Rag systems
Skill charlieviettq/awesome-agent-skill/.cursor/skills/ai-agent-systems/rag-systems
Curated skill pack for LLM agents in engineer and science workflow (Cursor & Claude ready).
npx -y skills add charlieviettq/awesome-agent-skill --skill rag-systemsAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 22 stars22 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Design retrieval-augmented generation pipelines—chunking, embeddings, retrieval, reranking, grounding, and evaluation. Use when building or improving doc Q&A, code search agents, or knowledge bases. Triggers: "RAG", "vector search", "embeddings", "retrieval", "knowledge base".
SKILL.md
1.8 KB, as published. Nobody here has run it
RAG systems
Pipeline stages
- Ingest — parse, clean, preserve structure (headings, tables).
- Chunk — size/overlap tuned to content type; keep metadata (source, section).
- Embed — consistent model; version stored with index.
- Retrieve — hybrid (keyword + vector) when recall matters.
- Rerank — optional cross-encoder for top-k precision.
- Generate — cite sources; refuse when context insufficient.
Chunking heuristics
| Content | Guidance |
|---|---|
| Docs | Split on headings; 300-800 tokens typical |
| Code | Function/class level; include path in metadata |
| Tables | Row batches or markdown table blocks |
Quality checks
- Retrieval hit rate on golden questions
- Answer grounded in retrieved chunks (no hallucinated citations)
- Latency budget per query documented
- Index refresh process defined
Failure modes
| Issue | Mitigation |
|---|---|
| Missed relevant doc | Hybrid search, query expansion, metadata filters |
| Wrong chunk | Smaller chunks, reranker, parent-child retrieval |
| Stale index | Version tag, scheduled re-embed |
| Prompt overflow | Summarize chunks, dynamic top-k |
Evaluation
- Golden Q&A set with expected source doc/section.
- Metrics: recall@k, answer correctness, citation accuracy.
- Regression when changing embed model or chunk strategy.
Security
- Respect ACLs at retrieval time; do not leak cross-tenant data in shared indexes.