How to connect MCP to GitHub Copilot in VS Code

Skeet integrates seamlessly with GitHub Copilot via Model Context Protocol (MCP) in Microsoft VS Code, enabling your AI coding assistant to connect with your favorite development tools and services.

What is Model Context Protocol?

Model Context Protocol (MCP) connects AI tools like GitHub Copilot in VS Code to external services through Skeet. MCP provides:

• Secure connections between AI tools and external services
• Context sharing across multiple applications
• Tool augmentation allowing AI models to perform real-world actions
• Enhanced capabilities beyond default AI functionality

With MCP, GitHub Copilot in VS Code can interact seamlessly with services like GitHub, Linear, Jira, and more, greatly enhancing your workflow.

Setting Up GitHub Copilot with Skeet

To integrate Skeet with GitHub Copilot in VS Code:

1. Obtain your Skeet MCP key from the Skeet dashboard
2. Configure VS Code to use the Skeet MCP server (VS Code MCP Documentation)
3. Connect integrations in your Skeet dashboard

Global Configuration

To configure tools globally for VS Code:

1. Open Up Add MCP Server Do this with cmd + shift + p and then search for "Add MCP Server" and press enter.

2. Select SSE MCP Server Copy and paste in your SSE URL from the dashboard or here (if you're logged in).

3. Name your mcp server There is an autogenerated name but you can change it to whatever you want.

4. Choose between User Settings or Workspace Settings. User Settings is recommended so that you don't have to setup MCP server for every workspace. Press enter to continue.

5. An mcp.json snippet will be created for you in your user settings:

{
  "mcp": {
    "servers": {
      "skeet.build": {
        "url": "https://skeet.sh/ot/YOUR_SKEET_MCP_KEY"
      }
    }
  }
}

Replace <YOUR_SKEET_MCP_KEY> with your actual Skeet MCP key from the dashboard.

6. List the MCP servers in VS Code's GitHub Copilot settings. You can do this with cmd + shift + p and then search for "List MCP Servers" and press enter.

7. Select the MCP server you just added to your GitHub Copilot settings.

8. Select the MCP Server you want to start Choose Start MCP Server and press enter.

9. Verify the connection with the output in the bottom right of VS Code.

Verifying Your Connection

After setting up your MCP connection:

1. Open your project in VS Code
2. Use GitHub Copilot to list Skeet integrations:

List my Skeet integrations

In order to use mcp tools, you must be in agent mode. Learn how to turn agent mode on here.

You should see available integrations and their statuses.

If your mcp server is setup correctly, you should see tools icons with the number of mcp tools available.

3. MCP Tool Call Click the down arrow and Allow Tool calls wihout repeatedly pressing continue. We recommend Always Allow (you can change this later).

When your copilot is in agent mode and uses an mcp tool, you should see a tool call in the output.

Best Practices

Follow these best practices:

• Connect multiple integrations for maximum utility
• Use clear prompts for AI interactions
• Verify actions before executing critical operations
• Regularly rotate MCP keys to maintain security
• Keep VS Code and Copilot updated for compatibility

Example Use Cases

GitHub Integration

Create a GitHub issue in owner/repo titled "Fix navigation bug" with body "Dropdown menu doesn't close when clicking outside."

Linear Integration

Create a Linear issue titled "User profile page" with high priority

Jira Integration

Create a Jira task in project KEY titled "Refactor authentication"

Integrations

Explore Skeet integrations compatible with VS Code and GitHub Copilot:

Troubleshooting

Common Problems

• Copilot not recognizing MCP commands: Verify MCP server in VS Code settings
• Authentication errors: Check MCP key accuracy
• Integration issues: Authorize integrations in Skeet dashboard

Getting Help

• See VSCode GitHub Copilot Troubleshooting docs
• See VSCode GitHub Copilot Agent Mode docs
• Review MCP documentation
• Contact Skeet Support for assistance