Does it support hybrid search implementations?

Yes, it specifically checks for hybrid search (dense + sparse) and reranking stages, which are critical for high-quality retrieval systems.

Can this skill help reduce LLM costs?

Yes, it identifies optimizations such as embedding caching, efficient dimension selection, and better retrieval filtering that can significantly lower operational costs.

What does a RAG audit evaluate?

It evaluates chunking strategies, embedding configurations, vector store setups, retrieval pipelines, generation prompts, and production readiness factors like error handling and logging.

How do I start a RAG audit in Claude Code?

Simply invoke the /rag-audit command. The skill will then search your workspace for RAG-related code and generate a detailed audit report.

Which vector databases are supported?

The skill identifies patterns for popular vector stores including Qdrant, Pinecone, Chroma, Weaviate, Milvus, and general vector store implementations.

RAG Pipeline Auditor

Name: RAG Pipeline Auditor
Author: floflo777

byfloflo777

•

Data Science & ML

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

The RAG Pipeline Auditor is a specialized tool for developers building AI-native applications that provides a comprehensive health check for retrieval-augmented generation systems. It automatically scans your codebase to identify embeddings, vector stores, and retrieval logic, checking them against industry best practices. By detecting common anti-patterns—such as poor chunking strategies, lack of reranking, or missing error handling—this skill helps you eliminate hallucinations, reduce costs, and improve the overall reliability of your AI features before they reach production.

Key Features

01Anti-pattern detection for chunking, embeddings, and vector search

02Comprehensive audit reports with severity scoring and priority fixes

03Automated codebase scanning for RAG-specific patterns and libraries

04Actionable code recommendations with before-and-after examples

05Production-readiness checks for error handling and observability

0617 GitHub stars

Use Cases

01Pre-deployment quality assurance to ensure RAG accuracy and reliability

02Optimizing vector store configurations for better cost efficiency and speed

03Debugging low-quality retrieval results or hallucination issues in LLM responses

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add floflo777/claude-rag-skills rag-audit

For use in Claude.ai and ChatGPT

Download Skill