Sherlock Domains MCP

Instructions

Implementation Reference

• src/sherlock_mcp/server.py:62-72 (handler)

  MCP tool handler function for 'claim_account'. It takes an email parameter, calls the underlying Sherlock._claim_account method via get_sherlock(), and wraps the response using handle_response.
  @mcp.tool() async def claim_account(email: str): """ Links an email address to an AI agent's account for web interface access and account recovery. Important notes: - Only accounts without an existing email can be linked - Each email can only be linked to one account - This method is rarely needed since emails are also set during domain registration """ return handle_response(get_sherlock()._claim_account(email))
• src/sherlock_mcp/server.py:62-72 (schema)

  Input schema defined by function parameter 'email: str' and comprehensive docstring describing usage, constraints, and purpose.
  @mcp.tool() async def claim_account(email: str): """ Links an email address to an AI agent's account for web interface access and account recovery. Important notes: - Only accounts without an existing email can be linked - Each email can only be linked to one account - This method is rarely needed since emails are also set during domain registration """ return handle_response(get_sherlock()._claim_account(email))
• src/sherlock_mcp/server.py:62-62 (registration)

  The @mcp.tool() decorator registers the claim_account function as an MCP tool on the FastMCP server instance.
  @mcp.tool()
• src/sherlock_mcp/server.py:10-17 (helper)

  Helper function to lazily create and return the Sherlock instance used by claim_account and other tools.
  def get_sherlock(): """Get or create a Sherlock instance. We want to create the class instance inside the tool, so the init errors will bubble up to the tool and hence the MCP client instead of silently failing during the server creation. """ return Sherlock()
• src/sherlock_mcp/server.py:19-31 (helper)

  Helper function to standardize responses from Sherlock methods, used in claim_account handler.
  def handle_response(response): """ Handle responses from Sherlock methods. Sherlock methods already process the response using _handle_response, which returns either a processed JSON object for successful requests or the response object itself. """ if hasattr(response, 'status_code'): # This is a raw response object try: return response.status_code, response.json() except: return response.status_code, response.text # This is already processed data (like a dictionary) return response