What are the hardware requirements for local RAG?

The skill requires at least 2GB of RAM to handle vector operations and local search efficiently within the Claude Code environment.

Does this skill send my documents to the cloud for indexing?

No, document indexing and vector storage happen entirely on your local machine using ChromaDB, though the final answer generation uses Claude Haiku via the Anthropic API.

Does this skill increase my Claude API token usage?

The retrieval bridge itself is designed for zero token use; tokens are only consumed when Claude Haiku generates the final synthesized answer based on the retrieved text.

What is the benefit of hybrid search mode?

Hybrid mode combines vector similarity for conceptual matching with keyword matching for specific technical terms, ensuring higher accuracy for technical documentation.

How do I add new documents to the RAG knowledge base?

You can index new content by running the provided indexing script pointing to your document folder, which populates the local vector database.

RAG Retrieval for Claude Code

Name: RAG Retrieval for Claude Code
Author: ItMeDiaTech

byItMeDiaTech

•

学习与文档

Integrates local document knowledge bases into Claude Code using semantic search and hybrid retrieval for context-aware development assistance.

关于

This skill empowers Claude Code with Retrieval-Augmented Generation (RAG) capabilities, allowing developers to query local documentation, codebase wikis, and private knowledge bases directly from the CLI. By leveraging ChromaDB for vector embeddings and a Multi-Agent Framework (MAF) for orchestration, it retrieves relevant context without extra token overhead for the bridge, providing accurate, cited answers powered by Claude Haiku while maintaining data privacy through local indexing.

主要功能

Multi-Agent Framework orchestration for intelligent info retrieval
Hybrid retrieval combining vector similarity and keyword matching
Zero-token bridge implementation for cost-efficient local context usage
Local semantic search with ChromaDB vector embeddings
6 GitHub stars
Built-in citation support for verifiable sources and document links

使用场景

Troubleshooting errors using locally indexed post-mortems and technical logs
Searching internal project documentation or API references without leaving the CLI
Onboarding to complex codebases by querying architectural guides and design docs

关于

主要功能

Multi-Agent Framework orchestration for intelligent info retrieval
Hybrid retrieval combining vector similarity and keyword matching
Zero-token bridge implementation for cost-efficient local context usage
Local semantic search with ChromaDB vector embeddings
6 GitHub stars
Built-in citation support for verifiable sources and document links

使用场景

Troubleshooting errors using locally indexed post-mortems and technical logs
Searching internal project documentation or API references without leaving the CLI
Onboarding to complex codebases by querying architectural guides and design docs