Skip to content
Get packs

RAG Architect

Designs and builds retrieval-augmented generation (RAG) search systems over your documents

Source Download skill
Repository
Jeffallan/claude-skills
Stars
12,176
Licence
MIT
Files
6 files

What it does

Teaches the agent how to build a system that lets an AI answer questions using your own documents. It covers splitting documents into chunks, turning them into embeddings (numerical representations for semantic search), storing them in a vector database, combining keyword and semantic search, reranking results, and measuring retrieval quality. The agent gets code patterns (Python with LangChain, Qdrant, OpenAI, Cohere and Ragas) and reference notes on comparing vector databases and embedding models.

Great for

  • Developers building a knowledge-base or document Q&A feature on top of an LLM
  • Engineers designing a semantic search or retrieval pipeline and wanting a structured checklist
  • Teams that want evaluation metrics (precision@k, recall@k, MRR) for retrieval quality

Skip it if

  • Non-developers looking for a finished app: it produces design guidance and code snippets, not a ready-to-run product
  • Projects handling sensitive documents without review: the examples send text to third-party embedding and reranking APIs, which may expose data
  • Simple keyword search needs: the skill assumes vector infrastructure and adds unnecessary complexity

Try asking

  • “Design a RAG architecture for our 50,000 internal support articles with multi-tenant filtering”
  • “Compare Pinecone, Qdrant and pgvector for a 10 million document semantic search system”
  • “Write chunking and embedding code for our PDF manuals and add a Cohere reranking step”
  • “Set up Ragas evaluation to measure context precision and recall on our question set”
Step-by-step guide: how to use RAG ArchitectInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 6 original files, the MIT licence and an attribution note.

    rag-architect.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip rag-architect.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip rag-architect.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the agent your real document types and scale so it can pick a vector store and chunk size that fit, rather than accepting the generic examples.
  • Keep API keys in environment variables or a secrets manager, as the skill advises, and review any code before running it against production data.
  • Ask for an evaluation set and metrics early; the skill's checkpoints (e.g. context_precision >= 0.7) are starting targets to validate on your own data.

What's inside

  • references/chunking-strategies.md
  • references/embedding-models.md
  • references/rag-evaluation.md
  • references/retrieval-optimization.md
  • references/vector-databases.md
  • SKILL.md

Redistributed unmodified from Jeffallan/claude-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.