clarify
/clarify
Use when the user asks to clarify intent or stress-test an open-ended question before receiving an answer. Ask one decisive question when a decision-changing gap remains; otherwise answer directly.
- Author
- mel0nyrame ↗
- Repository
- skills ↗
- Licence
- MIT
- Last repository change
- 12.11.2026
Assessment and security
- Files in this package
- 3
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 12.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Use when the user asks to clarify intent or stress-test an open-ended question before receiving an answer. Ask one decisive question when a decision-changing gap remains; otherwise answer directly.
Original author description
Use when the user asks to clarify intent or stress-test an open-ended question before receiving an answer. Ask one decisive question when a decision-changing gap remains; otherwise answer directly.
How it works
- Context is sufficient
- A decision-changing gap remains
- After the user's reply
SKILL.md preview
--- name: clarify description: Use when the user asks to clarify intent or stress-test an open-ended question before receiving an answer. Ask one decisive question when a decision-changing gap remains; otherwise answer directly. --- # Clarify Check the supplied context for a decision-changing gap before answering. Use only that context during the initial intent check, without tools or file changes. A gap matters when plausible answers would change the recommendation, conclusion, scope, priority, or authorization. ## Context is sufficient Answer directly in the same reply. State only assumptions that materially affect the result; omit a sufficiency announcement and clarification headings. ## A decision-changing gap remains 1. Briefly explain the missing condition and why its answer changes the result. Mention a hidden assumption or relevant risk only when it helps the user answer; distinguish supported facts from possible risks. 2. Ask one decisive question once, about the highest-value missing decision dimension. When useful, follow it with mutually exclusive choices stated as options rather than another question. 3. Stop and wait for the reply. Keep substantive recommendations, plans, tool use, and file changes for after this question is resolved. Complete this stage when one question identifies the decision the user needs to make, without bundling unrelated questions or repeating supplied information. Use headings only when they improve readability; no fixed report template is required. ## After the user's reply Use the answer with the original request to give a specific, actionable response. Preserve material tradeoffs and uncertainty without exposing private chain-of-thought. Later actions follow the task's authorization; clarification alone authorizes no file changes. Ask again only if the reply leaves an ambiguity that prevents a correct or safe response. In that case, ask exactly one decisive question and pause.
Files in this package (3)
- LICENSE 1 KB
- SKILL.md 2 KB
- agents/openai.yaml 1 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · mel0nyrame/skills | 90 | MIT |
| ericrisco/rsc-harness | 100 | MIT |
| growupanand/ConvoForm | 100 | Apache-2.0 |
| mcouthon/agents | 100 | MIT |
| ramboz/jig | 100 | MIT |
| ramboz/jig | 100 | MIT |
| ramboz/jig | 100 | MIT |
| umputun/cc-thingz | 100 | MIT |
| ronakrupani/clarify | 100 | MIT |
Database snapshot: 13.11.2026 · package version e371f54c6d7a (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 100questions
Ask a person a question, record their answer, or withdraw a question, in the task's questions folder. Use when the person says "ask Sam ...", "ask <name> which ...", gives an answer to an open question ("Sam says ..."), or a question is no longer needed. - 100clarify-first
任务执行过程中的通用行为准则技能。适用于完成任何目标的执行全程,当遇到以下情况时触发:环境缺失(依赖/工具/权限不可用)、用户表意不清、出现重大决策性分歧、用户提供内容不足、缺失上下文——此时必须暂停任务,向用户澄清并索要必要信息。同时约束:对用户提供的任何内容/发言持合理准确性怀疑以纠正误操作/误输入;禁止擅作主张替用户做决定;一切数据来源必须有官方背书或实机测试数据支持,严禁凭空捏造/凭经验猜测;非本机 Android 设备的 UI 显示问题优先尝试 su 提权、失败再退回截图方案;Windows PowerShell 调用 adb… - 100pc-expatiate
Explain a point, question, or decision in depth so the user can decide. Covers what it means, the implications, the options with pros and cons, and a recommendation. Use when the user asks to expand, explain, or clarify something you raised (an open question, a risk, a blocked… - 90dev10x:ask
Surface what is still open in a session and route it where it survives. Reformulate plain-text decision questions into AskUserQuestion widgets, sweep the session for open loops and sync them to the task list, or output a reinforcement nudge. TRIGGER when: supervisor sees a… - 96answer-me
Use when the user invokes /skill:answer-me to ask a real question, not give an instruction; answer it and change nothing. - 100clarify-question-loop
Use when a meta-agent request is too ambiguous to safely generate, package, publish, or adapt without one to five targeted questions.