# Skill of Skills

> The quality-ranked directory for AI coding skills — discovered, scored, and categorized across Claude Code, Cursor, Codex, Windsurf, Cline, and n8n. Unlike a plain awesome-list, every entry is ranked QUALITY-FIRST: a composite of structural quality (55%), reputation (15%), recency (15%), and momentum (15%) — not raw stars.

Currently indexing 1047 active skills across 10 skill types, of which 176 have reached the Curated tier (quality score >= 120).

## Quality tiers
- Curated (>=120): high-viability, with gotchas, examples, and verification
- Promising (80-119): good structure and several quality signals
- Experimental (40-79): basic skill, limited signals
- Review Required (<40): not surfaced by default

## How to query the directory (public JSON API)
Base URL: https://skills.911fund.io. No auth required; be reasonable with request frequency.

- GET /api/v1/tools — paginated tools. Params: page, limit (max 100), category (slug), type (tool_type), platform, search, sort (composite|quality|stars|trending|recent). To get the best skills for a job, prefer sort=quality or sort=composite.
- GET /api/v1/tools/[slug] — one tool with full detail (scores, risk level + reasons, platforms, maintenance).
- GET /api/v1/tools/trending — quality-gated momentum.
- GET /api/v1/categories — the 9 skill types with counts.
- GET /api/v1/categories/[slug] — one skill type.
- GET /api/v1/search?q= — full-text search across the directory.
- GET /api/v1/stats — directory totals (tools, categories, combined stars).

## MCP server (agent-native access)
Connect your agent directly: Streamable HTTP endpoint at https://skills.911fund.io/api/mcp (no auth).
Example: `claude mcp add --transport http skill-of-skills https://skills.911fund.io/api/mcp`
Tools: search_skills (quality-first ranked search), get_skill (full profile incl. risk assessment), compare_tools, list_collections, get_collection, demand_pulse (k-anonymous demand signal).
Notes: scores are evidence-based quality metadata, not endorsements; discovery tools exclude high/critical-risk entries by default; no install commands are served — evaluate via repo_url/profile_url.
Sub-registry (official MCP Registry v0.1 list shape, our mcp_server subset with quality _meta): GET https://skills.911fund.io/api/v0.1/servers

## Browse
- Directory: https://skills.911fund.io
- Collections (best skill per job): https://skills.911fund.io/collections
- Skill types: https://skills.911fund.io/categories
- Analytics: https://skills.911fund.io/analytics

## Collections (curated, quality-ranked, cross-platform)
- [MCP Servers](https://skills.911fund.io/collections/mcp-servers): Model Context Protocol servers that connect your agent to real tools and data.
- [Database & Migrations](https://skills.911fund.io/collections/database-migrations): Schema design, query authoring, and safe migrations across SQL and NoSQL.
- [CI/CD & DevOps](https://skills.911fund.io/collections/cicd-devops): Build, test, ship, and operate — pipelines, containers, and infra-as-code.
- [Testing & QA](https://skills.911fund.io/collections/testing-qa): Unit, integration, and end-to-end testing that actually catches regressions.
- [Browser Automation](https://skills.911fund.io/collections/browser-automation): Drive, scrape, and verify the web with Playwright, Puppeteer, and friends.
- [Security & Audit](https://skills.911fund.io/collections/security-audit): Find vulnerabilities, leaked secrets, and risky patterns before they ship.
- [Frontend & React](https://skills.911fund.io/collections/frontend-react): Components, styling, and UI work for React, Next.js, Vue, and Svelte.
- [Code Review & Quality](https://skills.911fund.io/collections/code-review): Adversarial review, linting, and refactoring that raise the bar on every PR.
- [Data & Analytics](https://skills.911fund.io/collections/data-analytics): Wrangle, pipeline, and analyze data — dataframes, ETL, and dashboards.

## Top skills right now (ranked quality-first)
1. openai/codex (quality 95, Claude Code) — https://skills.911fund.io/tools/openai-codex
2. openai/openai-agents-python (quality 109, Claude Code) — https://skills.911fund.io/tools/openai-openai-agents-python
3. affaan-m/ECC (quality 160, Claude Code) — https://skills.911fund.io/tools/affaan-m-ecc
4. graphify (quality 165, Claude Code) — https://skills.911fund.io/tools/safishamsi-graphify
5. openai/skills (quality 90, Claude Code) — https://skills.911fund.io/tools/openai-skills
6. ruflo (quality 175, Claude Code) — https://skills.911fund.io/tools/ruflo
7. garrytan/gstack (quality 160, Claude Code) — https://skills.911fund.io/tools/garrytan-gstack
8. anthropics/claude-agent-sdk-python (quality 106, Claude Code) — https://skills.911fund.io/tools/anthropics-claude-agent-sdk-python
9. last30days-skill (quality 149, Claude Code) — https://skills.911fund.io/tools/last30days-skill
10. anthropics/financial-services (quality 95, Claude Code) — https://skills.911fund.io/tools/anthropics-financial-services
11. skills (quality 147, Claude Code) — https://skills.911fund.io/tools/add-skill
12. openai/codex-plugin-cc (quality 82, Claude Code) — https://skills.911fund.io/tools/openai-codex-plugin-cc
13. colbymchenry/codegraph (quality 146, Claude Code) — https://skills.911fund.io/tools/colbymchenry-codegraph
14. cline (quality 80, Claude Code) — https://skills.911fund.io/tools/cline
15. superpowers (quality 132, Claude Code) — https://skills.911fund.io/tools/superpowers

## About
Maintained by the 911 Fund. Updated continuously by an automated discovery -> quality-scoring -> categorization pipeline (README and directory refresh every ~3 hours; metadata daily).
Source & methodology: https://github.com/the911fund/skill-of-skills