How to Write Haystack RAG Pipeline YAML from a Retrieval Brief

Haystack YAML fails when it pastes sk- and invents gpt-4o. The brief named LocalChatGenerator and top_k 4. Stop there.

The matching generator is the Haystack RAG Pipeline YAML from a Retrieval Brief (No Invented API Keys or Indexes) (Haystack Rag Pipeline Yaml From Retrieval Brief No Invented Keys) prompt. Browse related cards in the PromptDig library (Browse more prompts). When a filled run survives, share the version you actually use (Share a prompt).

Quote Haystack version or write UNKNOWN

Write a Haystack RAG pipeline YAML from a retrieval brief. No invented API keys, indexes, or model IDs. Start by filling Inputs, not by asking the model to remember last week's run. If a field is blank, write NONE or NOT IN INPUTS and leave it blank through Generate. The card is built so the model cannot honestly invent a number, owner, URL, or command that you did not paste.

Paste these fields before you hit run:

Retrieval brief (retriever type, top_k I lock, generator I lock): [Brief]
Haystack version I lock (or UNKNOWN): [Version]
Component names I allow: [Components]
Index or document store I lock (or UNKNOWN): [Index]
Words I must not use: [Banned]
What I must never invent (keys, models, indexes): [Never]
Output format: [Format]
Language for comments: [Lang]
Env var names I allow for secrets: [Env]

That inventory is the honesty ledger. Anything that does not appear there is forbidden in the draft. If you catch yourself adding a nice-to-have after the run, you are no longer using the card. You are ghostwriting. Put the extra fact in Inputs and run again.

Wire only named components

Generate is numbered on purpose. Do not skip a step because the first paragraph looked done. The early steps exist to stop later prose from smuggling claims.

Walk the Generate list in order:

1. Honesty ledger: Brief nouns, Version, Components, Index, Env, Lang. Forbidden: raw API keys, invented model IDs, invented indexes.
2. Version note: quote Version or write VERSION UNKNOWN.
3. YAML body: only Components. top_k from Brief. Generator from Brief. Index quoted or INDEX UNKNOWN.
4. Secrets: only Env names. If a key is needed and Env is empty, write SECRET NOT IN INPUTS.
5. Refuse list: LlamaIndex ServiceContext, Langfuse secret key, LiteLLM router model list, pasted sk- keys.
6. Connections: wire only named Components. Do not add a ranker not listed.
7. Never: do not invent an OpenAI model string if Brief did not name one.
8. Compliance pass: quote Banned and Never hits. Cut them. Format as Format.

If a step asks for a version lock, quote the version from Inputs in the output. If a step asks for a refuse list, keep the refuse list in the published artifact, not in a sidebar you delete. Reviewers should see what the model was not allowed to do.

Secrets stay as env names not raw keys

Most failures are the same shape: a missing field gets a confident fill. A conversion rate appears. A Gradle task appears. A flash point appears. A caption appears on a job that asked for slide text only. Your review is to search the draft for numbers, names, and commands, then grep Inputs. No match means cut.

Honor the constraints as hard stops, not vibes:

• Haystack YAML from Brief. Not LlamaIndex, not Langfuse, not LiteLLM.
• Never invent API keys, indexes, or model IDs.
• Only Components named in Inputs.
• No emojis.

When the card says not legal advice, not certification, not an exam dump, or not a caption engine, that sentence belongs at the top of the output. Deleting it to look more finished is how you inherit risk.

Refuse LlamaIndex Langfuse and LiteLLM dumps

Finish with the compliance pass the prompt already asks for. Quote the banned-word hits. Cut them. Print character counts when the job has a cap. Print word counts when the job has a budget. List gaps as gaps. Five missing facts are more useful than one smooth paragraph.

Tags on the card (haystack rag pipeline yaml, no invented api keys, retrieval brief haystack) are a reminder of the job shape, not an invitation to wander into a neighboring cluster. If you need a different surface, open a different PromptDig card rather than stretching this one.

Fill the card, then run

Replace every bracket. Run on ChatGPT, Claude, or Gemini. Read the ledger first, then the artifact. If the model invents a commit, KPI, DOI, PEL, bid, or logo, discard the run. Tighten Inputs. Run again. Share the filled card that survived, not the first draft that sounded done.