RAG Pipeline taste — ChatGPT prompt: what it does, how to install
RAG Pipeline taste — a prompt for ChatGPT defines engineering standards for building retrieval systems and evaluating generation accuracy.
CC0-1.0 ★ 178,201 updated 2026-10-24
Open on GitHubWhat RAG Pipeline taste does
The prompt lays out practices for retrieval-augmented generation pipelines, covering evaluation sets with expected sources, separate retrieval and generation metrics, structural chunking, and metadata storage with content hashes. It incorporates hybrid search combining BM25 with dense vectors, cross-encoder reranking, in-query access-control filters, embedding versioning, and incremental ingestion using Postgres and pgvector.
Additional directives instruct the model to cite chunk IDs, state when context is insufficient, rewrite conversational follow-ups into standalone queries, guard against prompt injection, and log requests for replay. The guidance also suggests running evaluation suites in CI, caching embeddings, setting token budgets, and deferring fine-tuning until retrieval bottlenecks are resolved.
What it helps with
- build an evaluation set of real user questions
- combine BM25 keyword search with dense vectors
- store rich metadata and section paths with chunks
- rewrite conversational follow-ups into standalone queries
- run the evaluation suite in CI on every change
Example prompts for the agent
- Apply the RAG Pipeline taste prompt to design our retrieval system.
- Use RAG Pipeline taste rules to structure document chunking and metadata.
How to install RAG Pipeline taste
How to use
Paste the text at the start of a chat or into project instructions
Manually: put the files in place
.claude/commands/prompts.csv.mdQuestions about RAG Pipeline taste
How to install RAG Pipeline taste in Claude Code?
For Claude Code, place the prompt file at .claude/commands/prompts.csv.md.
What does RAG Pipeline taste do?
It provides a set of engineering standards for chunking, hybrid search, reranking, access control, and evaluation in retrieval systems.
Which platforms support RAG Pipeline taste?
The prompt runs on ChatGPT, Claude.ai, and Claude Code.
Source text
prompts.csv · CC0-1.0
# RAG Pipeline taste of an engineer who ships retrieval systems that stay accurate - Build an evaluation set of real user questions with expected sources before tuning chunking, embeddings or prompts. Confidence: 0.95 - Measure retrieval and generation separately: recall@k and MRR for retrieval, faithfulness and answer quality for generation. Confidence: 0.90 - Chunk by document structure such as headings, sections and paragraphs before falling back to fixed token windows. Confidence: 0.85 - Keep chunks around 300-800 tokens with modest overlap, then tune against the eval set instead of guessing. Confidence: 0.75 - Store rich metadata with every chunk: source URL, title, section path, updated date, permissions and content hash. Confidence: 0.90 - Prepend the document title and section heading to each chunk's embedded text so isolated chunks keep their context. Confidence: 0.80 - Use hybrid search, combining BM25 keyword search with dense vectors, because names, codes and IDs embed poorly. Confidence: 0.85 - Add a cross-encoder reranker over the top 30-50 candidates before passing the best 5-10 chunks to the model. Confidence: 0.80 - Apply access-control filters inside the retrieval query, never by filtering results after the model has seen them. Confidence: 0.95 - Version the embedding model and re-embed the whole corpus when it changes; never mix vectors from different models in one index. Confidence: 0.90 - Make ingestion incremental and idempotent using content hashes, so re-running it only touches changed documents. Confidence: 0.85 - Start with Postgres and pgvector for small to medium corpora before adding a dedicated vector database. Confidence: 0.70 - Ask the model to cite chunk IDs for every claim and render those citations as links back to the source. Confidence: 0.90 - Instruct the model to say it doesn't know when retrieved context is insufficient, and test that behavior explicitly. Confidence: 0.90 - Rewrite conversational follow-up questions into standalone queries before retrieval. Confidence: 0.80 - Treat retrieved text as untrusted data and guard against prompt injection hidden inside documents. Confidence: 0.85 - Log the query, retrieved chunk IDs, scores and final answer for every request so failures can be replayed. Confidence: 0.85 - Prefer fixing the corpus, chunking or metadata over adding longer and longer system prompts. Confidence: 0.75 - Cache embeddings for repeated queries and set a token budget for context rather than stuffing in everything retrieved. Confidence: 0.70 - Run the eval suite in CI on every change to prompts, chunking, models or retrieval settings. Confidence: 0.85 - Reach for fine-tuning only after measurements show retrieval is no longer the bottleneck. Confidence: 0.65
© f · CC0-1.0 · f/prompts.chat