返回 MCP 目录
public公开dns本地运行

Higress管理服务器

一种模型上下文协议服务器,通过设计精良的代理流架构,实现对 Higress 的全面配置和管理。

article

README

Higress OPS MCP Server

A Model Context Protocol (MCP) server implementation that enables comprehensive configuration and management of Higress. This repository also provides an MCP client built on top of LangGraph and LangChain MCP Adapters, facilitating interaction with the Higress MCP Server through a well-designed agent flow architecture.

Demo

https://github.com/user-attachments/assets/bae66b77-a158-452e-9196-98060bac0df7

Config Environment Variables

Copy the .env.example file to .env and fill in the corresponding values.

Start MCP Client and MCP Server

In stdio mode, the MCP server process is started by the MCP client program. Run the following command to start the MCP client and MCP server:

uv run client.py

Add a new tool

Step 1: Create a new tool class or extend an existing one

  • Create a new file in the tools directory if adding a completely new tool category
  • Or add your tool to an existing class if it fits an existing category
from typing import Dict, List, Any
from fastmcp import FastMCP

class YourTools:
    def register_tools(self, mcp: FastMCP):
        @mcp.tool()
        async def your_tool_function(arg1: str, arg2: int) -> List[Dict]:
            """
            Your tool description.
            
            Args:
                arg1: Description of arg1
                arg2: Description of arg2

            Returns:
                Description of the return value
            
            Raises:
                ValueError: If the request fails
            """
            # Implementation using self.higress_client to make API calls
            return self.higress_client.your_api_method(arg1, arg2)

Step 2: Add a new method to HigressClient if your tool needs to interact with the Higress Console API

  • Add methods to utils/higress_client.py that encapsulate API calls
  • Use the existing HTTP methods (get, put, post) for actual API communication
def your_api_method(self, arg1: str, arg2: int) -> List[Dict]:
    """
    Description of what this API method does.
    
    Args:
        arg1: Description of arg1
        arg2: Description of arg2
        
    Returns:
        Response data
        
    Raises:
        ValueError: If the request fails
    """
    path = "/v1/your/api/endpoint"
    data = {"arg1": arg1, "arg2": arg2}
    return self.put(path, data)  # or self.get(path) or self.post(path, data)

Step 3: Register your tool class in the server

  • Add your tool class to the tool_classes list in server.py
  • This list is used by ToolsRegister to instantiate and register all tools
  • The ToolsRegister will automatically set logger and higress_client attributes
tool_classes = [
    CommonTools,
    RequestBlockTools,
    RouteTools,
    ServiceSourceTools,
    YourTools  # Add your tool class here
]

Step 4: Add your tool to SENSITIVE_TOOLS if it requires human confirmation

  • Tools in this list will require human confirmation before execution
# Define write operations that require human confirmation
SENSITIVE_TOOLS = [
    "add_route", 
    "add_service_source",
    "update_route",
    "update_request_block_plugin", 
    "update_service_source",
    "your_tool_function"  # Add your tool name here if it requires confirmation
]
help

运行方式说明

cloud

托管运行

托管运行通常表示这个 MCP Server 由服务方环境承载,用户一般按页面提供的连接方式或授权流程接入,不需要在本地长期启动一个 MCP 进程

  1. 打开服务方连接页
  2. 完成授权或复制端点
  3. 在 MCP 客户端中连接
terminal

本地运行 / 其它方式

本地运行通常需要用户在自己的电脑或服务器上安装依赖,把 server_config 复制到 MCP 客户端,并按 env_schema 补齐环境变量、密钥或其它配置

  1. 复制 server_config
  2. 安装所需依赖
  3. 补齐环境变量后重启客户端