PRD writing

Produce a tight, decision-oriented PRD. Favor clarity over length — a strong one-pager beats a vague ten-pager. Use start_draft to save it and link it to the ledger item.

Structure (drop sections that don't apply, never pad):

1. Title & one-liner — what this is, in a sentence.
2. Problem — the user/business problem and evidence it's real. Who hurts, how often, how much.
3. Goals & non-goals — what success looks like; what's explicitly out of scope.
4. Users & use cases — primary persona(s) and the top jobs-to-be-done.
5. Proposed solution — the approach at a conceptual level. Include the key flows; note alternatives considered and why not.
6. Success metrics — the 1–3 KPIs that will tell us this worked (see the kpi-definition skill). Include a target and a guardrail metric.
7. Scope & milestones — phased: an MVP that proves the core bet, then fast-follows.
8. Risks & open questions — the assumptions that, if wrong, sink this; and what you'd do to de-risk them (ties to the prove stage).
9. Dependencies — teams, data, or systems you need.

Quality bar:

• Every goal has a metric; every metric has a number.
• The MVP is genuinely minimal and maps to the riskiest assumption.
• A reader can tell in 30 seconds what we're building and why it matters.

After drafting, move the ledger item to the right stage (usually define) and tell your manager what you produced and the top open question.