A REA-prewired reverse-engineering workspace
Good labs start with wiring, not with a random binary. MADTreasures/reverse is a **workspace pattern**: MCP preconfigured, skill present, targets gitignored, Evidence notes separated from how-tos — shown on their own MAD Studio Electron build.
What it is. A multi-project reverse-engineering monorepo template (TypeScript). .mcp.json registers rea-agents MCP; .claude/skills/ vendors the REA skill; each project keeps README.md + RESEARCH.md with observation/conclusion/unknown split.
Relation to REA. Static JS/Electron analysis via CLI when MCP’s 60s limit is too tight; doctor/setup for native engines. Example Evidence from auditing their own shipped MAD Studio package.
Compliance note (important). The README mentions commercial DAW products only as UX inspiration for the author’s rebuild. On rea.run we do not turn that into a how-to against those products. Stick to own/authorized targets.
Evaluation. Excellent scaffolding for Claude Code teams. Early (0★, no license declared) — fork carefully and add your own license before sharing.
Compliance
Showcase the workspace pattern and the author’s own MAD Studio Electron audit only. Do not treat third-party commercial DAWs as reverse targets.
Takeaways
- Prewire MCP + skill before the first binary.
- Keep targets/ gitignored.
- Own Electron audits ≠ third-party commercial RE tutorials.
Related on rea.run
More on-site cases