game-reverse-engineering
Skill44 starsChanged 22 days agoMITView on GitHub
Decompile and analyze game binaries you have rights to.
In this repository9 other files it gives its agents.
- 3d-asset-pipelineSkill
- blender-pipelineSkill
- game-audioSkill
- game-studioSkill
- godot-devSkill
- pixel-art-pipelineSkill
- playtestingSkill
- ship-itSkill
- unity-devSkill
The file
Files
Game Reverse Engineering Skill
Decompilation workflows for modding, preservation, interoperability and decomp projects: Ghidra headless export of decompiled C, managed-code (Unity/.NET) decompilation, and engine pack inspection. Only for binaries the user owns or has the right to analyze; never for cheats in online games or DRM removal.
When to Use
Matching-decomp projects, modding a single-player game, recovering your own lost source, or understanding a file format for a fan tool.
Prerequisites
- Native code: Ghidra 11+ (
analyzeHeadlessin<ghidra>/support/) and a JDK. - Unity Mono games:
ilspycmd(dotnet tool install -g ilspycmd). IL2CPP games: Il2CppDumper first. - Godot games:
gdsdecomp(GDRE Tools) for .pck recovery.
How to Run
- Identify the target with
terminal:file <binary>; Unity if*_Data/Managed/Assembly-CSharp.dllexists, IL2CPP ifGameAssembly.dll/libil2cpp.so, Godot if a.pcksits beside the executable. - Native:
python scripts/ghidra_export.py --ghidra <ghidra dir> --binary <file> --out decomp/runs a headless import + analysis and writes one.cfile per function plusfunctions.json(name, address, size, callers). Start frommain/entry points and strings of interest viasearch_files. - Unity Mono:
ilspycmd -p -o decomp/ <Managed>/Assembly-CSharp.dll→ a buildable C# project. - Godot:
gdre_tools --headless --recover=<game.pck> --output-dir=decomp/. - Record findings in
decomp/NOTES.md: symbol names you assign, structs, file formats.
Pitfalls
- Respect licenses and EULAs; keep extracted assets out of public repos.
- Matching decomps need the original compiler; Ghidra output is a reading aid, not compilable source.
- Big binaries take Ghidra many minutes; run with a generous
terminaltimeout or in the background.
Verification
functions.json lists the expected entry points; spot-check a few .c files against the disassembly.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

