About this project

AgentSkillsHub (also known as Agent Skills Hub) is a continuously updated open-source directory that collects and indexes open-source AI agent skills, MCP servers, Codex skills, AI coding assistants, and agent frameworks from GitHub. It currently tracks over 117,000 projects, each classified into categories such as mcp-server, claude-skill, codex-skill, agent-tool, prompt-library, and ai-coding-assistant. Projects are evaluated on 10 weighted signals and 6 quality dimensions (completeness, clarity, specificity, examples, README structure, and agent readiness), producing a composite score from 0 to 100. The platform offers side-by-side comparison via /compare/ and scenario-based landing pages like /best/{scenario}/. Data is refreshed every 8 hours through an automated GitHub Actions pipeline. The architecture includes a Python 3.12 FastAPI backend, a React 18 + TypeScript frontend with Vite and TailwindCSS, and a Supabase (PostgreSQL) or SQLite database. The data pipeline consists of six phases: collection (GitHub API searches), cleaning (deduplication and classification), evaluation (quality analysis), scoring (composite score), and presentation (web and API). Key features include trending analysis (star velocity), rising projects, top-rated lists, verified creators, and a REST API with endpoints for skills, trending, stats, RSS feed, and sitemap. Deployment uses GitHub Pages and GitHub Actions, with optional Resend for email newsletters. Users can submit new tools via a form or GitHub issues, and community contributions are acknowledged. The project also provides documentation on its scoring algorithm and system architecture.