A Model Context Protocol (MCP) server that integrates GitHub Copilot CLI with MCP clients.

📣 Important Notice

Thanks to @leonardommello for the original work. This is an actively maintained fork to keep the project up-to-date with the latest Copilot features and supported models.

• 9 Tools - Interactive Copilot commands for coding assistance
• 2 Resources - Session history and management
• Full MCP Support - Compatible with Claude Desktop, Claude Code, Cline, and more
• Claude Code Plugin - Claude Code plugin with built-in workflow
• Quick Commands - Shortcut commands for rapid workflow (see below)

Add to your configuration file:

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Optional: Specify model preference

You can use the --prefer flag to choose between Claude or GPT models as defaults:

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server", "--prefer", "gpt"]
    }
  }
}

Available preferences:

• --prefer claude (default): Uses Claude models (Sonnet 4.5, Opus 4.5, Haiku 4.5)
• --prefer gpt: Uses GPT models (gpt-5.1-codex, gpt-5.1-codex-max, gpt-5.1-codex-mini, gpt-4.1)

Model defaults by preference:

Note: You can still override the default model for any tool by specifying the model parameter in individual tool calls.

You need GitHub Copilot CLI installed and authenticated:

npm install -g @github/copilot
copilot /login

Once configured, here's a simple way to get started:

In Claude Desktop (or your MCP client):

Use ask-copilot with prompt="Write a simple Hello World program in JavaScript"

Response:

console.log("Hello, World!");

That's it! You can now use all the Copilot tools through your AI client.

More examples:

Use copilot-explain to explain this code: console.log("Hello, World!");

Use copilot-suggest for task="List files in current directory"

Use copilot-debug with code="console.log(messge);" and error="ReferenceError: messge is not defined"

This repository includes ready-to-use plugins that extend functionality:

AI Collaboration Workflow Plugin - Automates a structured 5-stage development process between Claude and GitHub Copilot.

Features:

• 🔄 5-Stage Workflow: Analyze → Design → Implement → Review → Deliver
• 🤖 Smart Model Selection: Automatically selects optimal Copilot models based on task type
• 👀 Preview Mode: Shows execution plan before running
• 🔄 Recovery Mechanism: Resume interrupted workflows via session ID

Quick Install:

/plugin marketplace add Aykahshi/copilot-mcp-tool
/plugin install copilot-flow

Learn more: copilot-flow documentation

This MCP server works with any MCP-compatible client. Below are detailed setup instructions for popular AI coding assistants.

Claude Desktop is the most tested and recommended client for this MCP server.

Configuration Path:

• macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%\Claude\claude_desktop_config.json
• Linux: ~/.config/Claude/claude_desktop_config.json

Method 1: NPX (No Installation Required)

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Method 2: Global Installation

npm install -g @aykahshi/copilot-mcp-server

Then add to config:

{
  "mcpServers": {
    "copilot": {
      "command": "copilot-mcp-server"
    }
  }
}

Method 3: Local Development

{
  "mcpServers": {
    "copilot": {
      "command": "node",
      "args": ["/absolute/path/to/copilot-mcp-tool/dist/esm/copilot/index.js"]
    }
  }
}

After Setup:

1. Restart Claude Desktop
2. Look for the 🔌 icon in the bottom-right corner
3. Click it to see "copilot" in the connected servers list

Claude Code provides the fastest setup experience via CLI.

Quick Setup:

# Using npx (no installation)
claude mcp add copilot -- npx -y @aykahshi/copilot-mcp-server

# Or with global installation
npm install -g @aykahshi/copilot-mcp-server
claude mcp add copilot copilot-mcp-server

Import from Claude Desktop:

# If you already configured Claude Desktop
claude mcp add-from-claude-desktop

Verify Connection:

claude mcp list
# Should show: copilot (connected)

Usage in Chat:

/mcp  # Check server status

Cursor supports both one-click and manual installation.

