用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/rohitg00/agentbrain --skill question-ladder命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
Agent Brain /brain-brief: after intake, research, and grill have enough signal.
基于 SOC 职业分类
正在显示 SKILL.md
| name | question-ladder |
| description | Use when uncertainty remains and the minimum useful sequence of questions should be asked. |
Lifecycle stage: INTAKE
Use when missing answers block product, design, or engineering decisions.
Do not use when this trigger is absent; choose the command or skill that owns the requested state, artifact, and verification gate.
| Shortcut | Rebuttal |
|---|---|
| "This is obvious." | Write the assumption and evidence. If you cannot, it is not obvious. |
| "We can do this later." | If the missing step changes the decision, do it now or state the risk. |
| "The user wants speed." | Reduce scope; do not skip the quality bar. |
| "This does not need verification." | Every important claim or behavior needs proof. |
Open Questions plus assumptions
The artifact should be concise, auditable, and include evidence, blockers, and next action so another agent can resume.
Trigger: uncertainty remains after retrievable evidence has been checked. Action: ask the smallest useful next question or choose the safest documented default in noninteractive mode. Output artifact: intake or grill artifact with blockers and next action. Verification: cite evidence checked, why the question changes action, and stop condition if no answer is available.
A plan is blocked by "needs approval." The ladder should ask the smallest next question first: what side effect is approval-gated, what safe read-only alternative exists, what assumption can unblock planning, and what evidence would change the route.