human_hand_tool
Leverage human capabilities to perform physical tasks by sending instructions via the MCP server. Ideal for AI assistants needing human intervention for hands-on operations.
Instructions
人間が手を使って簡単な物理的操作を実行します。
Input Schema
TableJSON Schema
| Name | Required | Description | Default |
|---|---|---|---|
| instruction | Yes |
Implementation Reference
- human_mcp/mcp_server.py:63-81 (handler)The primary handler function for the 'human_hand_tool' tool. It is registered using the @mcp.tool() decorator. The function generates a unique task ID, adds the instruction to the database via db_utils, logs the task creation, polls asynchronously for task completion using wait_for_task_completion, and returns the result.@mcp.tool() async def human_hand_tool(instruction: str, ctx: Context) -> Dict[str, str]: """人間が手を使って簡単な物理的操作を実行します。""" task_id = str(uuid.uuid4()) formatted_instruction = f"✋ 手を使って操作: {instruction}" # タスクをデータベースに追加 db_utils.add_task(task_id, formatted_instruction) # ログ出力 sys.stderr.write(f"Human task created: {task_id}. Waiting for completion...\n") # 結果を待機（非同期ポーリング） result = await wait_for_task_completion(task_id) # ログ出力 sys.stderr.write(f"Human task {task_id} completed.\n") return {"result": result}
- human_mcp/tools.py:23-40 (schema)The JSON schema definition for the 'human_hand_tool' tool, including input schema requiring an 'instruction' string and output schema providing a 'result' string. Part of the HUMAN_TOOLS list.{ "name": "human_hand_tool", "description": "人間が手を使って簡単な物理的操作を実行します。", "input_schema": { "type": "object", "properties": { "instruction": {"type": "string", "description": "実行する物理的操作の指示"} }, "required": ["instruction"] }, "output_schema": { "type": "object", "properties": { "result": {"type": "string", "description": "操作の結果"} }, "required": ["result"] } },