How do I start an audit in Claude Code?

Simply run the `/rag-audit` command. Claude will then scan your codebase for patterns related to vector stores like Pinecone or Chroma and libraries like LangChain or LlamaIndex to begin the analysis.

What does the RAG Audit skill actually check?

It evaluates your entire RAG pipeline, including chunking sizes, overlap configuration, embedding model selection, vector index types, retrieval thresholds, and generation prompts to ensure they follow industry best practices.

Does it support specific vector databases?

Yes, it is designed to recognize and provide specific advice for popular vector stores including Qdrant, Pinecone, Chroma, Weaviate, and Milvus.

Can it help reduce my OpenAI or Anthropic API costs?

Yes, by identifying inefficiencies in chunking and context window utilization, the skill suggests ways to minimize token usage while maintaining or improving response quality.

RAG System Audit

Name: RAG System Audit
Author: davicqueiroz

bydavicqueiroz

•

Ciencia de Datos y ML

Analyzes and optimizes Retrieval-Augmented Generation (RAG) pipelines for performance, accuracy, and production readiness.

The RAG Audit skill provides a comprehensive diagnostic tool for developers building AI-driven search and generation systems within Claude Code. It automatically scans your codebase for RAG components—including chunking strategies, embedding configurations, and vector store setups—to identify critical anti-patterns like context window overflow, missing rerankers, or poor metadata preservation. By validating implementations against industry standards, it provides a structured report with actionable code fixes and performance optimizations to ensure your RAG pipelines are accurate, efficient, and reliable for production deployment.

Características Principales

01Production-readiness checks for error handling and hallucination guardrails

02In-depth analysis of chunking strategies and metadata preservation

03Automated detection of RAG anti-patterns and performance bottlenecks

04Retrieval optimization including hybrid search and reranking suggestions

051 GitHub stars

06Validation of embedding models, vector store indices, and distance metrics

Casos de Uso

01Optimizing vector search performance and reducing LLM token costs

02Debugging retrieval failures or low-quality generation in AI applications

03Reviewing RAG implementations for quality and accuracy before production deployment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add davicqueiroz/claude-rag-skills rag-audit

For use in Claude.ai and ChatGPT

Download Skill