Peacock for VS Code

Features • Tools • Setup • Configuring an MCP Host

MCP Server for the Peacock extension for VS Code, coloring your world, one Code editor at a time. The main goal of the project is to show how an MCP server can be used to interact with APIs.

Note: All data used by this MCP server is fetched from the official Peacock documentation.

🔧 Features

• Fetch Peacock docs: Get detailed info on Peacock.

🧰 Tools

1. fetch_peacock_docs 🔍🦸‍♂️

• Description: Fetches the Peacock for VS Code extension docs from its GitHub repository and answers questions based on the documentation
• Input:
  • prompt (query): The question about Peacock.
• Returns: Your answer!

🛠️ Setup

Install Peacock for VS Code HERE.

Running the MCP Server hosted in GitHub Copilot with VS Code Insiders

Note: If you already have the MCP server enabled with Claude Desktop, add chat.mcp.discovery.enabled: true in your VS Code settings and it will discover existing MCP server lists.

If you want to associate the MCP server with a specific repo, create a .vscode/mcp.json file with this content:

{
  "inputs": [],
  "servers": {
    "peacock-mcp": {
      "command": "npx",
      // "command": "node",
      "args": [
        "-y",
        "@johnpapa/peacock-mcp"
        // "_git/peacock-mcp/dist/index.js"
      ],
      "env": {}
    }
  }
}

If you want to associate the MCP server with all repos, add the following to your VS Code User Settings JSON:

"mcp": {
  "servers": {
    "peacock-mcp": {
      "command": "npx",
      // "command": "node",
      "args": [
        "-y",
        "@johnpapa/peacock-mcp"
        // "/Users/papa/_git/peacock-mcp/dist/index.js"
        // "_git/peacock-mcp/dist/index.js"
      ],
      "env": {}
    }
  }
}
"chat.mcp.discovery.enabled": true,

VS Code Manual Installation

Note: For quick installation, click the install buttons at the top of this README.

Using VS Code Settings

To manually install the Peacock MCP server in VS Code, follow these steps:

1. Open VS Code Settings (JSON) by pressing Cmd+Shift+P (macOS) or Ctrl+Shift+P (Windows/Linux) and searching for "Preferences: Open User Settings (JSON)"
2. Add the following JSON to your settings file:

{
  "mcp": {
    "servers": {
      "peacock-mcp": {
        "command": "npx",
        "args": ["-y", "@johnpapa/peacock-mcp"],
        "env": {}
      }
    }
  },
  "chat.mcp.discovery.enabled": true
}

Using CLI Commands

For VS Code Stable:

code --add-mcp '{"name":"peacock-mcp","command":"npx","args":["-y","@johnpapa/peacock-mcp"],"env":{}}'

For VS Code Insiders:

code-insiders --add-mcp '{"name":"peacock-mcp","command":"npx","args":["-y","@johnpapa/peacock-mcp"],"env":{}}'

Installing via Smithery

To install Peacock MCP Server for Claude Desktop automatically via Smithery:

npx -y @smithery/cli install @johnpapa/peacock-mcp --client claude

Run the MCP Server Locally with MCP Inspector

If you'd like to run MCP Inspector locally to test the server, follow these steps:

1. Clone this repository:

   git clone https://github.com/johnpapa/peacock-mcp
   

2. Install the required dependencies and build the project.

   npm install
   npm run build
   

3. (Optional) To try out the server using MCP Inspector run the following command:

   # Start the MCP Inspector
   npx @modelcontextprotocol/inspector node build/index.js
   

   Visit the MCP Inspector URL shown in the console in your browser. Change Arguments to dist/index.js and select Connect. Select List Tools to see the available tools.

Using Tools in GitHub Copilot

1. Now that the mcp server is discoverable, open GitHub Copilot and select the Agent mode (not Chat or Edits).

2. Select the "refresh" button in the Copilot chat text field to refresh the server list.

3. Select the "🛠️" button to see all the possible tools, including the ones from this repo.

4. Put a question in the chat that would naturally invoke one of the tools, for example:

   How do I set my VS Code accent colors?
   

   Note: If you see "Sorry, the response was filtered by the Responsible AI Service. Please rephrase your prompt and try again.", try running it again or rephrasing the prompt.