ghidra-mcp
Ghidra MCP server with 200+ tools that let AI assistants drive real reverse-engineering work
Native MCP plugin for x64dbg — let an AI drive the debugger: breakpoints, stepping, memory, registers
x64dbg-mcp-server is an open-source ai agents project by duty1g (duty1g/x64dbg-mcp-server on GitHub). It has 2,402 GitHub stars and 256 forks, is released under the MIT license, and was last updated on GitHub on 30 September 2026.
Have a binary you are authorized to analyze and a debugging goal — trace behavior, find a check, or understand a packed executable.
Download the latest release (or build from source with zig build) and copy the contents of dist/ into your x64dbg root folder.
Launch x64dbg — the MCP server auto-starts; open Plugins → x64dbg-MCP Server → Configure to set bind address, port and token if needed.
Add the Streamable HTTP endpoint (http://localhost:9094/) with the bearer token to your MCP client config, then ask the assistant to load your binary and break at the entry point.
The README documents a large toolset (80+ claimed in features; 72 listed in the tools table) covering disassembly, stepping, breakpoints (software/hardware/conditional/memory/exception), memory, registers, modules, threads, call stack, pattern scanning, strings, xrefs, symbols, PE analysis, OEP detection and dumping.
Full debugger event coverage — init, stop, breakpoint, exception, step, attach/detach, DLL load/unload, threads and more — with WaitForEvent long-polling.
No .NET, no Python, no runtime: drop the .dp32/.dp64 plugin into the plugins folder; one codebase builds both architectures and cross-compiles from any host.
A bearer token is auto-generated on first run and required on every request; IP/port/token are changeable from a config dialog that restarts the server on save.
The MCP server starts automatically when x64dbg launches.
LoadBinary, AttachProcess and ExecuteDebuggerCommand give the AI the same levers a human analyst has at the keyboard.
Clone x64dbg-mcp-server if building from source; release users download the prebuilt dist/ instead.
git clone https://github.com/duty1g/x64dbg-mcp-server.git
cd x64dbg-mcp-serverBuild both architectures into dist/ with Zig, then copy the contents of dist/ into your x64dbg root folder and launch x64dbg.
zig build -Doptimize=ReleaseSafe --prefix distConnect your client over Streamable HTTP with the auto-generated bearer token.
{
"mcpServers": {
"x64dbg": {
"type": "http",
"url": "http://localhost:9094/",
"headers": { "Authorization": "Bearer <redacted>" }
}
}
}duty1g maintains it; the project site is duty1g.online.
Yes, MIT-licensed. x64dbg itself is also free and open source.
No — the plugin is local. Your MCP client's model (Claude etc.) carries its own plan or key.
JADX-AI-MCP gives an LLM static, decompiled Android code. This plugin gives an LLM live control of a Windows debugger — execution, memory and registers — for native binaries.
Copy dist/ into your x64dbg folder, launch x64dbg, copy the generated token from the config dialog, and add the http://localhost:9094/ endpoint to your MCP client. Only analyze software you are authorized to debug, and don't expose the server to untrusted networks.
Ghidra MCP server with 200+ tools that let AI assistants drive real reverse-engineering work
Open-source Computer Use as an MCP service: let any agent drive macOS, Linux, and Windows desktops.
Control Gmail, Calendar, Docs, Sheets, Drive and more from AI — a full Google Workspace MCP server.
JADX plugin + MCP server that lets LLMs analyze Android APKs live — decompile, review, find vulnerabilities
Official MathWorks MCP server: drive MATLAB from AI coding agents.
High-performance code intelligence MCP server: your codebase as a persistent knowledge graph
The most-starred other AI Agents projects in the LibreAI catalogue — the realistic shortlist if x64dbg-mcp-server is not the right fit.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Free and open — grab the source, star the repo, and start creating.