This GitHub Repo Lets Claude Reverse Engineer Apps

REA GitHub repo lets Claude Code reverse engineer apps by reading shipped code, see how it traced Notion's copy feature and rebuilt DX-Ball. REA (Reverse Engineer Anything) is a free MCP add-on that gives Claude Code and other coding agents tools to inspect the files, bundles, and native instructions applications actually ship on your machine. Instead of guessing from screenshots, the agent reads exact locations in packaged code. The video traces how REA followed Notion's desktop app copy call through Electron's main process using only the shipped code, then shows what happens when decompiled native code gives wrong answers, using DX-Ball, a 1996 Windows brick-breaker, as a case study. REA pointed the agent at the sound-panning function's raw machine instructions, which contradicted the decompiler's pseudocode and revealed a missing input. Follow-up questions let the agent reconstruct the exact formula: brick position maps to a pan value from minus five hundred (left) to five hundred (right). The DX-Ball project proved the reconstruction by testing all 3,205 cases and byte-matching the compiled output against the original sixty-three-byte function. The project then rebuilt all 270 functions and reached a playable third-board game run. REA works best on JavaScript desktop apps like Notion (no extra tools needed), requires a decompiler like Ghidra for compiled programs, and cannot inspect server-side code. Every traced path is an inference, watching the program run confirms it actually happened. This video is for builders exploring AI agents, reverse engineering, and how agents verify claims against real code. Chapters: 0:00 Intro 1:09 Notion's Copy 2:24 Native Code 4:15 The Checks 5:43 Where It Stops 7:49 Verdict Tools & resources mentioned: - REA: https://github.com/morluto/rea - Ghidra: https://ghidra-sre.org - Claude Code: https://claude.ai - Hopper: https://www.hopperapp.com - IDA: https://hex-rays.com/ida-home About The Stack The Stack helps you build with AI. Each video takes one tool, model, or workflow and shows how it works in a few focused minutes, with the real benchmarks and real costs. We go deep on Claude Code and Cursor for AI coding, AI agents and MCP servers, the open-source AI tools and GitHub repos most people miss, RAG and vector search, fine-tuning, and running local LLMs on your own machine with Ollama and LM Studio. We compare models like ChatGPT and Claude, test AI automation with Zapier, Make, and n8n, and flag the tools that actually ship. Subscribe for new breakdowns: https://www.youtube.com/@the-stack-ai?sub_confirmation=1 #reverse engineering #claude code #ai automation