TECHNOLOGY
REA, an open-source tool, lets coding agents reverse engineer apps, binaries and websites
REA is an open-source project that gives AI coding agents a set of tools to inspect programs and explain what they do. It works through MCP, a protocol for connecting agents to local tools, and covers targets from native binaries to Android APKs.
Image: REA
REA, short for Reverse Engineer Anything, is an open-source project that gives a coding agent the tools to inspect a program and explain what it does.
Reverse engineering means working out how software behaves by examining the program itself. REA connects to an agent through MCP, a standard way for agents to call local tools. The agent asks about a target, REA inspects and traces it, and the results come back with their evidence.
The agent can then use those findings to ask follow-up questions, explain the behavior, or write and test an implementation. REA also has command-line tools that use the same workflows.
The project lists many kinds of targets. These include native binaries, offline ELF layouts, EVM bytecode (code for Ethereum smart contracts), recorded Linux crashes, JavaScript and Electron apps, websites, saved network captures, .NET assemblies, Android APKs, firmware, packages and process behavior. For native binaries, REA returns pseudocode, assembly, strings, symbols, calls and references. For JavaScript and Electron apps, it recovers modules and how they relate to each other.
Requirements depend on the target. REA needs Node.js 22.x (22.19 or later), 24.x (24.11 or later) or 26 and above, plus npm. Deep native analysis uses Hopper, Ghidra or IDA. Static JavaScript and .NET inspection work without any of them. Websites need a Chrome-family browser, and Android analysis needs headless JADX and a full JDK.
The setup is meant to be run from the agent itself. The project's suggested prompt tells the agent to run npx rea-agents@latest setup, show a setup plan for approval, and then verify the installation. Setup can install Hopper after approval, while Ghidra and IDA use existing installations. Any agent that supports local MCP servers can use REA, and other clients can register it manually.
The project makes a distinction between static and runtime work. Static JavaScript and .NET inspection read the supplied files without running the application. Runtime capture runs or interacts with the selected target using the user's own permissions, and each runtime guide describes its effects.
On privacy, the project says REA analyzes targets locally. The agent still receives the tool results, and the agent's model provider has its own data policy.
The project says REA is meant for lawful research, analysis and reconstruction. Users are responsible for getting any required authorization and following applicable laws. It does not endorse illegal or unauthorized use.