MCPDebugger
A lightweight MCP debugger designed for learning and experimentation. Supports Windows executables (x86 and x64).
claude mcp add --transport stdio fatmike-gh-mcpdebugger docker run -i fatmike/mcpdebugger \ --env MCP_DEBUGGER_TARGET="path to target executable (optional, to auto-attach)" \ --env MCP_SERVER_LOG_LEVEL="info (optional: set to debug to increase verbosity)"
How to use
MCPDebugger is a C# MCP server that exposes a lightweight C++ debugger backend via MCP tools. The server allows external clients to start and stop debugging sessions, control execution (continue, step into/over, continue to breakpoints), inspect memory and disassemble code, and query module information. Clients can leverage the provided MCP tools to drive debugging sessions, inspect disassembly with symbol information, read and write process memory, and resolve strings from the debuggee. The integration is designed for experimentation and static/dynamic analysis, using Zydis for instruction decoding to provide clear disassembly feedback during debugging sessions. To use, connect as an MCP client, start a Debugger_Start session with a target executable, and then invoke the lifecycle and execution-control tools to drive the debuggee through breakpoints and API interactions. The toolset includes memory and module inspection capabilities, which are essential for reverse engineering or in-depth static analysis workflows.
How to install
Prerequisites:
- Docker (for containerized deployment) or a Windows build environment with .NET SDK if you opt to run the MCP server directly.
- Access to the MCP client tooling (e.g., Gemini CLI) if you want to reproduce the quickstart demo.
Option A: Docker (recommended for quick startup)
- Install Docker on your system.
- Pull the MCPDebugger image: docker pull fatmike/mcpdebugger
- Run the MCP server container: docker run -d --name mcpdebugger -i fatmike/mcpdebugger
- Connect an MCP client to the running server using the appropriate endpoint (depending on your Docker networking configuration).
Option B: Native Windows deployment (requires .NET development environment)
- Open the MCPDebugger project in Visual Studio.
- Build the MCP.Server solution (MCPDebugger.sln).
- Run the MCP.Server.exe from the build output (Release x64 or x86).
- Ensure the server is listening for MCP client connections (configure endpoints as needed in your environment).
Prerequisites summary:
- Docker or .NET development environment
- MCP client tooling for testing (optional but recommended)
- Target executable for debugging if you plan to run a full session
Additional notes
Tips and notes:
- The MCP server exposes a set of tools under the Debugger, Execution, Code & Memory, Module, and Utilities categories. Use these to control the debuggee, inspect memory, and analyze modules.
- When running in containerized mode, ensure proper access to the host's target executable and any required symbol files.
- If the debuggee requires elevated privileges or special authentication, configure environment variables accordingly (for example, enabling verbose logging or providing target paths).
- Disabling heavy-state-transfer tools (as suggested in the README) can help manage token usage during long debugging sessions.
- For best results, provide realistic target executables and ensure symbol information is available to enrich disassembly output.
Related MCP Servers
mcp-language
mcp-language-server gives MCP enabled clients access semantic tools like get definition, references, rename, and diagnostics.
web-eval-agent
An MCP server that autonomously evaluates web applications.
mcp
🤖 Taskade MCP · Official MCP server and OpenAPI to MCP codegen. Build AI agent tools from any OpenAPI API and connect to Claude, Cursor, and more.
VibeShift
[MCP Server] The Security Agent for AI assisted coding
mcp-dap
MCP server to communicate with DAP servers allowing AI Agents the ability to debug live programs.
mcp-chain-of-draft
Chain of Draft Server is a powerful AI-driven tool that helps developers make better decisions through systematic, iterative refinement of thoughts and designs. It integrates seamlessly with popular AI agents and provides a structured approach to reasoning, API design, architecture decisions, code reviews, and implementation planning.