What problem does it solve? Turning a rough feature idea into a reviewable proposal takes PM discipline most developers skip in the local inner loop. This Skill captures feature requests as structured featurerequest.md files with problem statements, goals, non-goals, open questions, and assumptions, so a tech lead can take them straight to planning. ## Core Features & Use Cases - Structured Drafting: Creates a new featurerequest.md from a raw idea, transcript, or bullet list, resolving targets by full path or a numeric slot that expands to scratch/<MMDD>-<##>/featurerequest.md. - Iteration & Review: Reads an existing request, preserves untouched sections, and revises based on new feedback without silently dropping content. - GitHub Issue Seeding: Optionally fetches an existing issue via gh issue view to seed the draft title, problem, and notes. - Open Questions Walkthrough: Offers an interactive one-question-at-a-time walkthrough with recommended answers and rationales, applying each answer to the document. - Use Case: You have a one-line idea for a new CLI flag. Invoke the skill with slot 2 and your idea; it writes a complete proposal under scratch/, flags assumptions, and walks you through unresolved scope questions. ## Quick Start Ask the AI to draft a feature request for slot 3 describing your idea, for example: create a feature request in slot 3 for adding JSON output to the codegen CLI.