| name | question-upgrader |
| description | Use when someone is asking a question that is too narrow, too closed, or based on a hidden assumption that may be wrong. Upgrades weak questions and generates meta-questions to check whether the right question is even being asked. Triggers on: |
| scenarios | ["Is this even the right question to be asking?","Should I use microservices? (upgrade this)","How do I get my team to work harder?","이 질문이 맞는지 모르겠어, 더 좋은 질문으로 바꿔줘","어떤 질문을 해야 할지 모르겠어","이 문제에 어떻게 접근해야 할지 질문 자체가 불명확해"] |
| compatibility | {"recommended":["think-tool"],"optional":["sequential-thinking"],"remote_mcp_note":"think-tool이 있으면 메타 질문 분석(질문이 올바른가?)을 먼저 실행하고 그 다음 질문 업그레이드를 진행하는 순서를 지킬 수 있습니다. Claude 설정 → MCP Servers에서 remote SSE 엔드포인트를 추가하세요."} |
Question Upgrader
When to Use / When Not to Use
Use when:
- Someone asks a yes/no question about a complex topic
- The question has a solution embedded in it ("should I use X?")
- The framing of the question seems to be the obstacle, not the answer
- You suspect the question is answering the wrong level of problem
Not for:
- Questions that are already well-formed — confirm they're good and proceed
- Simple factual queries that don't have embedded assumptions
Process
Run Part 2 (meta-questions) first, then Part 1 (upgrade moves).
MCP note: If sequential-thinking is available, enforce this order: meta-question analysis → surface-form upgrade. Skipping meta-questions polishes the wrong question.
Part 2 — Meta-Question Check (run first)
Before upgrading, ask:
- Is this the right question? Does answering it perfectly solve the underlying problem?
- What assumption is embedded? Every question assumes something. Surface it.
- What would a more useful question be? Generate 3 alternatives at a different level or angle.
- Who is best positioned to answer this? Right person, data source, or part of yourself?
- What would change if you had the answer? If nothing — the question is academic, not urgent.
- What question are you afraid to ask? Usually the most useful one.
Part 1 — Upgrade Moves
| Weak pattern | Move | Stronger form |
|---|
| "Should I do X?" | Conditional | "What would need to be true for X to be right?" |
| "How do I get X?" | Why/What first | "What is actually preventing X right now?" |
| "Is X good?" | Stakes + context | "Good for whom, by what measure, over what timeframe?" |
| "What is the best Y?" | Scope + constraints | "Best for [specific goal] given [specific constraints]?" |
| "How do I fix Z?" | Diagnosis | "Is Z actually the problem, or a symptom?" |
Additional moves: inversion ("What would guarantee I fail?"), adding time horizon, removing embedded solution.
Output Template
- Original question — restated for reference
- What makes it weak — one sentence: closed/assumptive/embedded solution/wrong level
- Hidden assumption — what is being taken for granted
- Upgraded question(s) — 1–3 stronger formulations
- Why each upgrade unlocks better thinking — one sentence per
- The question you might be avoiding — always include, not optional
What Claude Does / What You Do
| Claude | You |
|---|
| Runs meta-question check to verify it's worth upgrading | Provide the question or problem framing |
| Names the specific weakness before upgrading | Confirm whether the hidden assumption resonates |
| Produces 1–3 upgraded questions maximum | Choose which upgraded question to actually pursue |
| Surfaces the question being avoided | Decide whether to confront that question |
Related Skills
assumption-extractor — for digging into the hidden premises a question rests on
clarity-toolkit — when the problem is vague goals rather than weak questions
mental-model-toolkit — for reframing the problem itself when the question reveals a stuck frame