rag-skills

Use this skill when building, debugging, or improving Retrieval-Augmented Generation systems, including chunking, vector database selection, hybrid search, reranking, multimodal RAG, code documentation RAG, retrieval latency, and production RAG architecture.

Chunk nested documents into parent-child levels so retrieval can move from broad sections to fine-grained passages.

Use semantic boundaries and embedding similarity to chunk text for higher-relevance retrieval.

Route RAG chunking decisions across semantic, hierarchical, sliding-window, contextual-header, and framework-selection strategies.

Use overlapping windows to preserve context across chunk boundaries while controlling retrieval size.

Reduce retrieval latency with caching, batching, and index-level optimization.

Route RAG performance work for latency, caching, indexing, filtering, batching, and query optimization.

Dynamically select retrieval strategies based on query type and characteristics.

Dynamically evaluate and correct retrieval quality using relevance scores and web search.

Provide traceability, source attribution, and explainable results for RAG systems.

Use knowledge graphs to connect related information across documents for enhanced retrieval.

Combine BM25 and dense retrieval signals to improve search quality across keyword and semantic queries.

Precompute hypothetical questions during indexing to eliminate runtime query expansion overhead.

Use a two-stage retrieval pipeline with reranking to balance recall, precision, and latency.

Use hierarchical clustering and summarization to create multi-level document trees for improved retrieval.

Use self-reflective loops to make dynamic retrieval decisions and assess response quality.

Route RAG retrieval quality work across hybrid search, reranking, query transformation, HyDE, Self-RAG, RAPTOR, CRAG, and Graph RAG.

Choose a vector database based on content type, metadata needs, and query patterns.

Route RAG vector database decisions across Qdrant setup, production operations, and datastore selection by data type.

Select the right chunking framework based on document type, pipeline architecture, and retrieval goals.

Version the vector index and reindex or migrate embeddings without downtime or quality regressions.

Serve, scale, and safely roll out a RAG pipeline with health checks, caching, and canary releases.

Route production RAG rollout work across serving, scaling, index versioning, and safe reindexing.

Score RAG answers for faithfulness, answer relevancy, and correctness using LLM-as-judge methods.

Choose and wire up a RAG evaluation framework and turn it into a CI regression gate.

Measure retrieval quality with recall, precision, MRR, nDCG, and context precision and recall.

Route RAG evaluation work across retrieval metrics, generation quality, and end-to-end evaluation frameworks.

Route each query to the right retrieval strategy, data source, or tool with an LLM-driven router.

Coordinate planner and specialist agents to decompose, retrieve, and synthesize complex RAG answers.

Route agentic RAG work across query routing, tool-use retrieval, and multi-agent orchestration.

Add higher-level context to chunks by prepending headers before embedding for better retrieval.

Handle code-aware retrieval by preserving symbols, file structure, and API context.

Index and retrieve text, images, tables, and other modalities with shared document context.

Route RAG handling for code documentation, APIs, images, tables, diagrams, and multimodal content.

Add surrounding chunks to retrieved results to provide more coherent and complete context.

Use LLM-generated hypothetical documents to bridge the query-document gap and improve retrieval relevance.

Use query rewriting, step-back prompting, and sub-query decomposition to improve retrieval quality and comprehensiveness.

Run Qdrant reliably in production with scaling, backups, monitoring, and operational tuning.

Set up Qdrant for RAG with collections, payload filtering, and batch ingestion.