Loading connector details…
Loading connector details…
Choose a unique username to continue using AgentHotspot
by samefarrar • Testing & QA
MCP-PDB enables integration of Python's pdb with Claude and LLMs for enhanced debugging capabilities.
Integrate Python debugging with language models for improved code analysis
Automatically manage breakpoints and debugging sessions
Conduct project-aware debugging with support for virtual environments
MCP-PDB allows users to interface with Python's debugger through the Model Context Protocol, optimizing the debugging process for coding agents like Claude and LLMs. Inspired by Microsoft's debug-gym, it offers tools for breakpoint management, variable inspection, and session control, benefiting environments that support the UV package manager. This tool is geared towards providing detailed insights during debugging sessions, especially for dynamic Python environments.
Start a debugging session on a Python file within its project context. Args: file_path: Path to the Python file or test module to debug. use_pytest: If True, run using pytest with --pdb. args: Additional arguments to pass to the Python script or pytest (space-separated).
Send a command to the running PDB instance. Examples: n (next line), c (continue), s (step into), r (return from function) p variable (print variable), pp variable (pretty print) b line_num (set breakpoint in current file), b file:line_num cl num (clear breakpoint number), cl file:line_num l (list source code), ll (list longer source code) a (print arguments of current function) q (quit) Args: command: The PDB command string.
Set a breakpoint at a specific line in a file. Uses relative path if possible. Args: file_path: Path to the file (can be relative to project root or absolute). line_number: Line number for the breakpoint.
Clear a breakpoint at a specific line in a file. Uses relative path if possible. Args: file_path: Path to the file where the breakpoint exists. line_number: Line number of the breakpoint to clear.
List breakpoints known by PDB and compare with internally tracked breakpoints.
Scores are informational only and provided “as is” without warranty. AgentHotspot assumes no liability for actions taken based on these ratings.