Python Debugger MCP

A Model Context Protocol (MCP) server that provides Python debugging capabilities for AI assistants like Claude. This tool allows AI assistants to debug Python code by providing interactive debugging session management, breakpoint control, variable inspection, and step-by-step execution.

Features

• Debug Session Management: Start, restart, and end debugging sessions
• Breakpoint Control: Set, clear, and list breakpoints
• Variable Inspection: Examine variable values during debugging
• Step Execution: Step through code line by line
• Environment Detection: Automatically detect Python environments and project structure
• Cross-Platform: Works on Windows, macOS, and Linux

Installation

pip install python-debugger-mcp

Usage

As an MCP Server

Run the server directly:

python-debugger-mcp

In Claude Desktop

Add this to your Claude Desktop configuration:

{
  "mcpServers": {
    "python-debugger": {
      "command": "python-debugger-mcp"
    }
  }
}

Available Tools

The MCP server provides these debugging tools:

• start_debug(script_path) - Start debugging a Python script
• send_pdb_command(command) - Send commands to the debugger
• set_breakpoint(file_path, line_number) - Set a breakpoint
• clear_breakpoint(file_path, line_number) - Clear a breakpoint
• list_breakpoints() - List all breakpoints
• examine_variable(variable_name) - Examine a variable's value
• restart_debug() - Restart the current debugging session
• end_debug() - End the debugging session
• get_debug_status() - Get current debugging status

Requirements

• Python 3.10 or higher
• MCP library support

Development

Installing for Development

git clone <repository-url>
cd python-debugger-mcp
uv sync

Running Tests

python run_tests.py --test-type all

License

MIT License - see LICENSE file for details.

Inspiration

This project was inspired by debug-gym, which provides debugging environments for AI agents.