What problem does it solve? Sprint reviews often rely on dev-report claims that were never independently checked, so acceptance criteria get marked done without evidence. This Skill objectively maps every acceptance criterion from scope.md or sprint-plan.md to concrete evidence and marks each one MET, NOT-MET, or UNVERIFIED before any ship decision is made. ## Core Features & Use Cases - Evidence aggregation: Inventories test-results.md, qa-report.md, demo-output.md, flow-graph.json, and dev-report.md, warning when evidence is missing or stale. - Gap-fill verification: Runs each criterion's declared empirical signal directly via Bash when no existing artifact covers it, never trusting dev-report prose as proof. - Conformance report: Seeds and updates conformance.md with a criteria table, MET/NOT-MET/UNVERIFIED summary, standing derived-artifact rows (ADR-002), and plan-quality notes. - Use Case: After the dev team finishes a sprint and QA has driven the UI, run this Skill on the sprint directory to produce an objective conformance report that /sprint-accept uses to make the ship or no-ship call. ## Quick Start Ask the agent to verify the sprint deliverables in a given sprint directory, for example by saying: verify the sprint in docs/sprints/2024-09-checkout against its acceptance criteria.
