Index local codebases and perform hybrid vector+keyword semantic search with SQLite.
npx skills add https://github.com/ddl2829/claude-skills --skill code-search-ddl2829
Please help me install this Agent Skill. Skill: Code Search Source: https://github.com/ddl2829/claude-skills/tree/main/skills/code-search Command: npx skills add https://github.com/ddl2829/claude-skills --skill code-search-ddl2829
This Skill allows you to quickly find relevant code snippets within your projects using natural language queries, overcoming the limitations of traditional keyword-based search.
Use the code search skill to find code related to handling payment processing in the 'app/services' directory.
High-intent search queries and answers about installing and using this skill.
Semantic code search uses local embeddings to find relevant snippets based on query meaning. It indexes your local codebase in a SQLite database, overcoming traditional keyword search limitations by understanding intent.
Yes, local repositories can be indexed securely without external data transfer. The process runs entirely on your machine utilizing local embeddings and the sqlite-vec extension to store and query vector data.
Semantic code search supports indexing and processing Python, PHP, TypeScript, JavaScript, and many other file types. It parses these languages to extract meaningful code structures for local embedding generation.
Incremental indexing and background indexing efficiently update the database as your codebase changes. This ensures large repositories remain searchable without requiring full rebuilds after every commit.
Hybrid search combines vector similarity with keyword matching to improve code navigation accuracy. This approach ensures you find functions related to concepts like 'user authentication' even when exact phrases are absent.
Yes, SQLite with the sqlite-vec extension and sentence-transformers are required dependencies. They provide the underlying local database storage and embedding generation capabilities needed for hybrid search functionality.
anderskevStore and query vector embeddings in SQLite using vec0 virtual tables.
blakebaumanSearches official AEM documentation for relevant feature details and implementation guidance.
junghan0611Synchronize semantic memory across local and remote hosts with incremental updates.
steipeteIndex local files and search them with BM25, vector, and hybrid queries.
