Case studies & GitHub projects

How a GPL authenticator turns time blocks + account keys into six-digit codes — an authorized Android packaging lab, not an account bypass.

Using REA inside a contest-style reverse challenge — Evidence first, rules always win.

A Breakout-style game lab: inspect behavior, then compare against a public reconstruction repo.

Same healthy pattern as DX-Ball: inspect, explain, rebuild in the open.

A first-session map for Electron-like apps: find where a feature actually lives before you rewrite it.

Why 200 + 10% can yield 220 — and how to check explanations against Microsoft’s public sources.

A practitioner used REA browser inspection on their own portal: what can an anonymous visitor see?

Gold-standard write-up: Ubuntu netcat-openbsd 1.234-1 inspected with REA layout tools and Ghidra — artifacts in-repo, no guesswork.

C++20 reimplementation of DICE Refractor 2 (BF2-era): engine only, bring-your-own install, REA wired via tools/rea.sh.

Claude Code plugin pairing REA MCP with TypeSafe Jev hooks — route, gate, score evidence, verify “done.”

Unofficial FR pedagogical repo: analyze apps/binaries with coding agents via MCP/CLI, Hopper or Ghidra, Evidence-first, legal notes included.

Passive Chrome DevTools Protocol session: inspect a public pricing layout and rebuild a clean single-file page with an agent.

Curated bilingual list of AI-assisted reverse-engineering tools — MCP, skills, plugins, agents — with REA in the local-agent picks.

CC0 curated list of reverse-engineering tools and learning resources — binaries, apps, firmware, formats, protocols — featuring REA.

moonfruit/homebrew-tap ships Formula/rea-agents.rb — npm tarball, bottles, shell completions — for brew users who want a formula not only npx.

Curated catalog entry `reverse-engineer-apps-and-binaries-with-rea` — MCP-oriented skill metadata pointing builders at setup and Evidence workflows.

Template repo: `.mcp.json` pins rea-agents, ships the REA skill, and keeps per-project RESEARCH.md — demonstrated on the author’s own MAD Studio Electron package.

Community Discussion #1435: median timings for analyze-javascript-application across Node lines, plus Ghidra/Temurin, Chromium executable, and TMPDIR tips.

yzua/nixos-config includes packages/rea/ with pinned npm integrity, Ghidra ZIP digests, JADX, and offline loader tests — a packaging story for Nix labs.

MIT source reconstruction of 東方永夜抄 ～ Imperishable Night 1.00d — Evidence-first agent routes in the same N0zoM1z0 family as th04.

Evidence-first control plane for verified Touhou reconstruction projects — shared IDA/Ghidra tooling, atomic evidence tools, read-only game adapters.

Partial vitest/TypeScript port of TraceBrief prepared for morluto/rea conventions — parity harness against real REA fixtures, explicit “what is not done.”

Excerpt focus: TechDoc/Reverse/REA-Review.md inside vibe-investing — Evidence v2, tool families, honest pros/cons, Getting Started — not the trading bots.

EstebanForge/AGENTS vendors `skills/reverse-engineer-anything` with META.md — a practical mirror for multi-agent skill directories.

AppitStudio/awesome-jev indexes TypeSafe Jev projects and includes a dedicated rea-jev tool page — discovery path into REA for Claude Code users.

Hands-on playbook: minimal Electron target, Evidence extractor with line refs, MCP probe, verified Decompile→Understand→Recreate loop.

Thin MCP stdio bridge for hosts that cannot filter REA’s full tools/list — 21 tools ≈ 6.7 KB, with CLI escape hatch for full coverage.

Apache-2.0 Rust/Bevy clean-room of GunZ: The Duel — loads maps/assets from your Steam install; REA+Ghidra scripts for investigation.

Codex skill that pairs REA with macOS Computer Use to rebuild behavior into a new source project — authorized/own targets only on rea.run.

Experimental Win64 C++ protect/verify toolkit for scoring multi-tool agents including REA — research prototype; anti-REA not established.