Summary

This project provides an AI-driven code auditing skill designed for AI coding agents like Codex, Copilot, and Gemini. It addresses the need for systematic code review, especially in complex or potentially problematic codebases, by enabling AI agents to generate evidence-based audit reports. The skill helps developers quickly identify, prioritize, and understand risks within a project, acting as a tireless, objective peer reviewer to surface potential issues before they become critical.

How It Works

The skill integrates directly into AI coding agents, allowing users to provide a repository link or install the skill locally. Upon activation, the AI performs an initial project profile, then conducts audits across numerous dimensions (e.g., security, stability, performance, maintainability) based on user-defined modes or a comprehensive full scan. It generates detailed reports including severity ratings, confidence levels, supporting evidence, impact assessments, and actionable remediation suggestions, complete with suggested testing strategies and estimated effort. Reports can be output in Markdown or HTML, featuring scoring panels and coverage matrices for clarity.

Quick Start & Requirements

• Installation: Clone the repository (git clone https://github.com/XiNian-dada/Fuck_My_Shit_Mountain.git) and copy the fuck-my-shit-mountain/ directory into the AI agent's designated skills path (e.g., ~/.codex/skills/ for Codex). Restart the AI agent.
• Prerequisites: Requires an AI coding agent that supports external skills (e.g., Codex, Claude Code, GitHub Copilot, Gemini CLI).
• Demo: A static demo showcasing report structure is available at: https://xinian-dada.github.io/Fuck_My_Shit_Mountain/

Highlighted Details

• Supports over 25 audit dimensions, selectable via natural language prompts (e.g., "security," "performance," "AI safety").
• Generates comprehensive reports with a scoring panel (0.0-10.0), coverage matrix, prioritized high-risk items, and detailed findings.
• Offers flexible output formats: Markdown or HTML, with the latter including interactive elements like sidebars and score bars.

Maintenance & Community

• The repository includes a "Star History" link. No specific community channels (like Discord or Slack) or detailed maintenance information are provided in the README.

Licensing & Compatibility

• License: MIT License.
• Compatibility: The MIT license is permissive, generally allowing for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

AI auditing is explicitly stated as a supplement, not a replacement for manual code review, testing, or real-world operational data. The tool aims to surface potential risks proactively but does not guarantee the absence of all issues.