to-prdlisted

This skill takes the current conversation context and codebase understanding and produces either: - a **full PRD** (default) — when the user wants to commit to building a feature and lock scope, or - a **defer capture** (Defer mode) — when the user wants to park a lead so future planning sessions can grab it without re-discovering the context. Pick by intent: PRD = "this is happening, lock requirements before we start". Defer = "remember this, not now". The issue tracker and triage label vocabulary should have been provided to you — run `/ai:setup` if not. ## Defer mode Trigger when the user says: - "defer this" - "track this for later" - "we'll do this in N" - "follow-up on …" - "open an issue + add to roadmap" - `/ai:to-prd defer <one-line lead>` Do **not** invoke Defer mode when the user wants the work done now — Defer is for deferral, not execution. Also distinct from [`to-issues`](../to-issues/SKILL.md), which breaks ONE plan into N tracer-bullet tickets. Defer is ONE surfaced lead → ONE issue + ONE roadmap paragraph. ### Process (Defer mode) #### 1. Read context Pull from the conversation: - The lead itself — one-line summary. - Why it surfaced — what review, plan, or research turn produced it. - Why-not-now — what blocks doing it in this session (scope, dependencies, decision pending). - Related issues — issue numbers mentioned recently, or sibling roadmap entries. If the lead is ambiguous, ask **one** clarifying question. Otherwise proceed. #### 2. Detect