dbgmcp

by ramenhost

Development/Debugging GDB LLDB PDB Debugging AI Agents

dbgmcp is a collection of MCP servers designed to connect various debuggers (GDB, LLDB, PDB) to LLM agents using the Model Context Protocol. It allows AI agents to interact with debuggers for enhanced debugging and analysis capabilities.

View on GitHub

Last updated: N/A

What is dbgmcp?

dbgmcp provides separate MCP servers for GDB, LLDB, and PDB, enabling LLM agents to interact with these debuggers. It allows loading programs, executing commands, and managing multiple connections.

How to use dbgmcp?

Pre-built binaries can be downloaded from the releases page. Alternatively, build from source using Rust and Cargo. Configure Claude Desktop or VS Code Github Copilot by adding the required MCP server configurations to their respective configuration files (claude_desktop_config.json or .vscode/mcp.json).

Key features of dbgmcp

Separate servers for each debugger
Load programs into the debugger
Execute arbitrary commands in the debugger
Supports multiple simultaneous connections

Use cases of dbgmcp

Debugging code with AI assistance
Analyzing program behavior through LLM interaction
Automated debugging workflows
Integrating debuggers into LLM-powered development environments

FAQ from dbgmcp

What debuggers are supported?

GDB (GNU Debugger), LLDB (LLVM Debugger), and PDB (Python Debugger) are supported.

How do I install dbgmcp?

You can download pre-built binaries from the releases page or build from source using Rust and Cargo.

How do I configure dbgmcp for Claude Desktop?

Add the required MCP server configurations to the claude_desktop_config.json file.

How do I configure dbgmcp for VS Code Github Copilot?

Enable MCP support in VS Code settings and add the required MCP server configurations to the .vscode/mcp.json file.

Is it safe to use dbgmcp?

AI agents can execute arbitrary commands inside debuggers, including shell commands. Use at your own risk.