How does it improve RAG accuracy?

By designing advanced retrieval strategies such as hybrid search, query expansion, and reranking to ensure high faithfulness and context relevance.

Does this skill suggest specific technologies?

Yes, it recommends specific embedding models, vector databases, and LLM serving patterns based on your specific latency, accuracy, and budget requirements.

What is the rag-design skill?

It is a Claude Code extension that helps developers architect Retrieval-Augmented Generation systems by defining data ingestion, storage, and retrieval strategies.

Can it handle complex enterprise use cases?

Absolutely. It covers everything from simple FAQ bots to high-complexity agentic RAG systems requiring multi-hop reasoning and sophisticated metadata filtering.

What are the primary outputs of this skill?

It provides a complete architectural blueprint, including ingestion/retrieval pipelines, a recommended tech stack, quality targets (recall/MRR), and monthly cost estimates.

RAG Architecture Designer

Name: RAG Architecture Designer
Author: melodic-software

bymelodic-software

•

データサイエンスとML

Designs comprehensive Retrieval-Augmented Generation (RAG) systems including ingestion pipelines, retrieval strategies, and optimized technology stacks.

The rag-design skill provides an end-to-end framework for architecting RAG systems tailored to specific business requirements, from customer support chatbots to complex legal document search. It automates the complex process of selecting chunking strategies, embedding models, vector databases, and reranking approaches while accounting for latency, cost, and accuracy targets. By leveraging specialized subagents, this skill generates detailed architectural blueprints that include ingestion workflows, retrieval logic, and performance trade-offs, helping developers build scalable and faithful AI applications with confidence.

主な機能

01Comprehensive cost and performance estimation including latency budgets

0238 GitHub stars

03Specialized subagent invocation for deep architectural reasoning and trade-off analysis

04End-to-end pipeline design from document extraction to context-aware retrieval

05Automated selection of embedding models, vector stores, and reranking strategies

06Support for various RAG patterns including hybrid search, HyDE, and agentic workflows

ユースケース

01Building enterprise-grade customer support bots using FAQ and knowledge base integration

02Architecting high-accuracy legal search tools with citation grounding

03Designing technical documentation Q&A systems with code-aware chunking

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins rag-design

For use in Claude.ai and ChatGPT

Download Skill