What problem does it solve? Turning discovery output (jobs, research plans, syntheses) into a delivery-ready feature definition is slow and error-prone: scope items lose traceability to user jobs, acceptance criteria end up vague, and obligations get dropped between PRD, spec, and tickets. This Skill produces a feature-level PRD where every scoped item traces to a JTBD id and every requirement carries testable acceptance criteria tracked in a canonical ledger. ## Core Features & Use Cases - JTBD-traced PRD drafting: Writes docs/features/<slug>/prd.md from a template, grouping user stories by job id and tracing every MoSCoW item and requirement to a [[NN]] job reference. - Acceptance obligation ledger: Owns acceptance.json with requirement IDs, obligation IDs, validation methods (automated, rendered-review, human-study, performance-benchmark), evidence provenance, and immutable revision history. - Pipeline integration: Consumes jobs-to-be-done output, ateam-plan.md, research-plan.md, and page briefs without rewriting them, then hands off requirement IDs to epics and ticket-writer. - Use Case: After discovery lands for an "assisted triage" feature, invoke this Skill to draft the PRD with ranked-queue requirements, requirement-level acceptance criteria, and a validated obligations ledger ready for the definition gate. ## Quick Start Ask the agent to write a PRD for your feature using the prd-writer skill, pointing it at your existing jobs-to-be-done files and research plan.
