RAG & Knowledge RetrievalPromptPlan
Query Decomposition For Multi-Hop Retrieval
Breaks a complex question into ordered atomic sub-queries optimized for a vector search retriever.
Opening lines · ~168 words in full
ROLE: You are a retrieval query planner for a multi-hop question-answering system. CONTEXT: Complex user question: [COMPLEX_QUESTION] Knowledge domain: [DOMAIN] Retriever type: [DENSE_VECTOR / BM25 …
The rest of “Query Decomposition For Multi-Hop Retrieval” opens on a plan
You are reading the opening lines. The full prompt (~168 words) — to copy, download as a ready .md file, or finish in the Studio — comes with the Library plan: every prompt and skill, .md downloads, 80 test runs a month.
- Built from
- Role
- Context
- Task
- Output format
- Constraints
How to use it
- Read it, then replace anything in [BRACKETS] with your details — the more concrete the context, the sharper the answer. The Studio lists the blanks for you and can add your project's background.
- Copy it (or download the .md) and paste it into the AI you already use — it knows your work, so that is where the prompt does the most.
- Not sure what it produces? Give it a test run in the Studio first, then refine with self-critique prompting.
Techniques in this prompt
Forces explicit intermediate reasoning instead of jumping to a conclusion, which improves accuracy on hard tasks.
Learn this techniquePins the response to a defined structure so it drops straight into your workflow.
Learn this techniqueAsks the model to reason step by step before answering — ideal for multi-step, logical, or analytical tasks.
Learn this techniqueWorks with
Any chat AI — ChatGPT, Claude, Gemini, Copilot, Grok, Mistral or a local model. The structure does the work, so you are not tied to one vendor or one model version.
New to structured prompts? Start with how to prompt AI, the RCTCO prompt framework this prompt is built on, and role prompting examples.