OpenSkillIndex← back to search
claude-coderesearch

Babysitter rag-query-transformation

Query expansion, HyDE, and multi-query generation for improved retrieval

install
source · Clone the upstream repo
git clone https://github.com/a5c-ai/babysitter
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/a5c-ai/babysitter "$T" && mkdir -p ~/.claude/skills && cp -r "$T/library/specializations/ai-agents-conversational/skills/rag-query-transformation" ~/.claude/skills/a5c-ai-babysitter-rag-query-transformation && rm -rf "$T"
manifest: library/specializations/ai-agents-conversational/skills/rag-query-transformation/SKILL.md
tags
#rag-enhancement#query-expansion#hyde-implementation#multi-query-generation#query-decomposition
source content

RAG Query Transformation Skill

Capabilities

  • Implement query expansion techniques
  • Configure Hypothetical Document Embeddings (HyDE)
  • Set up multi-query generation
  • Design query decomposition strategies
  • Implement step-back prompting
  • Configure query routing for specialized indices

Target Processes

  • advanced-rag-patterns
  • knowledge-base-qa

Implementation Details

Transformation Techniques

  1. Multi-Query Generation: Generate query variations
  2. HyDE: Generate hypothetical answer, embed that
  3. Query Decomposition: Break complex queries into sub-queries
  4. Step-Back Prompting: Generate higher-level queries
  5. Query Expansion: Add synonyms and related terms

Configuration Options

  • Number of query variations
  • LLM for query generation
  • Decomposition depth
  • Query routing rules
  • Result fusion strategy

Best Practices

  • Match technique to query complexity
  • Test with representative queries
  • Monitor retrieval quality changes
  • Balance latency vs quality tradeoffs

Dependencies

  • langchain
  • LLM provider