Dryxio

Dryxio / auto-re-agent

Public

Autonomous reverse engineering agent: dual-LLM loop, 11-signal parity engine, Ghidra backend

60
5
100% credibility
Found Feb 18, 2026 at 26 stars 2x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

re-agent automates reverse engineering binaries into C++ source code by coordinating AI agents with decompilation tools and quality checks.

How It Works

1
🕵️ Discover the Helper

You find a smart tool online that turns confusing machine code from games or programs into neat, readable instructions.

2
💻 Set It Up

You easily add the tool to your computer and prepare a simple note telling it where your project files are.

3
🔗 Connect Smart Parts

You link a thinking AI helper and your code-reading tool so everything can understand the binary secrets together.

4
Choose Your Target
🔍
Single Routine

Focus on one exact piece of code to convert right away.

👥
Group of Routines

Handle several related pieces at once for bigger progress.

5
✨ Let It Work Its Magic

Press go, and it reads the code, writes human-like instructions, checks for mistakes, and fixes them in a few tries.

6
📊 Track Your Wins

Peek at the progress list to see what's done, what's perfect, and what's next.

🎉 Enjoy Readable Code

You now have clean, verified instructions ready to drop into your project and make it come alive.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 26 to 60 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is auto-re-agent?

auto-re-agent is a Python tool that automates reverse engineering binaries into C++ source code using a dual-LLM agent loop— one to generate code from Ghidra decompilations, another to verify and fix it. It solves the drudgery of manual binary analysis by prioritizing functions via caller count, running bounded fix rounds, and applying an 11-signal parity engine for quality gates like call mismatches or stub detection. Users get CLI commands like `re-agent reverse --class CTrain` to target classes or addresses, plus `parity` checks and progress tracking.

Why is it gaining traction?

It stands out with its verify-then-fix workflow that iterates up to 4 rounds per function until LLM checks pass, plus configurable heuristics catching issues like tiny sources for large ASM or missing NaN logic. Safety features—no auto-commits, detailed logs, session isolation—make it reliable for real projects, unlike raw LLM prompting. As a github autonomous coding agent, it scales to entire classes while degrading gracefully if Ghidra features lack.

Who should use this?

Reverse engineers modding games like GTA SA, where hook patterns map addresses to stubs. Binary analysts tackling autonomous systems github repos or autonomous reverse engineering tasks. Teams with Ghidra projects needing quick C++ stubs for plugin::Call wrappers.

Verdict

Worth trying for niche binary reversal if you have Ghidra and LLM keys—solid CLI/docs for alpha stage, but 18 stars and 1.0% credibility signal early days; test on stubs first before production.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.