XiNian-dada/Fuck_My_Shit_Mountain

Source: https://github.com/XiNian-dada/Fuck_My_Shit_Mountain

Fuck My Shit Mountain 🏔️💩

Disclaimer: AI review results are for entertainment and reference only, and do not represent the actual quality, security, compliance, or release readiness of the project. Any conclusions should be validated through manual review, test results, and real runtime environments.

Evidence-based AI code audit skill. Professional output. Zero emotional bullshit.

Fuck My Shit Mountain is an AI skill (prompt framework) for generating rigorous, professional code review reports. The name is crude, but every report is calm, structured, evidence-driven, and actionable.

Workflow

1. AI first confirms necessary inputs — only asks for missing mode, report language, output format after a lightweight project portrait, recommending options like “full audit / security-focused / frontend experience / release & ops” in your language; if you’ve already specified, it proceeds directly.
2. AI systematically audits your codebase — covers first-party source code, tests, configs, dependencies, and release files, honestly recording exclusions.
3. AI generates a structured report — scores, findings, principle compliance, fix priority, quick wins.
4. HTML output — sidebar navigation + scroll tracking, colored score bars, findings table per dimension + verified checklist, design principle compliance table, fix priority table.

Demo Preview

If you want to show someone a “real-looking” review page, check out this static demo:

• docs/index.html is the GitHub Pages entry page
• The top of the page clearly marks it as Demo / Fictional / Not a real audit
• If the repo’s Pages source is set to main /docs, it auto-deploys on push; .nojekyll is already placed in docs/

Modes

Regular users can directly choose natural language options like “Full Audit”, “Security & Privacy”, “Frontend Experience”, “Release & Ops”, and the AI maps them to internal modes. Advanced users can specify single or combined modes (e.g., security, stability, type-safety). Full mode covers 25 audit dimensions.

Project Structure

fuck-my-shit-mountain/
├── SKILL.md              — Entry point & rules
├── agents/               — UI metadata
├── prompts/              — 26 audit mode prompts
├── references/           — Common report format, HTML, coverage, lint, tool references
├── rubrics/              — Severity, confidence, evidence, principles, scoring
├── scripts/              — Project portrait, report lint, skill packaging tools
├── templates/            — audit-report.md, audit-report.html, issue-card.md, remediation-plan.md
└── examples/             — Rust, Node.js, Vue audit examples

It is recommended to generate a clean skill package before distribution or installation, or use the same exclusion rules to sync to the local skill directory, avoiding inclusion of non-essential files like README, .DS_Store, cache files, etc.

python3 fuck-my-shit-mountain/scripts/package_skill.py --dry-run
python3 fuck-my-shit-mountain/scripts/package_skill.py

The generated dist/fuck-my-shit-mountain.zip excludes README, .DS_Store, cache files, and other non-essential items.

To update an installed skill in Codex locally, run from the repo root:

rsync -av --delete --delete-excluded \
  --exclude='README.md' \
  --exclude='.DS_Store' \
  --exclude='__pycache__/' \
  --exclude='*.pyc' \
  --exclude='*.pyo' \
  --exclude='dist/' \
  fuck-my-shit-mountain/ ~/.codex/skills/fuck-my-shit-mountain/

Example Score Panel

Security        ████████░░  8.0  A
Stability       ██████░░░░  6.0  B
Performance     ██████████  10.0 S
Testing         ████░░░░░░  4.0  C
Maintainability ███████░░░  7.0  A
Design          █████░░░░░  5.0  B
Release         ██████░░░░  6.0  B
─────────────────────────────────────
Overall         ██████░░░░  6.6  B

Higher is better (10 = clean, 0 = shit mountain). Scores are based on judgment, not formulas.

See fuck-my-shit-mountain/rubrics/scoring.md.

Natively Supported Platforms

This repository provides a standard SKILL.md + prompts + rubrics + templates skill directory, natively usable with the following tools:

Codex

1. Recommended: use the rsync command above to sync to ~/.codex/skills/fuck-my-shit-mountain/, or extract dist/fuck-my-shit-mountain.zip into ~/.codex/skills/
2. Restart Codex or start a new conversation to reload skill metadata
3. Request to use this skill in a Codex conversation

Claude Code

1. Personal install: copy to ~/.claude/skills/fuck-my-shit-mountain/
2. Project install: copy to .claude/skills/fuck-my-shit-mountain/
3. After starting claude, invoke via prompt or /fuck-my-shit-mountain

GitHub Copilot (CLI / VS Code Agent / Cloud Agent)

1. Personal install: copy to ~/.copilot/skills/fuck-my-shit-mountain/, or ~/.agents/skills/fuck-my-shit-mountain/
2. Project install: copy to .github/skills/fuck-my-shit-mountain/, .claude/skills/fuck-my-shit-mountain/, or .agents/skills/fuck-my-shit-mountain/
3. Run /skills reload in Copilot CLI
4. Confirm with /skills info fuck-my-shit-mountain

Gemini CLI

1. Personal install: copy to ~/.gemini/skills/fuck-my-shit-mountain/, or ~/.agents/skills/fuck-my-shit-mountain/
2. Project install: copy to .gemini/skills/fuck-my-shit-mountain/, or .agents/skills/fuck-my-shit-mountain/
3. Run /skills reload in Gemini CLI
4. Confirm with /skills list that the skill is discovered
5. If using project directory installation, ensure the workspace is trusted first.

When you update the skill content in this repo, prefer using package_skill.py or the rsync exclusion rules above to re-sync to the respective tool’s skills directory, then reload via each tool’s refresh method.

Example prompt:

Please use the fuck-my-shit-mountain skill to audit the current project
Mode: full
Report language: English
Output format: html

License

MIT

Learn AI on LinuxDo (https://linux.do/)