Hybrid Search Implementation

Combines vector similarity search with keyword matching using reciprocal rank fusion and linear score combination. Includes PostgreSQL implementations with pgvector for when you need both semantic understanding and exact term matching in RAG systems. The templates handle score normalization, metadata filtering, and cross-encoder reranking. Most useful when pure vector search misses specific names or technical terms that users search for literally. The RRF approach works well out of the box, while the weighted linear combination gives you more control when you know your data patterns.

skills.sh installs: 11.2K
GitHub stars: 42.4K
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add wshobson/agents --skill hybrid-search-implementation --agent claude-code

Installs into .claude/skills of the current project.

Files

Hybrid Search Implementation

Patterns for combining vector similarity and keyword-based search.

When to Use This Skill

Building RAG systems with improved recall
Combining semantic understanding with exact matching
Handling queries with specific terms (names, codes)
Improving search for domain-specific vocabulary
When pure vector search misses keyword matches

Core Concepts

1. Hybrid Search Architecture

Query → ┬─► Vector Search ──► Candidates ─┐
        │                                  │
        └─► Keyword Search ─► Candidates ─┴─► Fusion ─► Results

2. Fusion Methods

Method	Description	Best For
RRF	Reciprocal Rank Fusion	General purpose
Linear	Weighted sum of scores	Tunable balance
Cross-encoder	Rerank with neural model	Highest quality
Cascade	Filter then rerank	Efficiency