openEuler mcp-servers Repository

Introduction

MCP stands for Model Context Protocol. It aims to provide a universal protocol for foundation model context, enabling the invocation of various applications and extending the capabilities of foundation models. The openEuler mcp-servers repository stores different MCP servers, focusing on the operating system domain and improving the openEuler interaction experience together with DevStation and openEuler Intelligence.

Software Architecture

The repository contains a doc directory for documentation. Each folder under servers represents an independent MCP server.

mcp-servers/
├── servers
│   ├── oeDeploys/
│   │   ├── mcp_config.json
│   │   ├── mcp-rpm.yaml
│   │   └── src/
│   │       ├── icon.png
│   │       ├── mcp-oedp.py
│   │       ├── readme.md
│   │       └── requirements.txt
│   ├── oeGitExt/
│   └── xxxxx/
├── scripts/
└── doc/

Using oeDeploy as an example, each server directory must include:

1. MCP configuration file — contains MCP server configurations

2. Deployment configuration file — contains other related configuration files

3. Source code directory — contains the implementation code and resource files of the MCP server

The MCP Description in the doc directory provides recommended guidelines to improve the usability of MCP servers.

Usage Instructions

1. Configure an MCP client such as openEuler Intelligence, Roo Code, Cline, etc.

2. Add the MCP server configuration file to the MCP client.

Automatic Installation via yum (Recommended)

The openEuler community will package each MCP server as an RPM file. Users will be able to install them directly using yum install (coming soon).

Quick Start: How to Build Your Own MCP Server Using the MCP Python SDK

MCP Environment Setup

1. Install the uv Python management tool.

yum install -y uv

2. Install MCP.

yum install -y python3-mcp

Practical Example

1. Choose a small tool — oegitext.

oegitext is a small tool for interacting with Gitee. It can query repositories, issues, and PR information. It is preinstalled on DevStation. If not installed, you can install it using yum install oegitext after enabling the openEuler-25.03 EPOL repository. After installation, configure your Gitee token:

oegitext config -token ${access_token}

Then run oegitext show issue -p to query the issues:

2. Modify it using the Python SDK.

In the virtual environment created with uv, create a new file oegitext_mcp.py. Below is a simple example:

import subprocess
from mcp.server.fastmcp import FastMCP

mcp = FastMCP("Query issues in the openEuler community.")

@mcp.tool()
def get_my_openeuler_issue() -> str:
    """Collect statistics on the issues that I am responsible for in the openEuler community."""
    try:
        # Run the oegitext command and parse the result.
        result = subprocess.check_output(['oegitext', 'show', 'issue', '-p'], 
                                        text=True, 
                                        stderr=subprocess.STDOUT)
        
        return result
    except subprocess.CalledProcessError as e:
        return e
    except Exception as e:
        return e

if __name__ == "__main__":
    # Initialize
    mcp.run()

3. Set up the openEuler MCP environment.

Open the preinstalled vscodium on DevStation and install the Roo Code extension (openEuler Intelligence can also be used later). Configure the foundation model in settings. Here we use the DeepSeek V3 API:

API Provider: OpenAI Compatible
OpenAI Base URL: https://api.siliconflow.cn
API Key: your key
Model: Pro/deepseek-ai/DeepSeek-V3   # You can select a desired model.

In Roo Code, open the MCP Server tab and edit the global MCP configuration:

{
  "mcpServers": {
    "oegitext_mcp": {
      "command": "/usr/bin/uv",
      "args": [
        "--directory",
        "/home/xxx/oegitext_mcp",
        "run",
        "--python",
        "/usr/bin/python3",
        "oegitext_mcp.py"
      ],
      "disabled": false,
      "autoApprove": [],
      "alwaysAllow": []
    },
  }
}

Modify the paths as needed, save, and finish.

4. Check the MCP client invocation result.