Method 1: Manual Configuration

Edit ~/.cursor/mcp.json (create if it doesn't exist):

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Method 2: Settings UI

1. Open Cursor Settings (Cmd/Ctrl + ,)
2. Search for "MCP"
3. Click "Add MCP Server"
4. Name: copilot
5. Command: npx -y @aykahshi/copilot-mcp-server

Cline is a popular MCP-compatible VS Code extension.

Setup:

1. Install Cline extension from VS Code marketplace
2. Open Cline settings (gear icon in Cline panel)
3. Navigate to "MCP Servers" section
4. Add new server:

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Alternatively, edit VS Code settings.json:

{
  "cline.mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Zed has native MCP support built-in.

Configuration File: ~/.config/zed/mcp.json

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Or use Zed's UI:

1. Open Zed Settings (Cmd/Ctrl + ,)
2. Go to "Extensions" → "MCP"
3. Add server with command: npx -y @aykahshi/copilot-mcp-server

Configuration Path: ~/.windsurf/mcp.json

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Gemini CLI supports both project-wide and global MCP server installation.

Global Installation:

gemini mcp add copilot -- npx -y @aykahshi/copilot-mcp-server

Project-Specific:

# In your project directory
gemini mcp add --project copilot -- npx -y @aykahshi/copilot-mcp-server

Verify:

gemini mcp list

For IntelliJ IDEA, PyCharm, WebStorm, etc.

Setup:

1. Open Settings (Cmd/Ctrl + ,)
2. Navigate to: Tools → AI Assistant → Model Context Protocol (MCP)
3. Click "+" to add new MCP server
4. Configure:
   • Name: copilot
   • Command: npx
   • Arguments: -y @aykahshi/copilot-mcp-server

This server is compatible with any MCP-compliant client. Generic configuration:

{
  "mcpServers": {
    "copilot": {
      "command": "npx",
      "args": ["-y", "@aykahshi/copilot-mcp-server"]
    }
  }
}

Additional Compatible Clients:

• Amp - Configuration in ~/.amp/mcp.json
• Augment Code - MCP settings in IDE
• Roo Code - Via settings panel
• Qwen Coder - CLI: qwen mcp add copilot

This MCP server acts as a bridge between MCP clients and the GitHub Copilot CLI:

1. MCP Client (Claude Desktop) → Calls tool via MCP protocol
2. MCP Server (This package) → Translates to Copilot CLI command
3. GitHub Copilot CLI → Processes request and returns response
4. MCP Server → Returns formatted response to client

Benefits:

• Use Copilot's AI models directly in Claude conversations
• Maintain session history across interactions
• Access specialized Copilot features (explain, debug, review, etc.)
• No need to switch between tools

Scenario: Understanding complex code patterns

// Ask Copilot to explain
const result = arr.map(x => x * 2).filter(x => x > 10);

Response:

This code performs two operations on an array in sequence:

1. .map(x => x * 2) - Creates a new array by multiplying each element by 2
2. .filter(x => x > 10) - Filters that result to only keep values greater than 10

For example, if arr = [3, 5, 8], the result would be [16]

Scenario: Finding bugs in your code

// Buggy code
function sum(arr) {
  return arr.reduce((a) => a+b, 0);
}

Error: ReferenceError: b is not defined

Copilot's Fix:

function sum(arr) {
  return arr.reduce((a, b) => a+b, 0);
}

The reduce callback needs both the accumulator and the current value.

Scenario: Fixing SQL injection vulnerabilities

// Vulnerable code
function getUserData(id) {
  return db.query('SELECT * FROM users WHERE id = ' + id);
}

Copilot's Secure Version:

async function getUserData(id) {
  if (!id) return null;

  // Use parameterized query to prevent SQL injection
  const query = 'SELECT * FROM users WHERE id = ?';
  const user = await db.query(query, [id]);

  return user || null;
}

Key improvements: Parameterized queries, input validation, async/await

Scenario: Generate comprehensive tests

function isPrime(n) {
  if (n <= 1) return false;
  for (let i = 2; i * i <= n; i++) {
    if (n % i === 0) return false;
  }
  return true;
}

