Command Proxy

🚀 MCP Command Proxy Server

An MCP (Model Context Protocol) server acting as a proxy for command-line tools, specifically designed for Expo development but applicable to any command-line application.

🚀 Quick Start

How to Use in Cursor (Taking Expo as an Example)

1. Navigate to your Expo project directory.
2. Run the command npx mcp-command-proxy --prefix "ExpoServer" --command "expo start" --port 8383.
3. Open Cursor settings -> MCP -> Click "+ Add New MCP Server", as shown below:
4. Set the name to "ExpoServer", the type to "SSE", and the URL to http://localhost:8383/sse.
5. After clicking "Save", you can use the MCP server in Cursor. As shown below:

⚠️ Important Note

It is recommended to use the --port 8383 flag to avoid conflicts with other servers. You can also add the following instructions to the .cursorrules file:

💡 Usage Tip

You can use the MCP's getRecentLogs tool to obtain the latest logs of the Expo server. If necessary, use the sendKeyPress tool to send key presses to the running process.

✨ Features

• Command Proxy: Run any CLI command through the MCP server.
• Log Collection: Capture and store logs of the running process (configurable buffer size).
• Key Press Forwarding: Forward client key presses to the running process.
• Transparent Experience: End-users see the same command output as if they ran it directly.
• Interactive Command Support: Suitable for interactive CLI tools like Expo.
• MCP Integration: Built on the MCP SDK for easy integration with Claude and other MCP-enabled AI assistants.

🔧 Technical Details

How It Works

1. The server starts the specified command in a pseudo-terminal (PTY).
2. All stdout/stderr outputs are:
   • Streamed to the client in real-time.
   • Stored in a circular buffer (configurable size, default 300 lines).
3. Client key presses are forwarded to the running process.
4. The server provides the following tools:
   • View the collected logs.
   • Send key presses to the process.
   • Get the current state of the process.

📦 Use Cases

• Expo Development: Run expo start and interact with it while collecting logs.
• Build Process: Monitor the build process and analyze logs.
• Long-running Services: Monitor services and maintain recent log history.
• Remote Command Execution: Execute and monitor commands from a remote client.

📦 Installation

# Clone the repository
git clone https://github.com/hormold/mcp-command-proxy.git
cd mcp-command-proxy

# Install dependencies
pnpm install

# Build the project
pnpm build

# Run the server
pnpm start

💻 Usage Examples

Basic Usage

mcp-command-proxy [options]

Options

• --version, -v: Output the package version.
• --help, -h: Output usage help.
• --port, -p: Specify the server port, default is 3000.
• --prefix, -x: Specify the command prefix, e.g., "sudo".
• --command, -c: Specify the command to execute.

Examples

1. Run the basic server:

mcp-command-proxy --port 8080

2. Run with a command prefix:

mcp-command-proxy --prefix "sudo" --command "npm start"

📚 MCP Integration

This implementation includes the following MCP tools:

getRecentLogs

Get the recent log entries.

Parameters

• limit: The number of the latest log entries to return, default is 10.

sendKeyPress

Send a key press event to the running process.

Parameters

• key: The key to press, e.g., "ArrowUp" or "Enter".

More Examples

1. Run Expo:

mcp-command-proxy --command "expo start"

2. Use in Claude:

claude run expo start

💻 Development

# Clone the repository
git clone https://github.com/hormold/mcp-command-proxy.git
cd mcp-command-proxy

# Install dependencies
pnpm install

# Build the project
pnpm build

# Start the development mode
pnpm dev

📄 License

MIT