{"id":1635,"date":"2026-06-10T17:36:23","date_gmt":"2026-06-10T08:36:23","guid":{"rendered":"https:\/\/sheltie-garage.xyz\/tech\/?p=1635"},"modified":"2026-06-10T17:36:52","modified_gmt":"2026-06-10T08:36:52","slug":"python%e3%83%aa%e3%82%b9%e3%82%ad%e3%83%aa%e3%83%b3%e3%82%b0%ef%bc%9a%e7%b0%a1%e5%8d%98%e3%81%aamcp%e3%82%b5%e3%83%bc%e3%83%90%e3%83%bc%e3%82%92%e4%bd%9c%e3%81%a3%e3%81%a6%e3%81%bf%e3%82%8b","status":"publish","type":"post","link":"https:\/\/sheltie-garage.xyz\/tech\/2026\/06\/python%e3%83%aa%e3%82%b9%e3%82%ad%e3%83%aa%e3%83%b3%e3%82%b0%ef%bc%9a%e7%b0%a1%e5%8d%98%e3%81%aamcp%e3%82%b5%e3%83%bc%e3%83%90%e3%83%bc%e3%82%92%e4%bd%9c%e3%81%a3%e3%81%a6%e3%81%bf%e3%82%8b\/","title":{"rendered":"python\u30ea\u30b9\u30ad\u30ea\u30f3\u30b0\uff1a\u7c21\u5358\u306aMCP\u30b5\u30fc\u30d0\u30fc\u3092\u4f5c\u3063\u3066\u307f\u308b"},"content":{"rendered":"\n

\u7121\u8077\u671f\u9593\u3067\u30c8\u30ec\u30f3\u30c9\u6280\u8853\u306e\u5438\u53ce\u3068\u304b\u3001\u30ea\u30b9\u30ad\u30ea\u30f3\u30b0\u3092\u3061\u307e\u3061\u307e\u884c\u3063\u3066\u3044\u307e\u3059<\/p>\n\n\n\n

\u4eca\u56de\u306f\u4ee5\u524d\u304b\u3089\u6c17\u306b\u306a\u3063\u3066\u3044\u305f\u300cLangChain, MCP\u300d\u3092\u89e6\u3063\u3066\u3044\u304d\u307e\u3059
\u7c21\u5358\u306a\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u3092\u4f5c\u6210\u3057\u3001\u30cf\u30f3\u30ba\u30aa\u30f3\u3067\u9032\u3081\u308b\u3053\u3068\u306b\u3057\u307e\u3057\u305f<\/p>\n\n\n\n

\u4f5c\u308b\u3082\u306e<\/h2>\n\n\n\n

\u96fb\u5353\u3067\u8a08\u7b97\u304c\u3067\u304d\u3066\u3001\u5929\u6c17\u3092\u805e\u3051\u3070\u5929\u6c17\u3092\u8fd4\u3057\u3066\u304f\u308c\u308b\u3001\u305d\u3093\u306aAI\u30a2\u30b7\u30b9\u30bf\u30f3\u30c8(\u7121\u80fd\u30fb\u30fb\u30fb)\u3092\u4f5c\u308a\u307e\u3059<\/p>\n\n\n\n

Step1.\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u6e96\u5099<\/h2>\n\n\n\n

\u4ee5\u4e0b\u306e\u30b3\u30de\u30f3\u30c9\u3092\u9806\u756a\u306b\u5b9f\u884c\u3057\u307e\u3059<\/p>\n\n\n\n

mkdir ai-agent-sample\ncd ai-agent-sample\n\nuv init\n\nuv add langchain\nuv add langchain-openai\nuv add mcp\nuv add python-dotenv<\/code><\/pre><\/div>\n\n\n\n
uv\u306fpython\u306e\u30d1\u30c3\u30b1\u30fc\u30b8\u7ba1\u7406 & \u5b9f\u884c\u74b0\u5883\u3067\u3001pip\u3068venv\u304c\u6df7\u3056\u3063\u305f\u3088\u3046\u306a\u30c4\u30fc\u30eb
\u300cuv init\u300d\u3067\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u3092\u521d\u671f\u5316\u3057\u3001\u521d\u671f\u30d5\u30a1\u30a4\u30eb\u304c\u81ea\u52d5\u751f\u6210\u3055\u308c\u308b
