Ashby MCP Server

Instructions

Implementation Reference

• src/ashby/server.py:407-418 (handler)

  The handle_call_tool function serves as the central handler for all tools, including 'candidate_list'. It maps the tool name to an Ashby API endpoint via TOOL_ENDPOINT_MAP and executes the request.

  @server.call_tool()
  async def handle_call_tool(name: str, arguments: dict[str, Any]) -> list[types.TextContent]:
      """Route tool calls to the correct Ashby endpoint, passing arguments directly."""
      endpoint = TOOL_ENDPOINT_MAP.get(name)
      if not endpoint:
          return [types.TextContent(type="text", text=f"Unknown tool: {name}")]
  
      try:
          # Pass arguments straight through -- tool schemas already use Ashby's
          # camelCase param names so no translation is needed.
          response = ashby.post(endpoint, data=arguments if arguments else None)
          return [types.TextContent(type="text", text=json.dumps(response, indent=2))]

• src/ashby/server.py:112-122 (schema)

  Registration and input schema definition for the 'candidate_list' tool.

  types.Tool(
      name="candidate_list",
      description="List all candidates with cursor pagination.",
      inputSchema={
          "type": "object",
          "properties": {
              "limit": {"type": "integer", "description": "Max results per page (default/max 100)"},
              "cursor": {"type": "string", "description": "Cursor for next page"},
          },
      },
  ),

• src/ashby/server.py:374-378 (registration)

  Definition of the mapping between the 'candidate_list' tool name and the underlying Ashby API endpoint (/candidate.list).

  TOOL_ENDPOINT_MAP = {
      "job_list": "/job.list",
      "job_info": "/job.info",
      "job_search": "/job.search",
      "candidate_list": "/candidate.list",