What problem does it solve? Plans and designs often hide unexamined assumptions that surface only during implementation. This Skill forces every material decision to be made explicitly through a structured one-question-at-a-time interview before any code is written. ## Core Features & Use Cases - Decision-tree interview: Asks exactly one question at a time, each with a recommended answer and its main tradeoff, until no unresolved decisions, assumptions, acceptance criteria, or risks remain. - Evidence-based questioning: Inspects the filesystem, repository, and configured tools directly instead of asking the user for discoverable facts. - Confirmed plan artifact: After explicit user confirmation, writes a structured Markdown plan (context, decisions, scope, acceptance criteria, risks, implementation outline, Jira fields) to .context/plans/YYYY-MM-DD-<slug>.md without touching tracked files. - Use Case: Before refactoring a service architecture, invoke the skill to be grilled on scaling constraints, data migration, and rollback strategy, then receive a dated plan file ready for implementation. ## Quick Start Ask the AI to use the grilling skill to interview you one decision at a time about your planned feature and produce a confirmed implementation plan in .context/plans/.