\u30e9\u30a4\u30d6\u30e9\u30ea\u3092\u30a4\u30f3\u30b9\u30c8\u30fc\u30eb\u3059\u308b\u3068\u304d\u306f\u300cuv add\u300d\u3067\u8ffd\u52a0\u3059\u308b\u3053\u3068\u3067\u3001pyproject.toml\u306b\u4f9d\u5b58\u95a2\u4fc2\u304c\u4fdd\u5b58\u3055\u308c\u308b
\u5b9f\u884c\u6642\u306f\u300cuv run xxx.py\u300d\u3067\u5b9f\u884c\u53ef\u80fd<\/p>\n\n\n\n
Step2.MCP\u30b5\u30fc\u30d0\u30fc\u4f5c\u6210<\/h2>\n\n\n\n
\u30b7\u30f3\u30d7\u30eb\u306aMCP\u30b5\u30fc\u30d0\u30fc\u3092\u4f5c\u6210\u3057\u3001\u8d77\u52d5\u3057\u307e\u3059<\/p>\n\n\n\n
\u30d7\u30ed\u30b8\u30a7\u30af\u30c8\u76f4\u4e0b\u306b\u300cserver.py\u300d\u3092\u4f5c\u6210\u3057\u3001\u4ee5\u4e0b\u306e\u5185\u5bb9\u3092\u8a18\u8ff0\u3057\u307e\u3057\u305f
(\u30c7\u30d5\u30a9\u30eb\u30c8\u3067\u4f5c\u6210\u3055\u308c\u308bmain.py\u306f\u524a\u9664)<\/p>\n\n\n\n
from mcp.server.fastmcp import FastMCP\n\nmcp = FastMCP("sample-agent")\n\n# \u8db3\u3057\u7b97\u30c4\u30fc\u30eb\n@mcp.tool()\ndef add(a: int, b: int) -> int:\n    return a + b;\n\n#\u304a\u5929\u6c17\u30c4\u30fc\u30eb\n@mcp.tool()\ndef wether(city: str) -> str:\n    return f"{city}\u306f\u6674\u308c\u3067\u3059"\n\nif __name__ == "__main__":\n    mcp.run()\n<\/code><\/pre><\/div>\n\n\n\n
FastMCP\u306b\u6e21\u3055\u308c\u308b\u5f15\u6570\u306f\u3001mcp\u30b5\u30fc\u30d0\u30fc\u306e\u540d\u524d\u306e\u3088\u3046\u306a\u6271\u3044
@mcp.tool\u30c7\u30b3\u30ec\u30fc\u30bf\u30fc\u3067\u3001\u30c4\u30fc\u30eb\u3068\u3057\u3066\u5916\u90e8\u516c\u958b\u3055\u308c\u308b<\/p>\n\n\n\n
Step3.MCP\u30b5\u30fc\u30d0\u30fc\u8d77\u52d5<\/h2>\n\n\n\n
\u30b5\u30fc\u30d0\u30fc\u8d77\u52d5\u306f\u4ee5\u4e0b\u306e\u30b3\u30de\u30f3\u30c9\u3067\u5b9f\u65bd
\u4e00\u5148\u305a\u3001\u4f55\u3082\u8868\u793a\u3055\u308c\u306a\u3051\u308c\u3070\u8d77\u52d5\u6210\u529f<\/p>\n\n\n\n
uv run python server.py<\/code><\/pre><\/div>\n\n\n\n
\u30b5\u30fc\u30d0\u30fc\u304c\u8d77\u52d5\u3057\u305f\u3089\u3001\u5225\u30bf\u30fc\u30df\u30ca\u30eb\u3067\u4ee5\u4e0b\u3092\u5b9f\u884c<\/p>\n\n\n\n
npx @modelcontextprotocol\/inspector<\/code><\/pre><\/div>\n\n\n\n
\u30d6\u30e9\u30a6\u30b6\u304c\u7acb\u3061\u4e0a\u304c\u308b\u306e\u3067\u3001\u4ee5\u4e0b\u306e\u3088\u3046\u306b\u8a2d\u5b9a\u3057\u3066\u30c4\u30fc\u30eb\u3092\u78ba\u8a8d\u3059\u308b
\u4ee5\u4e0b\u306e\u3088\u3046\u306b\u5165\u529b\u3092\u884c\u3044\u3001Connect\u30dc\u30bf\u30f3\u3067MCP\u30b5\u30fc\u30d0\u30fc\u306b\u63a5\u7d9a\u3059\u308b<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
server.py\u3067\u4f5c\u6210\u3057\u305f\u30c4\u30fc\u30eb\u304c\u4e00\u89a7\u3067\u8868\u793a\u3055\u308c\u307e\u3057\u305f
add\u3092\u30af\u30ea\u30c3\u30af\u3059\u308b\u3068\u53f3\u5074\u306b\u30c6\u30b9\u30c8\u3059\u308b\u305f\u3081\u306e\u30a4\u30f3\u30bf\u30fc\u30d5\u30a7\u30fc\u30b9\u304c\u8868\u793a\u3055\u308c\u308b<\/p>\n\n\n\n
