Your personal skill library for Claude Code. Organize, analyze, and search skills with built-in security analysis. Zero dependencies — clone and go.
SaferSkills independently audited skill-vault (Agent Skill) and scored it 82/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 2 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 2 flagged
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
Your personal library for Claude Code skills.
Imagine you have a big toy box, but instead of toys, it holds skills — special abilities that make Claude Code smarter.
People are sharing thousands of skills online every day. The problem? They're scattered everywhere. You find one on GitHub, another on skills.sh, another someone shared on X... and then you forget where you saved them. Or worse — some of them might be harmful.
Skill Vault is your organized, safe toy box for skills. It does three things:
That's it. Clone it, open it with Claude, and start collecting.
git clone https://github.com/Hainrixz/skill-vault.git
cd skill-vault
claudeNo installs. No setup. No dependencies. Just open it and go.
When you open this project with Claude Code, the Vault Master agent takes over. Think of it as your personal librarian who:
| Command | What it does |
|---|---|
/vault-add | Save a new skill (from URL, paste, or name) |
/vault-search | Search your vault by keyword |
/vault-recommend | Describe your project, get matching skills |
/vault-discover | Find new skills from the internet |
/vault-list | See everything in your vault |
/vault-stats | Vault stats + health check |
/vault-remove | Remove a skill |
Skills are saved in category folders. Each skill gets its own file with metadata, security analysis, install instructions, and the original content preserved.
categories/
├── automation/ # Browser bots, scripts, task runners
├── code-quality/ # Linting, refactoring, reviews
├── design-ui/ # UI components, design systems, CSS
├── devops-deploy/ # Docker, CI/CD, infrastructure
├── documentation/ # READMEs, API docs, changelogs
├── organization/ # Project management, planning
├── productivity/ # Text tools, meeting notes, workflows
├── research/ # Web research, data gathering
├── testing/ # Unit tests, E2E, QA
└── web-development/ # React, Next.js, Vue, APIsDon't see a category that fits? The Vault Master creates new ones automatically.
Every skill is scanned before it enters your vault. The Vault Master checks for:
rm -rf, pipe-to-shell execution)| Rating | What it means |
|---|---|
| SAFE | Clean — no risky operations detected |
| CAUTION | Some risk — review the findings before using |
| DANGEROUS | Red flag — you'll be warned before it's saved |
Full methodology in security-rubric.md.
You: /vault-add https://github.com/anthropics/skills/tree/main/skill-creator
Vault: Analyzing skill... Rating: SAFE
Saved to: categories/productivity/skill-creator.md
Added to catalog.
You: /vault-recommend I'm building a Next.js e-commerce app
Vault: From your vault, I recommend:
- shadcn-ui (design-ui) [SAFE] — UI components
- playwright-cli (testing) [CAUTION] — E2E testing
- vercel-deploy (devops-deploy) [SAFE] — Deployment
You: /vault-discover SEO optimization
Vault: Found 3 skills on skills.sh:
- seo-audit [SAFE] — Technical SEO analysis
- meta-tags-generator [SAFE] — Auto meta tags
Add any of these? (y/n)/vault-add or create entries manually in categories/ using the template at templates/skill-entry.mdTodo de IA by @soyenriquerocha — Building tools for the AI community.
MIT
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.