Grill
Use this skill only for an explicit request to grill, pressure-test, or stress-test a plan or design, or when the user directly invokes this skill. Complexity, ambiguity, or missing details alone never activate it.
Interview Contract
- Research discoverable facts in the repository, issue, docs, and code before asking the user. Ask only for judgments or facts that cannot be discovered.
- Ask one decision-forcing question at a time. State the recommended answer and the reason briefly, then wait for the answer.
- Prefer the host's structured input surface for bounded decisions:
- On TBH or Codex, use
request_user_input when it is exposed and the decision fits 2-3 short, mutually exclusive choices. Put the recommended answer first.
Send one question per call with id, header, question, and options; omit Other and None because the client provides a free-form escape. Never set the top-level auto_resolution_ms in a grilling interview: a grilling decision always waits for the human answer, because auto-resolving to the recommended default is never safe when the whole point of the interview is the user's judgment.
- On Claude Code, use
AskUserQuestion for the same bounded choice.
- Use plain chat only when the structured surface is unavailable or the answer needs free-form discussion.
Never pose a bounded checkpoint as plain assistant text or as a free-form question before offering structured choices.
- Use comparison tables only when the user explicitly requested one or the question concerns agent-product behavior, such as Claude Code versus Codex.
- Follow dependent decisions until the skill decides the decision tree is exhausted. Never ask a final "are we done?" meta-question.
- Summarize the settled contract: goals, non-goals, decisions, constraints, risks, validation, and unresolved items.
Ending the interview never authorizes implementation. Implement only after a separate explicit user request.
1---2name: grill3description: Run a decision-forcing interview only when the user explicitly asks to be grilled, pressure-tested, or stress-tested.4---56# Grill78Use this skill only for an explicit request to grill, pressure-test, or stress-test a plan or design, or when the user directly invokes this skill. Complexity, ambiguity, or missing details alone never activate it.910## Interview Contract11121. Research discoverable facts in the repository, issue, docs, and code before asking the user. Ask only for judgments or facts that cannot be discovered.132. Ask one decision-forcing question at a time. State the recommended answer and the reason briefly, then wait for the answer.143. Prefer the host's structured input surface for bounded decisions:15 - On TBH or Codex, use `request_user_input` when it is exposed and the decision fits 2-3 short, mutually exclusive choices. Put the recommended answer first.16 Send one question per call with `id`, `header`, `question`, and `options`; omit `Other` and `None` because the client provides a free-form escape. Never set the top-level `auto_resolution_ms` in a grilling interview: a grilling decision always waits for the human answer, because auto-resolving to the recommended default is never safe when the whole point of the interview is the user's judgment.17 - On Claude Code, use `AskUserQuestion` for the same bounded choice.18 - Use plain chat only when the structured surface is unavailable or the answer needs free-form discussion.19 Never pose a bounded checkpoint as plain assistant text or as a free-form question before offering structured choices.204. Use comparison tables only when the user explicitly requested one or the question concerns agent-product behavior, such as Claude Code versus Codex.215. Follow dependent decisions until the skill decides the decision tree is exhausted. Never ask a final "are we done?" meta-question.226. Summarize the settled contract: goals, non-goals, decisions, constraints, risks, validation, and unresolved items.2324Ending the interview never authorizes implementation. Implement only after a separate explicit user request.
shunkakinoki