\"\"<\/figure>\n\n\n\n
Step4.MCP\u30af\u30e9\u30a4\u30a2\u30f3\u30c8\u306e\u4f5c\u6210<\/h2>\n\n\n\n
client.py\u3092\u4f5c\u6210\u3057\u3066\u3001\u30c4\u30fc\u30eb\u4e00\u89a7\u306e\u53d6\u5f97\u3068add\u306e\u5b9f\u884c\u3092\u884c\u3063\u3066\u307f\u307e\u3059<\/p>\n\n\n\n
import asyncio\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\nasync def main():\n    server_params = StdioServerParameters(\n        command="uv",\n        args=["run", "python", "server.py"]\n    )\n\n    async with stdio_client(server_params) as (read, write):\n        async with ClientSession(read, write) as session:\n            await session.initialize()\n\n            tools = await session.list_tools()\n\n            print("===== Tooles =====")\n                  \n            for tool in tools.tools:\n                print(tool.name)\n\n            print("===== Add =====")\n\n            result = await session.call_tool(\n                "add",\n                {\n                    "a": 1,\n                    "b": 2,\n                }\n            )\n            print(result)\n                    \n\nif __name__ == "__main__":\n    asyncio.run(main())<\/code><\/pre><\/div>\n\n\n\n
\u51fa\u529b\u7d50\u679c\u4f8b\u3067\u3059<\/p>\n\n\n\n
 uv run python client.py\nProcessing request of type ListToolsRequest\n===== Tooles =====\nadd\nweather\n===== Add =====\nProcessing request of type CallToolRequest\nmeta=None content=[TextContent(type='text', text='3', annotations=None, meta=None)] structuredContent={'result': 3} isError=False<\/code><\/pre><\/div>\n\n\n\n
\u5c11\u3057\u898b\u306b\u304f\u3044\u3067\u3059\u304c\u30011+2\u304c\u8a08\u7b97\u3055\u308c\u30663\u304c\u8fd4\u5374\u3055\u308c\u3066\u3044\u307e\u3059
\u73fe\u5728\u306f\u30c6\u30b9\u30c8\u306e\u305f\u3081\u3001python\u30d7\u30ed\u30b0\u30e9\u30e0\u3067\u76f4\u63a5mcp\u306eadd\u3092\u8aad\u3093\u3067\u3044\u307e\u3059\u304c\u3001\u6b21\u306e\u30b9\u30c6\u30c3\u30d7\u3067LangChain\u306b\u7f6e\u304d\u63db\u3048\u308b\u3053\u3068\u3067\u3001LLM\u3068MCP\u30b5\u30fc\u30d0\u30fc\u306e\u9023\u643a\u304c\u5b9f\u73fe\u3057\u307e\u3059<\/p>\n\n\n\n
Step5.LLM\u30e2\u30c7\u30eb\u3092\u547c\u3073\u51fa\u3059<\/h2>\n\n\n\n
\u30ed\u30fc\u30ab\u30eb\u306eOllama\u3092\u5229\u7528\u3057\u3066MCP\u3092\u5b9f\u9a13\u3057\u307e\u3059
\u307e\u305a\u306fOllama\u306b\u5bfe\u3057\u3066\u30c1\u30e3\u30c3\u30c8\u3092\u9001\u3063\u3066\u307f\u307e\u3059<\/p>\n\n\n\n
llm_test.py<\/p>\n\n\n\n
