← all entries
A-TIER tool

REA MCP — agent-driven reverse engineering with a Ghidra backend

backsla.sh · 2026-10-19T16:06:56.000Z · in AI Saves, Vol. 8
Open post ↗

100k ft — What is this?

REA MCP — agent-driven reverse engineering with a Ghidra backend
toolA-tier

50k ft — Why care?

REA is an MCP server that lets an AI agent reverse engineer apps and stripped native binaries, with Ghidra doing the decompilation while the agent drives and cites evidence per call. The creator decompiled a closed-source program in three to four minutes on Ubuntu. Setup is deferred to a follow-up video.
agentsopen sourcesecurityvideo production

10k ft — Key takeaways

• Evidence-per-call output makes agent RE findings auditable • Pair an agent with Ghidra over MCP to recover logic from stripped binaries • Use only on software you are authorized to analyze

5k ft — Tools & techniques

Ghidra, REA, Ubuntu AI agents, MCP, MCP server, decompilation

1k ft — Deep-dive / try it

REA - an MCP that lets an AI agent reverse engineer anything, from app behavior down to native binaries. Point it at a stripped binary with no source and it recovers the logic, then shows you the evidence behind every call. Ghidra runs the backend while your agent drives. Work that used to take a red team days now takes minutes. For educational purposes only.
Transcript excerpt
Now that all the impatient people are gone. Here's how you can reverse engineer anything I just built an AI powered reverse engineering lab that decompiled my closed source program in about three to four minutes You only need a Ubuntu desktop and this github repo right here Dude, that is a flippin time saver. This is a very dangerous tool and an ethical hackers arsenal Which is why I'll be showing you my exact setup in my next video like and follow if you want to become a better ethical hacker

Appears in

Collection: AI Saves, Vol. 8