/* LEEWAY HEADER — DO NOT REMOVE

REGION: AI.RAG TAG: AI.RAG.RAG_SYSTEMS

COLOR_ONION_HEX: NEON=#20B2AA FLUO=#23D8D8 PASTEL=#E8F5E9

ICON_ASCII: family=lucide glyph=zap

5WH: WHAT = rag systems skill for Leeway-compliant AI systems WHY = Provides capabilities for rag-knowledge within the AIskills ecosystem WHO = Leeway Industries (By Leonard Jerome Lee) WHERE = skills/rag-knowledge/rag-systems/SKILL.md WHEN = 2026 HOW = Leeway-governed skill.md definition with structured capabilities and tags

AGENTS: ASSESS AUDIT

LICENSE: MIT */

RAG (Retrieval-Augmented Generation)

Expert in: Building knowledge-grounded AI systems that retrieve and synthesize information from external sources.

Capabilities

• Design and implement RAG architectures
• Build document indexing and chunking strategies
• Implement semantic search and retrieval
• Create knowledge graph construction and querying
• Handle multi-source document fusion
• Implement retrieval evaluation and optimization
• Create feedback loops for RAG quality improvement
• Handle hallucination mitigation through grounding

Use this skill when:

• Building AI systems grounded in your data
• Creating knowledge question-answering systems
• Implementing document analysis and synthesis
• Building research assistants
• Creating domain-specific AI systems
• Reducing model hallucinations
• Handling context length limitations

Key techniques

• Vector embeddings and similarity search
• BM25 and sparse retrieval
• Hybrid search strategies
• Knowledge graph construction
• Entity extraction and linking
• Document chunking strategies (recursive, semantic)
• BLIP-2 and multi-modal retrieval
• ColBERT and learned ranking

Tags

rag retrieval knowledge-management embedding semantic-search grounding