from langchain_ollama import ChatOllama\n\nllm = ChatOllama(\n    model="gemma4:e2b"\n)\n\nresponse = llm.invoke("\u3053\u3093\u306b\u3061\u306f")\n\nprint(response.content)<\/code><\/pre><\/div>\n\n\n\n
\u5b9f\u884c\u3059\u308b\u3068\u3001\u300c\u3053\u3093\u306b\u3061\u306f\u3002\u300d\u306e\u3088\u3046\u306a\u6328\u62f6\u304c\u8fd4\u3063\u3066\u304f\u308b<\/p>\n\n\n\n
\u7d9a\u3044\u3066\u3001LangChain\u306e\u52d5\u304d\u3082\u898b\u3066\u307f\u307e\u3059<\/p>\n\n\n\n
tool_test.py<\/p>\n\n\n\n
from langchain.tools import tool\n\n@tool\ndef weather(city: str) -> str:\n    """\u6307\u5b9a\u3057\u305f\u5e74\u306e\u5929\u6c17\u3092\u5e30\u3057\u307e\u3059"""\n    \n    print(f"[TOOL\u5b9f\u884c] weather({city})")\n\n    return f"{city}\u306f\u6674\u308c\u3067\u3059"\n\nprint(weather.name)\nprint(weather.description)\nprint(weather.args)<\/code><\/pre><\/div>\n\n\n\n
\u5b9f\u884c\u7d50\u679c\u306f\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u611f\u3058<\/p>\n\n\n\n
weather\n\u6307\u5b9a\u3057\u305f\u5e74\u306e\u5929\u6c17\u3092\u5e30\u3057\u307e\u3059\n{'city': {'title': 'City', 'type': 'string'}}<\/code><\/pre><\/div>\n\n\n\n
@tool\u3092\u5229\u7528\u3059\u308b\u3068\u3001docstring\u304c\u5fc5\u9808\u3068\u306a\u308b\u3088\u3046\u3067\u3001\u30c9\u30ad\u30e5\u30e1\u30f3\u30c8\u3092\u6307\u5b9a\u3057\u306a\u3044\u3068\u5b9f\u884c\u6642\u30a8\u30e9\u30fc\u3068\u306a\u308a\u307e\u3057\u305f<\/p>\n\n\n\n
\u6b21\u306b\u6700\u5c0f\u9650\u306e\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u3092\u6e96\u5099\u3057\u3066\u3001LLM\u304c\u4f5c\u6210\u3057\u305f\u30c4\u30fc\u30eb\u3092\u9078\u629e\u3059\u308b\u69d8\u5b50\u3092\u89b3\u5bdf\u3057\u307e\u3059<\/p>\n\n\n\n
agent_test.py<\/p>\n\n\n\n
from langchain.tools import tool\nfrom langchain_ollama import ChatOllama\n\n@tool\ndef weather(city: str) -> str:\n    """\u6307\u5b9a\u3057\u305f\u90fd\u5e02\u306e\u5929\u6c17\u3092\u5e30\u3057\u307e\u3059"""\n\n    print(f"[Tool\u5b9f\u884c] weather({city})")\n\n    return f"{city}\u306f\u6674\u308c\u3067\u3059"\n\nllm = ChatOllama(\n    model="gemma4:e2b"\n)\n\nllm_with_tools = llm.bind_tools([weather])\n\nresponse = llm_with_tools.invoke(\n    "\u6771\u4eac\u306e\u5929\u6c17\u3092\u6559\u3048\u3066"\n)\n\nprint(response)<\/code><\/pre><\/div>\n\n\n\n
\u5b9f\u884c\u7d50\u679c\u306f\u4ee5\u4e0b\u306e\u901a\u308a<\/p>\n\n\n\n
