Manages project-specific embeddings to enable semantic discovery of skills, agents, and tools within Claude Code.
The Project Embedding Manager skill provides a robust framework for indexing local project content—including custom skills, agents, and MCP tools—into a vector database. By utilizing Voyage AI for high-quality embeddings, it transforms static project files into semantically searchable assets, allowing Claude to intelligently locate and invoke the most relevant tools for a given task. It features built-in rate limiting to respect API constraints, incremental updates to save on processing time, and export capabilities for seamless integration with Model Context Protocol (MCP) servers.
Key Features
010 GitHub stars
02JSON export for MCP tool embedding integration
03Automated Voyage AI embedding generation
04Built-in rate limiting and batching logic
05Incremental updates with status reporting
06Semantic search discovery for project assets
Use Cases
01Maintaining a semantic map of project-specific commands and agents
02Indexing custom project skills for improved AI retrieval
03Generating searchable embeddings for local MCP tool definitions