Solid foundation for building RAG systems that actually work in production. Covers the essential stack: vector databases like Pinecone and Chroma, embedding models including Anthropic's recommended voyage-3-large, and retrieval strategies from basic semantic search to advanced patterns like HyDE and parent document retrieval. The LangGraph examples show real implementation patterns, not toy demos. Includes hybrid search with BM25 plus embeddings, multi-query retrieval for better recall, and contextual compression to reduce noise. If you're building document Q&A, chatbots with external knowledge, or trying to reduce hallucinations with grounded responses, this covers the techniques that matter.
npx skills add https://github.com/wshobson/agents --skill rag-implementation