ai-agent-sample) MacBook-Pro ai-agent-sample % uv run python agent_test.py\ncontent='' additional_kwargs={} response_metadata={'model': 'gemma4:e2b', 'created_at': '2026-06-10T07:06:42.457207Z', 'done': True, 'done_reason': 'stop', 'total_duration': 15426462624, 'load_duration': 518820958, 'prompt_eval_count': 64, 'prompt_eval_duration': 1254592000, 'eval_count': 187, 'eval_duration': 13640234000, 'logprobs': None, 'model_name': 'gemma4:e2b', 'model_provider': 'ollama'} id='lc_run--019eb05a-9549-73f0-ae2e-f5d83fdafe93-0' tool_calls=[{'name': 'weather', 'args': {'city': '\u6771\u4eac'}, 'id': '90d50f05-c77b-425f-a341-cfe6accac6c3', 'type': 'tool_call'}] invalid_tool_calls=[] usage_metadata={'input_tokens': 64, 'output_tokens': 187, 'total_tokens': 251}<\/code><\/pre><\/div>\n\n\n\n
\u30ec\u30b9\u30dd\u30f3\u30b9\u3067tool_calls\u304c\u8fd4\u5374\u3055\u308c\u3066\u3044\u3066\u3001\u30c4\u30fc\u30eb\u3092\u5229\u7528\u3057\u3088\u3046\u3068\u3057\u305f\u3053\u3068\u304c\u78ba\u8a8d\u3067\u304d\u308b<\/p>\n\n\n\n
Step6.LLM\u304b\u3089\u547c\u3073\u51fa\u3057<\/h2>\n\n\n\n
\u3044\u3088\u3044\u3088LLM\u304b\u3089MCP\u30b5\u30fc\u30d0\u30fc\u306b\u8a2d\u5b9a\u3057\u305f\u30c4\u30fc\u30eb\u3092\u547c\u3073\u51fa\u3057\u307e\u3059
\u307e\u305a\u306f\u624b\u52d5\u3067\u5229\u7528\u3059\u308btool\u3092\u6307\u5b9a\u3059\u308b\u30d1\u30bf\u30fc\u30f3<\/p>\n\n\n\n
import asyncio\n\nfrom langchain.tools import tool\nfrom langchain_ollama import ChatOllama\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\n@tool\ndef weather(city: str) -> str:\n    """\u6307\u5b9a\u3057\u305f\u90fd\u5e02\u306e\u5929\u6c17\u3092\u5e30\u3057\u307e\u3059"""\n    return "dummy"\n\nasync def main():\n    llm = ChatOllama(\n        model="gemma4:e2b"\n    )\n\n    llm_with_tools = llm.bind_tools([weather])\n\n    response = llm_with_tools.invoke(\n        "\u6771\u4eac\u306e\u5929\u6c17\u3092\u6559\u3048\u3066"\n    )\n\n    print(response.tool_calls)\n\n    tool_call = response.tool_calls[0]\n\n    server_params = StdioServerParameters(\n        command="uv",\n        args=["run", "python", "server.py"],\n    )\n\n    async with stdio_client(server_params) as (read, write):\n        async with ClientSession(read, write) as session:\n            await session.initialize()\n\n            result = await session.call_tool(\n                tool_call["name"],\n                tool_call["args"]\n            )\n\n            print(result)\n\nif __name__ == "__main__":\n    asyncio.run(main())<\/code><\/pre><\/div>\n\n\n\n
\u30d7\u30ed\u30b0\u30e9\u30e0\u5185\u3067\u8a2d\u5b9a\u3055\u308c\u3066\u3044\u308bweather\u30c4\u30fc\u30eb\u306f\u30c0\u30df\u30fc\u95a2\u6570\u3068\u306a\u3063\u3066\u3044\u3066\u3001dummy\u304c\u8fd4\u5374\u3055\u308c\u307e\u3059
\u3067\u3082\u3001\u4eca\u30d7\u30ed\u30b0\u30e9\u30e0\u3092\u52d5\u304b\u3059\u3068<\/p>\n\n\n\n
