Skills

REA: Reverse Engineer Native Binaries, Electron Apps, .NET Assemblies, and Websites with an Agent

Best suited for letting an agent reverse engineer native binaries, JavaScript/Electron apps, .NET assemblies, and websites via an MCP server. Before use, you must confirm the Node.js version requirements (22.x >=22.23.4, 24.x >=24.22.1, or 26.12.1+), and note that setup rewrites agent configuration while native analysis depends on an existing Hopper/Ghidra/IDA.

Official source · morluto/rea

Install

npx rea-agents setup

Who It's For

REA ("Reverse Engineer Anything") delivers reverse engineering capabilities to agents through an MCP server, covering native binaries, JavaScript/Electron apps, .NET assemblies, and website behavior. It suits users who need to analyze a binary, an Electron app, or a website directly in conversation.

Prerequisites: your machine needs Node.js 22.x (>=22.23.4), 24.x (>=24.22.1), or 26.12.1+, plus npm. For native analysis, you can use your existing Hopper, Ghidra, or IDA installation, and setup can also install Hopper on your behalf once approved; static JavaScript and .NET analysis do not require a native engine. On Linux, static JavaScript/Electron analysis requires a procps-compatible ps on your PATH.

REA works with any agent that supports local MCP servers; the official documentation says setup supports Claude Code, Codex, Cursor, Gemini CLI, Grok Build, and others. The evidence does not state which operating systems are supported, nor the specifics of which file formats are supported, so it's advisable to confirm in the official repository before use.

First Attempt in 3 Minutes

The officially recommended main path is to run setup directly with npx:

  1. Confirm that Node.js and npm meeting the version requirements are installed (on Linux, also confirm ps is on your PATH).
  2. Run in a terminal:
npx rea-agents setup
  1. Setup will ask you to choose which agent to configure, show the changes it will make, and ask for your approval; once confirmed, it adds REA's MCP server and accompanying workflow instructions, and backs up your existing configuration.
  2. Restart your agent, and REA becomes available.

Alternatively, you can install globally: npm install --global rea-agents, then run setup the same way; but this is an alternative method given by the official documentation, and the main path remains npx rea-agents setup.

For your first attempt, it's recommended to pick a local binary you own with low risk and give the agent a read-only analysis task, such as asking it to describe the binary's structure and exported symbols. The exact tool names and output format depend on the tools actually available inside your agent; the evidence does not list them item by item. Note: setup itself modifies the chosen agent's configuration file (the official documentation says it makes a backup), so read the changes it presents before approving; native analysis that relies on Ghidra/IDA requires you to have them installed, and approving the Hopper installation is an action that changes your system. Which processes read or run the binary during analysis is not stated in the evidence and remains unknown.

Use Cases

Case 1: Analyzing a local .NET assembly. Send the agent a task: "Use REA to analyze the .NET assembly MyApp.exe and list its main types and methods." This scenario uses static .NET analysis, which per the official documentation does not require native engines like Hopper/Ghidra/IDA and does not depend on the procps ps. The expected result is that the agent, using REA's MCP tools, provides a structural description of the assembly; the exact granularity of the output is not shown in the evidence.

Case 2: Analyzing an Electron app. Scenario: you want to understand the internal behavior of an Electron app. Send the agent: "Use REA to analyze this Electron app's JavaScript." This scenario uses static JavaScript analysis; on Linux, first confirm that a procps-compatible ps is on your PATH. The expected result is an analysis of the app's scripts. Similar tasks involving the native binary portions would use your existing Ghidra/IDA/Hopper or a Hopper installed with your approval.

Boundaries of Use

  • Configuration side effects: setup modifies the chosen agent's configuration file and adds workflow instructions; the official documentation says it backs up the existing configuration. Installing Hopper requires your approval.
  • Runtime environment: specific versions of Node.js and npm are required; static JavaScript/Electron analysis on Linux requires a procps-compatible ps.
  • Native analysis dependencies: uses your existing Hopper, Ghidra, or IDA; setup can install Hopper with your approval.
  • Client prerequisites: the agent must support local MCP servers; the agent needs to be restarted after configuration.
  • Unconfirmed information: whether the analysis reads/writes user files, executes commands, or accesses the network is not stated in the official evidence; the supported range of text and binary languages, and supported operating systems, are also not given in the evidence. If you have security concerns, verify these behaviors in the official repository before use.