What is Retrieval-Augmented Generation (RAG)?

RAG is a technique that provides an LLM with relevant external context retrieved from a knowledge base, allowing it to generate more accurate and factual responses without retraining.

Which vector databases are supported by this skill?

This skill includes implementation patterns for several popular databases including Pinecone, Weaviate, Chroma, Milvus, Qdrant, and FAISS.

How does this skill help reduce AI hallucinations?

By forcing the LLM to use retrieved source documents as the primary context for its answers and providing citation patterns, the skill ensures the AI remains grounded in fact.

Does this skill provide code for hybrid search?

Yes, it includes implementation examples for combining semantic (dense) search with keyword (sparse/BM25) search to improve retrieval accuracy.

Can I use this for local development?

Absolutely. The skill provides configurations for local vector stores like Chroma and local embedding models from the Sentence Transformers library.

RAG Implementation Guide

Name: RAG Implementation Guide
Author: drgaciw

bydrgaciw

0•

Ciencia de Datos y ML

Builds high-performance Retrieval-Augmented Generation (RAG) systems to ground LLM responses with proprietary or external data.

This skill provides comprehensive patterns and best practices for implementing Retrieval-Augmented Generation (RAG) systems in AI applications. It covers the entire lifecycle of RAG development, from document chunking and embedding generation to advanced retrieval strategies like hybrid search and reranking. Whether building documentation assistants, proprietary knowledge bases, or research tools, this skill helps developers minimize hallucinations and ensure accurate, context-aware responses by integrating LLMs with scalable vector databases and efficient semantic search architectures.

Características Principales

01Evaluation frameworks for measuring accuracy, retrieval quality, and groundedness

02Multi-strategy document chunking including recursive, semantic, and markdown headers

03Advanced retrieval techniques including hybrid search and multi-query generation

04Vector database configuration for Pinecone, Weaviate, Chroma, and FAISS

050 GitHub stars

06Result optimization through reranking, contextual compression, and MMR

Casos de Uso

01Building a Q&A chatbot for proprietary internal documentation

02Implementing semantic search to improve information discovery in enterprise databases

03Developing research tools that provide source-cited answers from large datasets

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add drgaciw/academic-compliance-hub-glm rag-implementation

For use in Claude.ai and ChatGPT

Download Skill