hybrid-search-implementation

This skill provides patterns for combining vector similarity and keyword-based search to enhance retrieval accuracy. It is particularly useful when building Retrieval Augmented Generation (RAG) systems, improving search engine performance, or when neither vector search nor keyword search alone provides satisfactory results.

By combining these approaches, the skill addresses scenarios where semantic understanding needs to be coupled with exact matching, especially when dealing with queries containing specific terms like names or codes. It also improves search accuracy for domain-specific vocabulary, ensuring comprehensive results.

This skill guides you through clarifying goals, applying best practices, validating outcomes, and providing actionable steps, ensuring a robust and effective hybrid search implementation.

What It Does

Combines vector similarity and keyword-based search techniques to improve information retrieval, especially in RAG systems and search engines, yielding better recall and precision.

When To Use

Use when building RAG systems, combining semantic understanding with exact matching, handling queries with specific terms, or improving search for domain-specific vocabulary.

Installation

Copy SKILL.md to your skills directory