specialist
   goal: >
     Gather comprehensive information from public sources about the target: {target}, that has the following affiliations: {affiliations}. 
+    If no information is found, just say "No information found.".
   backstory: >
     You are an expert OSINT researcher with deep experience in 
     finding and connecting information about people and businesses. You know how to
     follow information trails and identify valuable data sources.
+    You have the ability to search the web using search tools.
   llm: openai/gpt-4o
 
 fincrime_analyst:

investigators/src/investigators/config/tasks.yaml

@@ -10,11 +10,13 @@ research_target:
     3. Public records and news mentions
     4. Legal or regulatory issues
     
-    Run multiple searches until your findings are adequate. 
+    Run multiple searches until your findings are adequate.
     Organize your findings clearly and be sure to track your sources.
+    If no information is found, just say "No information found.".
   expected_output: >
     Comprehensive research findings on {target}
   agent: researcher
+  output_file: output/research_{target}
 
 analyze_target:
   description: >
@@ -53,12 +55,12 @@ reporting_task:
     Your report should include:
     
     1. Executive Summary
-       - Brief overview of the target
+       - Overall risk assessment score [between 0.0 (no risk) to 1.0 (high risk)]
        - Summary of key findings and risk assessment
     
     2. Target Profile
        - Background information
-       - Known associates and relationships
+       - Known associates and relationships (if applicable)
        - Business activities and corporate structure (if applicable)
     
     3. Red Flag Analysis
@@ -81,4 +83,5 @@ reporting_task:
   agent: osint_reporter
   context:
     - analyze_target
+    - research_target
   output_file: output/osint_report_{target}.md

investigators/src/investigators/crew.py

@@ -1,8 +1,31 @@
+from typing import Tuple, Any
 from crewai import Agent, Crew, Process, Task
 from crewai.project import CrewBase, agent, crew, task
+from crewai import TaskOutput
 from crewai_tools import SerperDevTool
+from crewai_tools import ScrapeWebsiteTool
+from crewai_tools import BraveSearchTool
+NO_INFORMATION_FOUND="no information found"
 
+def validate_researcher_content(result: TaskOutput) -> Tuple[bool, Any]:
+    """Validate research content meets requirements."""
+    try:
+        # print("In validate_researcher_content...")
+        # Check word count
+        word_count = len(result.raw.split())
+        # print("Word Count:", word_count)
+        # print(result.raw)
 
+        if word_count < 100:
+            return (False, "Not enough content")
+
+        if NO_INFORMATION_FOUND in result.raw.lower():
+            return (False, "Model says that no information was found")
+
+        return (True, result)
+    except Exception as e:
+        print("The exception was:", e)
+        return (False, "Unexpected error during validation")
 
 @CrewBase
 class Investigators():
@@ -16,9 +39,9 @@ class Investigators():
         return Agent(
             config=self.agents_config['researcher'],
             verbose=True,
-            tools=[SerperDevTool()],
+            tools=[SerperDevTool(save_file=True)],
             retry_on_fail=True,  # Enable retry
-            max_retries=3  # Set maximum retries
+            max_retries=3,  # Set maximum retries
         )
 
     @agent
@@ -38,10 +61,15 @@ class Investigators():
     # Tasks
     tasks_config = 'config/tasks.yaml'
 
+    
+
     @task
     def research_target(self) -> Task:
+        
+        
         return Task(
             config=self.tasks_config['research_target'],
+            guardrail=validate_researcher_content,
         )
 
     @task

pyproject.toml

@@ -10,6 +10,8 @@ dependencies = [
     "openai-agents>=0.0.6",
     "python-dotenv>=1.0.1",
     "crewai[tools]>=0.119.0,<1.0.0",
+    "requests>=2.32.3",
+    "beautifulsoup4>=4.13.4",
 ]
 
 [dependency-groups]

uv.lock

@@ -2308,11 +2308,13 @@ name = "osint-agent"
 version = "0.1.0"
 source = { virtual = "." }
 dependencies = [
+    { name = "beautifulsoup4" },
     { name = "crewai", extra = ["tools"] },
     { name = "gradio" },
     { name = "openai" },
     { name = "openai-agents" },
     { name = "python-dotenv" },
+    { name = "requests" },
 ]
 
 [package.dev-dependencies]
@@ -2322,11 +2324,13 @@ dev = [
 
 [package.metadata]
 requires-dist = [
+    { name = "beautifulsoup4", specifier = ">=4.13.4" },
     { name = "crewai", extras = ["tools"], specifier = ">=0.119.0,<1.0.0" },
     { name = "gradio", specifier = ">=5.22.0" },
     { name = "openai", specifier = ">=1.68.2" },
     { name = "openai-agents", specifier = ">=0.0.6" },
     { name = "python-dotenv", specifier = ">=1.0.1" },
+    { name = "requests", specifier = ">=2.32.3" },
 ]
 
 [package.metadata.requires-dev]