ai-agent-sample % uv run python agent_test.py \n[{'name': 'weather', 'args': {'city': '\u6771\u4eac'}, 'id': '366711c7-4dfb-4566-9dca-c279b9649423', 'type': 'tool_call'}]\nProcessing request of type CallToolRequest\nProcessing request of type ListToolsRequest\nmeta=None content=[TextContent(type='text', text='\u6771\u4eac\u306f\u6674\u308c\u3067\u3059', annotations=None, meta=None)] structuredContent={'result': '\u6771\u4eac\u306f\u6674\u308c\u3067\u3059'} isError=False<\/code><\/pre><\/div>\n\n\n\n
\u304d\u3061\u3093\u3068\u300c\u6771\u4eac\u306f\u6674\u308c\u3067\u3059\u300d\u3068\u3044\u3046\u30c6\u30ad\u30b9\u30c8\u304c\u8fd4\u5374\u3055\u308c\u3066\u3044\u307e\u3059
\u3053\u308c\u306f\u3001@tool\u3067\u6307\u5b9a\u3055\u308c\u3066\u3044\u308bweather\u306f\u300c\u540d\u79f0\u3084\u30d1\u30e9\u30e1\u30fc\u30bf\u306a\u3069\u306e\u547c\u3073\u51fa\u3057\u60c5\u5831\u306e\u307f\u300d\u53d6\u5f97\u3059\u308b\u305f\u3081\u306b\u5229\u7528\u3055\u308c\u3001\u5b9f\u969b\u306b\u306f\u5229\u7528\u3055\u308c\u3066\u3044\u306a\u3044\u305f\u3081\u3067\u3059<\/p>\n\n\n\n
\u5b9f\u969b\u306b\u306fMCP\u30b5\u30fc\u30d0\u30fc\u306b\u8a2d\u5b9a\u3057\u305fweather\u30c4\u30fc\u30eb\u304c\u5b9f\u884c\u3055\u308c\u3066\u3044\u308b\u305f\u3081\u3001\u6b63\u3057\u3044\u6587\u7ae0\u304c\u8fd4\u5374\u3055\u308c\u3066\u3044\u307e\u3059<\/p>\n\n\n\n
Step7.\u30c4\u30fc\u30eb\u306e\u81ea\u52d5\u8a2d\u5b9a<\/h2>\n\n\n\n
MCP\u30b5\u30fc\u30d0\u30fc\u304b\u3089\u30c4\u30fc\u30eb\u4e00\u89a7\u3092\u53d6\u5f97\u3057\u3001\u81ea\u52d5\u3067\u8a2d\u5b9a\u3067\u304d\u308b\u3088\u3046\u306b\u3057\u307e\u3059<\/p>\n\n\n\n
uv add langchain-mcp-adapters<\/code><\/pre><\/div>\n\n\n\n
\u4ee5\u4e0b\u306e\u3088\u3046\u306a\u30b3\u30fc\u30c9\u3068\u3057\u307e\u3057\u305f<\/p>\n\n\n\n
agent_test.py<\/p>\n\n\n\n
import asyncio\n\nfrom langchain.tools import tool\nfrom langchain_ollama import ChatOllama\nfrom langchain_mcp_adapters.tools import load_mcp_tools\n\nfrom mcp import ClientSession, StdioServerParameters\nfrom mcp.client.stdio import stdio_client\n\n@tool\ndef weather(city: str) -> str:\n    """\u6307\u5b9a\u3057\u305f\u90fd\u5e02\u306e\u5929\u6c17\u3092\u5e30\u3057\u307e\u3059"""\n    return "dummy"\n\nasync def main():\n    llm = ChatOllama(\n        model="gemma4:e2b"\n    )\n\n    server_params = StdioServerParameters(\n        command="uv",\n        args=["run", "python", "server.py"],\n    )\n\n    async with stdio_client(server_params) as (read, write):\n        async with ClientSession(read, write) as session:\n            await session.initialize()\n\n            # \u30c4\u30fc\u30eb\u4e00\u89a7\u306e\u53d6\u5f97\n            tools = await load_mcp_tools(session)\n            