Quack MCP Server

🚀 Quack Guide

Quack is a Python code analysis server built on the Model Context Protocol (MCP). It supports multiple code analysis tools and integrates with Visual Studio Code through C Extension, providing real-time feedback for Python development.

🚀 Quick Start

Run the Server

python3 quack.py --stdio

By default, the server listens on port 8000 of the local IP address. To specify a different host or port, use the following command:

python3 quack.py --stdio --host <host> --port <port>

Usage Example Code

The following is example code containing various issues, used to test the server's functionality:

# Example of incorrect usage:
unused_var = 42  # Unused variable
x = 10  # Single-letter variable name
print("Hello, World!")  # Missing newline character

# Example of type error:
def add(a: int, b: int) -> int:
    return a + b

result = add("5", 10)  # Type error: Adding a string and an integer

Paste this code into the C Extension, and the Quack server will automatically detect and report the issues.

✨ Features

• Supports multiple code analysis tools, providing comprehensive code analysis capabilities.
• Integrates with Visual Studio Code through C Extension, offering real-time feedback for Python development.

📦 Installation

Install Dependencies

Before running Quack, make sure the following tools are installed:

• Python 3.8 or higher
• Docker (only for SSE mode)
• pip

Then install the required Python packages:

pip install quack

Run the Quack Server

The Quack server has two running modes: Standard Input/Output mode and Streaming Exchange Extension mode.

Standard Input/Output Mode

Run the Python script directly:

python3 -m quack.server --stdio

By default, the server listens on port 8000 of the local IP address. To specify a different host or port, use the following command:

python3 -m quack.server --stdio --host <host> --port <port>

Streaming Exchange Extension Mode

Run through Docker:

docker run -p 8000:8000 quack/mcp-server

📚 Documentation

Configure the Quack Server

Create a configuration file quack_server_config.json and add the following content:

{
    "name": "Python Linter",
    "command": "python3 -m quack.server --stdio",
    "args": [],
    "stdinHandler": {
        "type": "stdio"
    },
    "schemes": ["file"],
    "language": "python",
    "dllLoadMessage": ""
}

Add this configuration file to Visual Studio Code's settings.json:

{
    "extensions.csharpEnableCodeLens": false,
    "extensions.csharpShowSnippetsOnHover": false,
    "extensions.pythonVSQuackServer": "/path/to/quack_server_config.json"
}

💻 Usage Examples

Standard Input/Output Mode

Run in the terminal:

python3 -m quack.server --stdio

Then use the following commands to interact with the server:

# Submit a new job
curl -X POST http://localhost:8000/queue \
     -H "Content-Type: application/json" \
     -d '{"tool":"pylint","content":"print(\"Hello, World!\")"}'

# Check the job status
curl http://localhost:8000/queue/<JOB_ID>

# Get the analysis results
curl http://localhost:8000/results/<RESULT_ID>

Streaming Exchange Extension Mode

Access through the following command:

http://localhost:8000/sse

Tool Integration

Quack supports multiple code analysis tools, including:

• Pylint: Used to detect code style and quality issues.
• Mypy: Used for static type checking.

More Information

For more information, please visit Official Documentation.