Claude Skills
Enables AI models and coding assistants to semantically search and retrieve Claude Agent Skills using vector embeddings and the Model Context Protocol.
About
This MCP server brings Anthropic's powerful Agent Skills framework to the broader AI ecosystem, allowing any MCP-compatible AI model or coding assistant (like Cursor, Codex, or Gemini) to leverage specialized skills. It offers intelligent search capabilities by using vector embeddings for semantic similarity, implementing Anthropic's progressive disclosure architecture. This means skills are made available through multi-level loading, from metadata to full content and associated files, enhancing AI applications with curated scientific and official Claude skills without complex configuration.
Key Features
- Semantic Search for intelligent skill discovery
- Progressive Disclosure for multi-level skill loading
- Zero Configuration with curated skills out-of-the-box
- Multi-Source loading from GitHub repositories and local directories
- Fast & Local operation with automatic GitHub caching
- 9 GitHub stars
Use Cases
- Intelligently search for relevant skills based on task descriptions
- Retrieve specific skill files (scripts, data, references) from discovered skills
- Explore a complete inventory of all available Claude Agent Skills
Sponsored