Habilidades rag-implementation
📦

rag-implementation

Seguro

Build RAG Systems Step by Step

Também disponível em: wshobson

Implementing RAG systems requires coordinating multiple components including embeddings, vector databases, and retrieval pipelines. This workflow provides a structured approach to build production-ready RAG applications.

Suporta: Claude Codex Code(CC)
🥉 74 Bronze
1

Baixar o ZIP da skill

2

Upload no Claude

Vá em Configurações → Capacidades → Skills → Upload skill

3

Ative e comece a usar

Testar

A utilizar "rag-implementation". Help me start building a RAG system

Resultado esperado:

  • Phase 1: Requirements Analysis - Define your use case (document Q&A, semantic search, etc.), identify data sources, set accuracy requirements and latency targets
  • Phase 2: Embedding Selection - Evaluate embedding models, test domain relevance, measure embedding quality, consider cost and latency tradeoffs
  • Phase 3: Vector Database Setup - Choose vector database (Pinecone, Weaviate, Chroma), design schema, configure indexes, set up connections

A utilizar "rag-implementation". Which embedding model should I use?

Resultado esperado:

  • Consider these factors: domain relevance (test with your data), embedding dimension requirements, latency constraints, and cost per 1K tokens
  • Common options: OpenAI text-embedding-ada-002 for general use, Cohere for multilingual, sentence-transformers for self-hosted
  • Recommendation: Test top 3 candidates with your actual data using retrieval accuracy as the deciding metric

Auditoria de Segurança

Seguro
v1 • 2/24/2026

Static analysis flagged 45 potential issues (external_commands, weak crypto, system reconnaissance). Manual evaluation confirms these are false positives. The skill contains only markdown documentation with code block examples referencing other skills. No actual shell commands, cryptographic algorithms, or system reconnaissance code exists. This is a documentation-only skill providing RAG implementation workflow guidance.

1
Arquivos analisados
198
Linhas analisadas
3
achados
1
Total de auditorias
Problemas de Baixo Risco (3)
SKILL.md:31-196
False Positive: Markdown Code Block Backticks
Static analyzer flagged backticks as Ruby/shell command execution. These are markdown code fences (```) used for skill references like `Use @ai-product`. No actual shell commands present.
SKILL.md:3-130
False Positive: Weak Crypto Detection
Static analyzer flagged 'weak cryptographic algorithm' on skill reference lines. Manual review confirms these are skill names (ai-product, rag-engineer) with no cryptographic content.
SKILL.md:104-119
False Positive: System Reconnaissance Detection
Static analyzer flagged lines referencing other skills (similarity-search-patterns, hybrid-search-implementation) as system reconnaissance. These are legitimate skill references for workflow integration.
Auditado por: claude

Pontuação de qualidade

38
Arquitetura
100
Manutenibilidade
87
Conteúdo
50
Comunidade
99
Segurança
91
Conformidade com especificações

O Que Você Pode Construir

Building a Document Q&A System

Create a knowledge-base Q&A application that retrieves relevant documents and generates answers using an LLM.

Implementing Semantic Search

Set up semantic search capabilities over document collections using embeddings and vector similarity.

Optimizing RAG Pipeline

Evaluate and improve existing RAG system performance through systematic retrieval optimization.

Tente Estes Prompts

Start RAG Project 
Help me start building a RAG system. Use @rag-implementation to guide me through the initial phases.
Select Embedding Model 
I need to choose an embedding model for my RAG system. Walk me through Phase 2 of the RAG implementation workflow.
Configure Vector Database 
Set up a vector database for my RAG application. Follow the Phase 3 vector database setup workflow.
Evaluate RAG System 
Help me evaluate my RAG system quality. Use the Phase 8 evaluation workflow to define metrics and test approach.

Melhores Práticas

  • Start with Phase 1 requirements analysis before diving into technical implementation
  • Test retrieval quality with representative queries before integrating with LLM
  • Use evaluation metrics from Phase 8 to validate each phase completion

Evitar

  • Skipping requirements analysis and jumping directly to embedding selection
  • Using default chunking without considering document structure and query patterns
  • Integrating LLM before verifying retrieval quality with simple tests

Perguntas Frequentes

Does this skill execute code?
No. This is a workflow guide that provides guidance and references to other skills. It does not run any code or interact with external systems.
What other skills does this reference?
The workflow references skills like ai-product, rag-engineer, embedding-strategies, vector-database-engineer, similarity-search-patterns, and llm-evaluation.
Which vector databases are supported?
The workflow is database-agnostic. It provides guidance for setting up any vector database including Pinecone, Weaviate, Chroma, Milvus, and Qdrant.
Can I use this for production systems?
Yes. This workflow provides structured guidance suitable for production RAG implementations. Follow quality gates at each phase.
How long does full implementation take?
Implementation time varies based on data complexity and requirements. The 8-phase structure allows incremental progress over days or weeks.
Is this compatible with Claude Code?
Yes. This skill works with Claude, Codex, and Claude Code. Use it to guide AI assistants building RAG applications.

Detalhes do Desenvolvedor

Autor

sickn33

Licença

MIT

Repositório

https://github.com/sickn33/antigravity-awesome-skills/tree/main/skills/rag-implementation

Referência

main

Estrutura de arquivos

📄 SKILL.md