What problem does it solve? Ambiguity in feature requests often surfaces mid-implementation, causing rework and scope creep. This Skill front-loads the discovery process by interviewing the user one question at a time before any code or plan is written, so decisions about scope, data models, security rules, and release steps are settled up front. ## Core Features & Use Cases - Structured Interview Process: Walks a branching decision tree covering goal, reuse, files affected, Firestore data layer, Cloud Functions, gating, mobile, release ritual, rollback, and open questions — one question per turn. - Standardized Plan Output: Produces a single Markdown plan file at CodeLog/ClaudesPlan/V{next}_{slug}.md with required sections including Goal, Approach, Files modified, Data model changes, Release ritual, Rollback, and Verification checklist. - Context Orientation: Reads recent git history, wishlist items, and prior plans before asking questions to avoid re-asking already-answered decisions. - Use Case: When a user says "let's plan V3.2 — add a missing-PDF queue to the library", the Skill grills them on scope, Firestore paths, security rules, and deploy order, then writes the versioned plan file for implementation in a later session. ## Quick Start Ask the assistant to grill you on your next feature idea and write the versioned plan before any implementation begins.