Generated Jest Tests:

describe('isPrime', () => {
  it('should return false for numbers <= 1', () => {
    expect(isPrime(-5)).toBe(false);
    expect(isPrime(0)).toBe(false);
    expect(isPrime(1)).toBe(false);
  });

  it('should return true for prime numbers', () => {
    expect(isPrime(2)).toBe(true);
    expect(isPrime(3)).toBe(true);
    expect(isPrime(97)).toBe(true);
  });

  it('should return false for composite numbers', () => {
    expect(isPrime(4)).toBe(false);
    expect(isPrime(100)).toBe(false);
  });
});

Scenario: Update legacy JavaScript to modern ES6+

// Legacy code
function processData(data) {
  var result = [];
  for (var i = 0; i < data.length; i++) {
    result.push(data[i] * 2);
  }
  return result;
}

Modern Version:

function processData(data) {
  return data.map(item => item * 2);
}

Benefits: Cleaner, modern arrow function, immutable array operation

Task: "List all files recursively and count them"

PowerShell:

(Get-ChildItem -Recurse -File).Count

Bash:

find . -type f | wc -l

General-purpose coding assistant

📝 Prompt: "Write a TypeScript function to debounce user input"

🤖 Response: Complete implementation with TypeScript types, error handling, and usage examples

Code explanation and education

📝 Code to explain:
const memoize = fn => {
  const cache = new Map();
  return (...args) => {
    const key = JSON.stringify(args);
    return cache.has(key) ? cache.get(key) : cache.set(key, fn(...args)).get(key);
  };
};

🤖 Explanation: "This is a memoization function that caches results..."

Command-line suggestions

📝 Task: "Find all TypeScript files modified in the last 7 days"

🤖 Windows: Get-ChildItem -Recurse -Filter *.ts | Where-Object {$_.LastWriteTime -gt (Get-Date).AddDays(-7)}
🤖 Linux: find . -name "*.ts" -mtime -7

Bug identification and fixes

📝 Buggy async code:
async function fetchData() {
  const data = await fetch('/api/data');
  return data.json;
}

❌ Error: "data.json is not a function"

✅ Fix: return data.json() // Add parentheses

Code quality improvements

📝 Goal: "Improve performance"

// Before
const result = array.filter(x => x > 0).map(x => x * 2);

// After
const result = array.reduce((acc, x) => {
  if (x > 0) acc.push(x * 2);
  return acc;
}, []);

💡 Benefit: Single pass through array instead of two

Automated test creation

📝 Framework: "jest"

🧪 Generates: Unit tests, edge cases, integration tests, mock data
📊 Coverage: Positive cases, negative cases, boundary conditions

Code review with focus areas

📝 Focus: ["security", "performance"]

🔍 Reviews:
- SQL injection vulnerabilities
- XSS vulnerabilities
- N+1 query problems
- Memory leaks
- Inefficient algorithms

Conversation tracking

📝 Start session: Creates unique session ID
💾 Track history: All prompts and responses saved
🔍 View history: Retrieve past conversations
♻️ Resume context: Continue previous discussions

Generate Code:

Use ask-copilot to write a Python function that validates email addresses

Explain Code:

Use copilot-explain to explain this regex: /^[a-z0-9]+@[a-z0-9]+\.[a-z]{2,}$/i

Debug:

Use copilot-debug with code="function sum(arr) { return arr.reduce((a) => a+b, 0); }" and error="ReferenceError: b is not defined"

Generate Tests:

Use copilot-test-generate with code="function isPrime(n) { return n > 1 && ![...Array(n).keys()].slice(2).some(i => n % i === 0); }" and framework="jest"

Review Code:

Use copilot-review with code="..." and focusAreas=["security", "performance"]

Session Management:

Use copilot-session-start to begin a new tracked conversation
Use copilot-session-history to view conversation history

Select from available models:

• claude-sonnet-4.5 (default)
• claude-opus-4.5
• claude-haiku-4.5
• gpt-5.1
• gpt-5.1-codex
• gpt-5.1-codex-max
• gpt-5.1-codex-mini
• gpt-5.2
• gpt-5-mini
• gpt-4.1
• gemini-3-pro-preview

You can see all available models via copilot cli with /model command.

