Rea (Reverse Engineer Anything) is an open-source MCP toolkit with 350,000+ stars on GitHub that lets AI agents reverse engineer software — from high-level app behavior down to native binaries.
See a feature in an app and want to understand exactly how it works, without source code? Rea connects your agent to professional analysis tools through a single MCP server: the agent inspects the target, traces the relevant code and returns findings together with the evidence and limitations behind each conclusion, so you can explain the behavior or rebuild the feature in your own project. Analysis runs locally, and static JavaScript and .NET inspection never executes the target. Native analysis builds on your existing Hopper, Ghidra or IDA installation, and the same workflows are available from a terminal CLI. Documented showcases include reconstructing a sound-pan calculation in DX-Ball, tracing Notion’s Electron clipboard bridge and recovering a bullet-ring calculation from a DOS game.
Rea is free and open source under the MIT license, distributed through the rea-agents npm package. Native analysis uses your own existing Hopper, Ghidra or IDA setup — Ghidra is itself free — so there is no Rea subscription or usage fee.
Security researchers, reverse engineers and developers who need to understand how closed software behaves — for interoperability, security auditing, malware analysis or reimplementing a feature they admire.