morluto/rea

What it solves

REA (Reverse Engineer Anything) allows AI agents to inspect and understand software applications without needing access to their original source code. It bridges the gap between high-level AI reasoning and low-level binary analysis, enabling users to ask an agent to explain how a specific feature works, provide evidence from the code, or even help reconstruct that feature for a new project.

How it works

REA functions as an MCP (Model Context Protocol) server that connects your preferred AI agent to a suite of reverse-engineering tools. When you ask a question about an application, the agent uses REA to perform static or dynamic analysis on the target. Depending on the file type, REA interfaces with specialized tools like Ghidra, Hopper, IDA, or native JavaScript/Electron analysis engines to extract pseudocode, assembly, module relationships, and runtime behavior. The agent then processes these findings to answer your questions or generate code.

Who it’s for

This tool is designed for developers, security researchers, and software engineers who want to leverage AI to accelerate the process of understanding legacy code, proprietary applications, or complex binary structures.

Highlights

• Supports a wide range of targets including native binaries (x86/ARM), JavaScript/Electron apps, .NET assemblies, Android APKs, and firmware.
• Integrates directly with popular AI agents via the Model Context Protocol (MCP).
• Provides evidence-based results, returning actual pseudocode, assembly, and structural metadata rather than just guesses.
• Runs analysis locally, ensuring that sensitive application data remains on your machine.
• Offers both an agent-integrated workflow and a standalone CLI for manual inspection.