A demo of multi mcp server tool use trajectory evaluation Frame work.
{
"mcpServers": {
"langchain-mcp-deepeval-trajectory-eval": {
"command": "<see-readme>",
"args": []
}
}
}No install config available. Check the server's README for setup instructions.
Are you the author?
Add this badge to your README to show your security score and help users find safe servers.
A demo of multi mcp server tool use trajectory evaluation Frame work.
Is it safe?
No package registry to scan.
No authentication — any process on your machine can connect.
MIT. View license →
Is it maintained?
Last commit 5 days ago. 8 stars.
Will it work with my client?
Transport: stdio. Works with Claude Desktop, Cursor, Claude Code, and most MCP clients.
基于 LangChain 和 LangGraph 构建的 MCP (Model Context Protocol) Agent,集成高德地图导航和自定义工具服务,支持 Human-in-the-loop 人机交互确认。
# 使用 uv 安装依赖
uv sync
# 复制环境变量示例文件
cp .env.example .env
# 编辑 .env 文件,填入你的 API Key
vim .env
直接运行自定义 MCP 服务器,不可中断
# 启动自定义 MCP 服务器(提供 GPS、联系人、拨号等功能)
uv run python -m agent_call.custom_mcp_server
方式一:使用 LangGraph CLI(推荐)
# 启动 LangGraph 开发服务器
uv run langgraph dev
方式二:运行 Agent 脚本测试
# 直接运行 Agent 测试
uv run python -m agent_call.agent
.
├── README.md # 本文件
├── pyproject.toml # Python 项目配置和依赖
├── langgraph.json # LangGraph CLI 配置文件
├── uv.lock # uv 依赖锁定文件
├── .env.example # 环境变量示例
├── main.py # 项目入口文件
└── agent_call/ # Agent 核心代码包
├── __init__.py # 包初始化文件
├── agent.py # Agent 主逻辑和构建函数
├── mcp_config.py # MCP 客户端配置
├── custom_mcp_server.py # 自定义 MCP 服务器实现
└── middleware.py # Human-in-the-loop 中间件配置
LangGraph CLI 的配置文件,定义了图的入口点。
{
"dependencies": ["."],
"graphs": {
"my_mcp_agent": "agent_call/agent.py:build_graph"
},
"env": ".env"
}
graphs: 定义可用的图,键是图名称,值是 模块路径:函数名env: 指定环境变量文件路径graphs 中添加新的键值对核心 Agent 实现文件,包含 build_graph() 函数用于构建 Agent。
主要功能:
修改建议:
system_prompt 变量ChatOpenAI 的参数InMemorySaver 替换为 PostgresSaver配置 MCP 客户端连接的服务列表。
当前配置的服务:
Phone-use: 自定义 MCP 服务器(本地 8000 端口)Amap: 高德地图 MCP 服务器修改建议:
# 添加新的 MCP 服务
client = MultiServerMCPClient(
{
"新服务名": {
"url": "http://your-mcp-server/mcp",
"transport": "http",
},
# ... 现有服务
}
)
实现自定义工具的 MCP 服务器,提供三个核心工具:
| 工具名 | 功能 | 说明 |
|---|---|---|
get_gps | 获取 GPS 位置 | 返回固定坐标(模拟成都位置) |
get_contact_phone | 获取联系人电话 | 支持张三、李四、王五 |
make_call | 拨打电话 | 模拟拨号功能 |
send_sms | 发送短信 | 模拟短信发送(随机成功/失败) |
修改建议:
@mcp.tool() 装饰器定义新函数CUSTOM_MCP_PORT 环境变量CONTACTS 字典配置 Human-in-the-loop 中间件,定义哪些工具需要人工确认。
当前配置:
interrupt_on={
"get_gps": {"allowed_decisions": ["approve", "reject"]},
"get_contact_phone": True, # 允许编辑、批准、拒绝
"send_sms": True,
"make_call": {"allowed_decisions": ["approve", "reject"]},
}
修改建议:
True: 允许所有操作(批准、拒绝、编辑){"allowed_decisions": ["approve", "reject"]}: 只允许批准或拒绝Python 项目配置文件,包含:
主要依赖:
langchain>=1.2.15 - LangChain 框架langchain-mcp-adapters>=0.2.2 - MCP 适配器langchain-openai>=1.1.12 - OpenAI/Kimi 模型支持langgraph-cli[inmem]>=0.4.21 - LangGraph CLIfastmcp>=3.2.3 - MCP 服务器开发复制 .env.example 为 .env 并填写以下配置:
MOONSHOT_API_KEY=your_moonshot_api_key_here
MOONSHOT_BASE_URL=https://api.moonshot.cn/v1
MOONSHOT_MODEL=kimi-k2.5
No automated test available for this server. Check the GitHub README for setup instructions.
This server is missing a description. Tools and install config are also missing.If you've used it, help the community.
Add informationNo known vulnerabilities.
Have you used this server?
Share your experience — it helps other developers decide.
Sign in to write a review.
Dynamic problem-solving through sequential thought chains
A Model Context Protocol (MCP) server and CLI that provides tools for agent use when working on iOS and macOS projects.
The official Python SDK for Model Context Protocol servers and clients
An open-source AI agent that brings the power of Gemini directly into your terminal.
MCP Security Weekly
Get CVE alerts and security updates for Langchain Mcp Deepeval Trajectory Eval and similar servers.
Start a conversation
Ask a question, share a tip, or report an issue.
Sign in to join the discussion.