docs: improve check_url_official tool docstring for LLM usage

app.py

@@ -61,5 +61,55 @@ async def get_public_org_domains_csv() -> str:
     response.raise_for_status()
     return response.text
 
+@mcp.tool(
+    name="check_url_official",
+    description="Check if a given URL is present in the official French government URL list (sitesgouv.txt)."
+)
+async def check_url_official(url: str) -> dict:
+    """
+    Tool: check_url_official
+    -------------------------
+    Checks if a given URL is present in the official French government URL list (sitesgouv.txt).
+
+    Parameters:
+        url (str): The URL to check. The URL can be with or without protocol (http/https), and with or without a trailing slash. Example: "https://www.gouvernement.fr" or "gouvernement.fr".
+
+    Returns:
+        dict: A dictionary with the following keys:
+            - url (str): The input URL as provided.
+            - is_official (bool): True if the normalized URL is present in the official list, False otherwise.
+
+    Normalization:
+        - The tool normalizes both the input URL and the official list by:
+            * Lowercasing all characters
+            * Removing the protocol (http:// or https://)
+            * Removing any trailing slashes
+        - This ensures that URLs are compared in a consistent way, regardless of formatting.
+
+    Example usage:
+        >>> check_url_official("https://www.gouvernement.fr/")
+        {"url": "https://www.gouvernement.fr/", "is_official": True}
+
+        >>> check_url_official("example.com")
+        {"url": "example.com", "is_official": False}
+
+    Notes for LLMs:
+        - Use this tool to verify if a website is officially recognized as a French government site.
+        - The official list is fetched live from https://www.auracom.fr/gouv/sitesgouv.txt for each call.
+        - This tool is useful for compliance, trust, or filtering tasks involving French government domains.
+    """
+    response = requests.get("https://www.auracom.fr/gouv/sitesgouv.txt", timeout=10)
+    response.raise_for_status()
+    official_urls = set(line.strip() for line in response.text.splitlines() if line.strip())
+    # Simple normalization: ignore protocol and trailing slashes
+    def normalize(u):
+        return u.lower().replace("http://", "").replace("https://", "").rstrip("/")
+    normalized_official = set(normalize(u) for u in official_urls)
+    normalized_input = normalize(url)
+    return {
+        "url": url,
+        "is_official": normalized_input in normalized_official
+    }
+
 if __name__ == "__main__":
     mcp.run(transport="sse", host="0.0.0.0", port=7860)