Unlimited Model (0x cost in Copilot usage): gpt-5-mini and gpt-4.1 are available with unlimited usage for GitHub Copilot Pro and above subscriptions.

Example:

Use ask-copilot with model="claude-opus-4.5" and prompt="Explain async/await"
Use ask-copilot with model="gpt-5-mini" and prompt="Quick code review"

• Node.js: >= 22.0.0
• npm: >= 10.0.0

• GitHub Copilot subscription: Required (Get Copilot)
• GitHub Copilot CLI: Must be installed and authenticated

  npm install -g @github/copilot
  copilot /login

❌ "copilot command not found"

# Install GitHub Copilot CLI
npm install -g @github/copilot

# Verify installation
copilot --version

❌ "Not authenticated"

# Login to GitHub Copilot
copilot /login

# Follow the authentication flow in your browser

❌ "Node.js version too old"

# Check your Node.js version
node --version  # Must be >= 22.0.0

# Update Node.js
# Using nvm (recommended)
nvm install 22
nvm use 22

# Or download from nodejs.org

❌ "MCP server not responding"

# Test the server directly
npx -y @aykahshi/copilot-mcp-server

# Check Claude Desktop logs
# macOS: ~/Library/Logs/Claude/
# Windows: %APPDATA%\Claude\logs\

❌ "Permission denied" on Windows

# Run as administrator or use npx without global install
npx -y @aykahshi/copilot-mcp-server

Q: Do I need a GitHub Copilot subscription? A: Yes, this MCP server requires an active GitHub Copilot subscription and the Copilot CLI installed.

Q: Can I use this with Claude Desktop? A: Yes! This is the primary use case. Just add the configuration to claude_desktop_config.json.

Q: Does this work with VS Code? A: Yes, through the Cline extension or any other MCP-compatible VS Code extension.

Q: What's the difference between this and using Copilot directly? A: This allows you to use Copilot's capabilities within Claude conversations, combining both AI assistants.

Q: Is my code sent to both GitHub and Anthropic? A: Code you share in conversations goes through Claude's MCP protocol to Copilot CLI, which processes it according to GitHub's privacy policy.

Q: Can I use this offline? A: No, both GitHub Copilot and MCP clients require internet connection.

# Clone the repository
git clone https://github.com/Aykahshi/copilot-mcp-server.git
cd copilot-mcp-server

# Install dependencies
npm install

# Build the project
npm run build

# Run locally
npm start

# Run tests
npm test

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

src/
├── copilot/          # Main Copilot CLI integration
│   ├── index.ts      # Entry point
│   ├── server.ts     # Server configuration
│   ├── cli.ts        # Copilot CLI execution
│   ├── session.ts    # Session management
│   ├── constants.ts  # Configuration & models
│   ├── types.ts      # Type definitions
│   ├── tools/        # MCP Tools
│   └── resources/    # MCP Resources
├── server/           # MCP Server core
├── shared/           # Shared utilities
└── types.ts          # Global types

The Model Context Protocol (MCP) is an open protocol that enables AI applications to securely access data and tools from different sources. Think of it as a universal connector for AI assistants.

Key concepts:

• Servers (like this package): Provide tools, resources, and prompts
• Clients (like Claude Desktop): Use these capabilities in conversations
• Tools: Actions the AI can perform (like calling GitHub Copilot)
• Resources: Data the AI can access (like session history)
• Prompts: Templates for common workflows

Learn more at modelcontextprotocol.io

• 📦 npm Package: https://www.npmjs.com/package/@aykahshi/copilot-mcp-server
• 💻 GitHub Repository: https://github.com/Aykahshi/copilot-mcp-tool
• 🐛 Report Issues: https://github.com/Aykahshi/copilot-mcp-tool/issues
• 🤖 GitHub Copilot: https://github.com/features/copilot
• 🔗 Model Context Protocol: https://modelcontextprotocol.io

MIT License - see LICENSE file for details

Original created by Leonardo M. Mello (@leonardommello) Forked and maintained by Aykahshi (@Aykahshi)