What is RAG in LLM application development?

RAG (Retrieval-Augmented Generation) is a technique that retrieves relevant information from external knowledge bases and provides it to the LLM as context to produce more accurate and grounded responses.

How does this skill help reduce AI hallucinations?

By implementing grounding prompts and rigorous retrieval-based context, the skill ensures the LLM relies on verified provided facts rather than attempting to generate information from its training data alone.

Can I use this for local document indexing?

Yes, the skill includes guidance for both local vector stores like Chroma and FAISS as well as enterprise cloud-based solutions.

Which vector databases does this skill support?

This skill provides implementation patterns and configurations for major vector stores including Pinecone, Weaviate, Chroma, Milvus, Qdrant, and FAISS.

What retrieval strategies are included in this skill?

It covers dense retrieval, sparse retrieval (BM25), hybrid search, multi-query expansion, HyDE, and reranking with cross-encoders to ensure the most relevant context is found.

RAG Implementation & LLM Knowledge Integration

Name: RAG Implementation & LLM Knowledge Integration
Author: Krosebrook

byKrosebrook

•

Data Science & ML

Builds sophisticated Retrieval-Augmented Generation (RAG) systems to ground AI responses in external knowledge bases and private documentation.

This skill empowers developers to implement production-grade Retrieval-Augmented Generation (RAG) architectures within the Claude Code environment. It provides comprehensive patterns for vector database integration, embedding selection, and advanced retrieval strategies like hybrid search and reranking. By leveraging diverse document chunking methods and specialized prompt engineering, the skill helps you build factual Q&A systems, documentation assistants, and research tools that significantly reduce hallucinations and ensure AI outputs are grounded in specific, proprietary, or real-time data sources.

Key Features

01Seamless vector database integration for Pinecone, Weaviate, Chroma, and Milvus

02Multi-strategy document chunking including Recursive, Semantic, and Markdown splitters

031 GitHub stars

04Advanced retrieval strategies including hybrid search and cross-encoder reranking

05Comprehensive evaluation metrics for measuring groundedness and retrieval quality

06Retrieval-focused prompt engineering for accurate grounding and source citations

Use Cases

01Building a Q&A chatbot over internal company documentation and wikis

02Creating a research assistant that cites specific sources from a PDF library

03Implementing semantic search engines for proprietary technical datasets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add krosebrook/source-of-truth-monorepo rag-implementation

For use in Claude.ai and ChatGPT

Download Skill