heavy3-ai

heavy3-ai / code-audit

Public

Heavy3 Code Audit: Agent skill that uses multi-model consensus to review plans, code, and PRs for coding agents

33
10
100% credibility
Found Feb 02, 2026 at 19 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

An open-source Claude AI skill that uses multiple AI models via OpenRouter to perform detailed code reviews, plan assessments, and pull request audits.

How It Works

1
🔍 Discover the code review helper

You stumble upon Heavy3 Code Audit, a free tool that lets AI experts check your code for problems.

2
📥 Bring it home

You download it and place it in your Claude chat helper folder with a few simple steps.

3
🔑 Link the smart thinkers

You grab a free passcode from a website and add it so the AIs can share their wisdom.

4
💬 Chat with Claude

In your Claude conversation, you type a simple command like /h3 to start a review.

5
🤖 Pick your review team

Choose a quick free check or gather a council of three expert AIs for deeper insights on bugs, speed, and safety.

6
📝 Share your code

Paste your recent changes, a plan, or a project update, and the experts dive in right away.

✅ Receive expert advice

You get a clear report with fixes and tips, making your code stronger and ready to shine.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 19 to 33 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 code-audit?

Heavy3 Code Audit is a Python tool for ai code audit github workflows, integrating as a skill into coding agents like Claude Code, Cursor, or Gemini CLI. It automates source code audit github tasks by reviewing plans, diffs, uncommitted changes, or PRs via OpenRouter APIs—using single models for quick checks or multi-model consensus for deeper analysis on correctness, performance, and security. Developers get structured feedback on bugs, vulnerabilities, and optimizations without manual code auditing certification or expensive code audit services.

Why is it gaining traction?

Its killer hook is dirt-cheap code audit cost: free tier with rotating models, single reviews at ~$0.01, or council mode (3 specialized LLMs) at ~$0.10, all BYOK. Commands like `/h3 pr 123` or `/h3 --council` deliver consensus-driven audits faster than waiting for humans, with web search for context-aware insights. Stands out from basic linters by understanding agent-generated code intent and providing actionable, role-specific critiques.

Who should use this?

AI coding agent users validating agent outputs before commit, like solo full-stack devs auditing code from Claude or Cursor. Open source maintainers triaging PRs quickly, or indie hackers doing pre-deploy code audits on tight budgets. Ideal for backend teams spotting N+1 queries or security holes in prototypes.

Verdict

Grab it for low-stakes code audition 2025 experiments—the 1.0% credibility score and 28 stars signal early maturity with solid docs but no tests, so treat as a smart second opinion, not a replacement for human code auditors. MIT-licensed and dead simple to try. (187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.