rea.run

Community guide

Give your coding agent eyes for software you may inspect.

rea.run is an unofficial bilingual home for REA (Reverse Engineer Anything) — MCP + CLI workflows, install notes, and original teaching examples. Source of truth stays on GitHub.

Agent prompt

Paste into your coding agent

› Install REA for this agent with npx rea-agents@latest setup. Show the setup plan for my approval, then verify the MCP connection. Use only targets I am authorized to analyze.

CLI

Or run setup yourself

$ npx rea-agents@latest setup

What is reverse engineering?

Understand a rule by reading the program — not the brochure.

Original puzzle: a local job runner retries failed tasks. Why do waits look like 1s → 2s → 4s → stop?

Backoff lab

Micro-demo of the observed wait pattern — not a live REA session.

1s2s4sstop
01
Find the scheduler

Hunt config keys and log lines by hand.

02
Decode the delay

Linear, capped, or exponential?

03
Name the stop rule

Max attempts? Max delay? Both?

Examples

Two teaching blocks — original scenarios

open-source-cli-flag

Trace a CLI flag in an open-source tool

Ask your agent to locate where a documented flag is parsed — on a repo you already cloned.

Call-site reveal

flag --verbose
parse parseArgs()
use Logger.setLevel()
Open example →

your-electron-export-path

Find an export path in your Electron app

Goal
List entry modules and the function that writes the export file.
Why reverse engineer
Support asked “where does CSV leave the app?” and you want the agent to gather the graph first.
Sample prompt
Use REA to inspect /absolute/path/to/our-electron-app (or the .asar we ship). Show export-related entry points, calls, and source locations. Mark anything unresolved.
Open example →

Capabilities

Chunked by target surface

Compliance