llm_with_tools = llm.bind_tools(tools)\n\n            response = llm_with_tools.invoke(\n                "1+2\u3092\u8a08\u7b97\u3057\u3066\u304f\u3060\u3055\u3044"\n            )\n\n            print("===== Tool Calls =====")\n            print(response.tool_calls)\n\n            tool_call = response.tool_calls[0]\n\n            result = await session.call_tool(\n                tool_call["name"],\n                tool_call["args"]\n            )\n\n            weather_result = result.content[0].text\n\n            print("===== Tool Result =====")\n            print(weather_result)\n\n            print("===== Final Content =====")\n            final_response = llm.invoke(\n                f"""\n                \u30e6\u30fc\u30b6\u30fc\u306e\u8cea\u554f:\n                \u8a08\u7b97\u3092\u884c\u3063\u3066\n\n                \u30c4\u30fc\u30eb\u5b9f\u884c\u7d50\u679c\n                {weather_result}\n\n                \u30c4\u30fc\u30eb\u5b9f\u884c\u7d50\u679c\u3092\u4f7f\u3063\u3066\u56de\u7b54\u3057\u3066\u304f\u3060\u3055\u3044\n                """\n            )\n            print(final_response.content)\n\nif __name__ == "__main__":\n    asyncio.run(main())<\/code><\/pre><\/div>\n\n\n\n
LLM\u306b\u6295\u3052\u308b\u30d7\u30ed\u30f3\u30d7\u30c8\u306f\u30d7\u30ed\u30b0\u30e9\u30e0\u5185\u306b\u56fa\u5b9a\u306a\u306e\u3067\u3001\u30c6\u30b9\u30c8\u3059\u308b\u3068\u304d\u306f\u76f4\u63a5\u5909\u3048\u3066\u304f\u3060\u3055\u3044\u30fb\u30fb\u30fb<\/p>\n\n\n\n
\u5929\u6c17\u3092\u805e\u3044\u3066\u898b\u305f\u3068\u304d\u306e\u7d50\u679c<\/p>\n\n\n\n
MacBook-Pro ai-agent-sample % uv run python agent_test.py  \nProcessing request of type ListToolsRequest\n===== Tool Calls =====\n[{'name': 'weather', 'args': {'city': '\u6771\u4eac'}, 'id': '243ede14-c039-4f69-b5d4-840fd12cbec4', 'type': 'tool_call'}]\nProcessing request of type CallToolRequest\n===== Tool Result =====\n\u6771\u4eac\u306f\u6674\u308c\u3067\u3059\n===== Final Content =====\n\u6771\u4eac\u306f\u6674\u308c\u3067\u3059\u3002<\/code><\/pre><\/div>\n\n\n\n
\u8a08\u7b97\u3092\u884c\u308f\u305b\u3066\u307f\u305f\u3068\u304d\u306e\u7d50\u679c<\/p>\n\n\n\n
MacBook-Pro ai-agent-sample % uv run python agent_test.py\nProcessing request of type ListToolsRequest\n===== Tool Calls =====\n[{'name': 'add', 'args': {'a': 1, 'b': 2}, 'id': 'ac56a447-3786-41d7-9df9-7dbd5e6f57ce', 'type': 'tool_call'}]\nProcessing request of type CallToolRequest\n===== Tool Result =====\n3\n===== Final Content =====\n\u6050\u308c\u5165\u308a\u307e\u3059\u304c\u3001**\u3069\u306e\u3088\u3046\u306a\u8a08\u7b97\u3092\u3059\u308c\u3070\u826f\u3044\u304b**\u3001\u307e\u305f\u306f**\u8a08\u7b97\u306b\u5fc5\u8981\u306a\u60c5\u5831\uff08\u554f\u984c\u3084\u30c7\u30fc\u30bf\uff09**\u3092\u3054\u63d0\u793a\u304f\u3060\u3055\u3044\u3002\n\n\u300c\u30c4\u30fc\u30eb\u5b9f\u884c\u7d50\u679c\u300d\u3068\u3057\u3066\u300c3\u300d\u304c\u63d0\u4f9b\u3055\u308c\u307e\u3057\u305f\u304c\u3001\u3053\u306e\u6570\u5b57\u3092\u4f7f\u3063\u3066\u4f55\u3092\u8a08\u7b97\u3059\u308c\u3070\u3088\u3044\u306e\u304b\u304c\u5206\u304b\u308a\u307e\u305b\u3093\u3002\n\n\u5177\u4f53\u7684\u306a\u3054\u8cea\u554f\u3092\u304a\u77e5\u3089\u305b\u3044\u305f\u3060\u3051\u308c\u3070\u3001\u559c\u3093\u3067\u8a08\u7b97\u3044\u305f\u3057\u307e\u3059\u3002<\/code><\/pre><\/div>\n\n\n\n
\u8a08\u7b97\u7d50\u679c\u306e\u5fdc\u7b54\u7d50\u679c\u304c\u304a\u304b\u3057\u3044\u3067\u3059\u304c\u3001\u8a08\u7b97\u81ea\u4f53\u306f\u3042\u3063\u3066\u3044\u307e\u3057\u305f<\/p>\n\n\n\n
\u4ee5\u4e0a<\/h2>\n\n\n\n
\u3068\u3044\u3046\u3053\u3068\u3067\u3001\u30b7\u30f3\u30d7\u30eb\u306aMCP\u30b5\u30fc\u30d0\u30fc\u3092\u4f5c\u3063\u3066\u307f\u308b\u5185\u5bb9\u3067\u3057\u305f
LLM\u304b\u3089MCP\u30b5\u30fc\u30d0\u30fc\u306e\u30c4\u30fc\u30eb\u304c\u547c\u3070\u308c\u308b\u5224\u65ad\u306fLLM\u5074\u3067\u884c\u3063\u3066\u3044\u308b\u3068\u3044\u3046\u3053\u3068\u306f\u3061\u3087\u3063\u3068\u9a5a\u304d\u3067\u3057\u305f<\/p>\n\n\n\n
MCP\u306e\u4ed5\u7d44\u307f\u306f\u306a\u3093\u3068\u306a\u304f\u7406\u89e3\u3067\u304d\u305f\u306e\u3067\u3001\u6b21\u306fBlenderMCP\u306a\u3069\u306e\u3044\u308d\u3044\u308d\u306aMCP\u304c\u3069\u3046\u3084\u3063\u3066\u5b9f\u88c5\u3055\u308c\u3066\u3044\u308b\u306e\u304b\u8abf\u3079\u3066\u307f\u3066\u3001\u5b9f\u30a2\u30d7\u30ea\u3068\u306e\u9023\u643a\u65b9\u6cd5\u3092\u7406\u89e3\u3057\u3066\u3044\u304d\u305f\u3044\u3068\u601d\u3044\u307e\u3057\u305f<\/p>\n\n\n
\n\t
\n\t\t
\n\t\t\t\t\t\t\t<\/a>\t\t\t\t\t<\/div>\n\t\t
\n\t\t\t
\n\t\t\t\t\t\t\t\t\t\u30bc\u30ed\u304b\u3089\u5b66\u3076 MCP&A2A\u30d7\u30ed\u30b0\u30e9\u30df\u30f3\u30b0\u5165\u9580 AI\u30a8\u30fc\u30b8\u30a7\u30f3\u30c8\u6642\u4ee3\u3092\u5207\u308a\u958b\u304f\u6b21\u4e16\u4ee3\u30d7\u30ed\u30c8\u30b3\u30eb \uff08Compass Books\u30b7\u30ea\u30fc\u30ba\uff09 [ \u5e03\u7559\u5ddd\u82f1\u4e00 ]<\/a>\t\t\t\t\t\t\t<\/div>\n\t\t\t
\n\t\t\t\t\t\t\t
created by Rinker<\/a><\/div>\n\t\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\u00a53,520<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t(2026\/6\/10 17:35:38\u6642\u70b9 \u697d\u5929\u5e02\u5834\u8abf\u3079-<\/span>\u8a73\u7d30)<\/span>\n\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t<\/div>\n\t\t\t\t\t\t