🚀 GitHub MCP Server

A production-ready Model Context Protocol (MCP) Server built with Python that enables AI assistants and MCP clients to interact with GitHub repositories through the GitHub REST API.

This project demonstrates how to build a custom MCP server capable of performing GitHub operations such as listing repositories, retrieving repository information, managing issues, and extending to pull requests, workflows, and code management.

📌 Features

Repository Management

• List all repositories for the authenticated user
• Retrieve repository details
• Access repository metadata

Issue Management

• List repository issues
• Create new issues
• Close existing issues (future enhancement)

MCP Integration

• Built using the official MCP Python SDK
• Compatible with MCP Inspector
• JSON-RPC based communication
• Local STDIO transport support

🏗️ Architecture

┌─────────────────────┐
│   MCP Client        │
│ (Inspector / AI)    │
└──────────┬──────────┘
           │
           ▼
┌─────────────────────┐
│   GitHub MCP Server │
│      (Python)       │
└──────────┬──────────┘
           │
           ▼
┌─────────────────────┐
│   GitHub REST API   │
└──────────┬──────────┘
           │
           ▼
┌─────────────────────┐
│   GitHub Account    │
└─────────────────────┘

📂 Project Structure

github-mcp/
│
├── server.py
├── .env
├── pyproject.toml
├── uv.lock
├── .gitignore
└── README.md

⚙️ Prerequisites

• Python 3.10+
• UV Package Manager
• GitHub Account
• GitHub Personal Access Token (PAT)
• Node.js (for MCP Inspector)

🔑 GitHub Token Setup

Generate a Personal Access Token from:

https://github.com/settings/tokens

Recommended permissions:

Repository Access
Issues: Read & Write
Contents: Read
Pull Requests: Read & Write

Create a .env file:

GITHUB_TOKEN=your_github_token_here

📦 Installation

Clone the repository:

git clone <repository-url>
cd github-mcp

Initialize project:

uv init

Install dependencies:

uv add mcp requests python-dotenv

▶️ Running the Server

Start the MCP server:

uv run server.py

The server will wait for MCP client connections.

🔍 Using MCP Inspector

Start MCP Inspector:

npx @modelcontextprotocol/inspector

Open:

http://localhost:6274

Connection Settings:

Transport : STDIO

Command:
uv

Arguments:
run server.py

Click Connect.

🛠️ Available Tools

list_repositories

Returns all repositories accessible by the authenticated user.

Example

{}

Response

[
  {
    "name": "github-mcp",
    "full_name": "username/github-mcp",
    "private": false
  }
]

get_repository

Returns details for a specific repository.

Input

{
  "owner": "username",
  "repo": "github-mcp"
}

Response

{
  "name": "github-mcp",
  "description": "GitHub MCP Server",
  "stars": 10,
  "forks": 2
}

🔄 MCP Request Flow

User
 │
 ▼
MCP Client
 │
 ▼
Tool Call
 │
 ▼
GitHub MCP Server
 │
 ▼
GitHub REST API
 │
 ▼
Response Returned

🚧 Planned Enhancements

• Create GitHub Issues
• Close Issues
• Pull Request Management
• Branch Management
• File Operations
• GitHub Actions Integration
• Repository Search
• Code Search
• OAuth Authentication
• Remote MCP Deployment

🧠 Learning Outcomes

This project demonstrates:

• Model Context Protocol (MCP)
• MCP Tools
• JSON-RPC Communication
• GitHub REST API Integration
• Authentication using PAT
• MCP Inspector Usage
• Production MCP Architecture
• Local MCP Server Development

📄 License

MIT License

👨‍💻 Author

Built as a learning and production-ready MCP project for understanding MCP architecture, GitHub integrations, and AI tooling ecosystems.