Q CLI MCP Server

A FastMCP server implementation that allows other AI agents to interact with Amazon Q CLI through the Model Context Protocol (MCP).

Features

• q_chat: Execute Q CLI chat commands with AI assistant
• q_whoami: Get current Q CLI user information
• q_diagnostic: Run Q CLI diagnostic tests

Package Structure

q_mcp_server/
├── q_cli_mcp/              # Main package directory
│   ├── __init__.py         # Package initialization
│   └── server.py           # MCP server implementation
├── pyproject.toml          # Package configuration
├── requirements.txt        # Dependencies
├── add_to_q.sh            # Installation script
└── README.md              # This file

Installation

Method 1: Using the installation script

./add_to_q.sh

Method 2: Manual installation

1. Install the package dependencies:

uv sync

2. Add to Q CLI:

q mcp add q-cli-proxy --command "uvx" --args "--from /path/to/q_mcp_server q-cli-mcp-server"

Verify Installation

q mcp list
q mcp status --name q-cli-proxy

Usage

Running the server directly

# Using uv (in project environment)
uv run q-cli-mcp-server

# Using uvx (isolated environment)
uvx --from . q-cli-mcp-server

Using with Q CLI

Once added to Q CLI, you can use the tools:

q "Use the q_chat tool to ask: What is 1+1?"
q "Use the q_whoami tool to get user information"
q "Use the q_diagnostic tool to run diagnostics"

Implementation Details

The server uses FastMCP with proper subprocess handling for Q CLI commands:

• Non-interactive mode with trusted tools
• Proper stdin/stdout management
• Environment variable handling
• Robust error handling and logging
• Proper Python package structure with entry points

Example Usage

# Call q_chat tool
result = await session.call_tool("q_chat", {"prompt": "1+1"})

# Call q_whoami tool
result = await session.call_tool("q_whoami", {})

# Call q_diagnostic tool
result = await session.call_tool("q_diagnostic", {})

Files

• q_cli_mcp/server.py: Main MCP server implementation
• q_cli_mcp/__init__.py: Package initialization
• pyproject.toml: Project and dependency configuration with entry points
• requirements.txt: Python dependencies
• add_to_q.sh: Installation script

Configuration

The MCP server configuration is stored in .amazonq/mcp.json and uses uvx for dependency isolation with proper package entry points.

q-cli-mcp-server