RAG Patterns

Use this skill to design, review, or optimize Retrieval-Augmented Generation systems, document Q&A, knowledge retrieval, agent memory, multimodal retrieval, and LLM context usage.

Invocation Triggers

• User asks for RAG architecture, chunking, embeddings, retrieval, reranking, vector databases, evaluation, or memory design.
• User is building document Q&A, agent memory, multimodal document processing, or retrieval tuning.
• User wants to reduce context usage through retrieval, caching, or memory patterns.

Core Workflow

1. Clarify data shape, query types, freshness needs, latency/cost limits, and quality bar.
2. Choose architecture tier: naive, advanced, agentic, or multimodal.
3. Select chunking strategy and metadata model.
4. Select embeddings and storage.
5. Design retrieval: vector, keyword, hybrid, query expansion, reranking, or routing.
6. Add memory, cache, and multimodal handling only when the use case needs them.
7. Define evaluation metrics and feedback loop.