What problem does it solve?
Ensures a running autospec implementation actually satisfies the written product specification, preventing “it looks done” releases by auditing UI behavior, validation, state transitions, API effects, accessibility, and evidence-backed test coverage.
Core Features & Use Cases
- Spec traceability QA matrix: Produces a requirement-by-requirement PASS/FAIL/PARTIAL/NOT TESTED report so gaps are explicit instead of anecdotal.
- UI + workflow revalidation: Audits user-input controls (enabled/disabled, discovery, valid/invalid/boundary inputs), workflow controls, failure banners, and cross-feature behavior.
- No-mock deployed smoke + evidence-first verdicts: Requires deployed/dev, no-mock smoke coverage for user workflow controls and enforces a PARTIAL verdict when required smoke paths are NOT TESTED.
- Test regeneration/strengthening plan: Regenerates or strengthens tests when coverage is missing or weak, converting QA findings into actionable follow-up issues.
- Required artifacts and reliability gating: Produces machine-readable proof artifacts (proof matrix, reliability contract, control-intent ledger, mutation-proof, canary results when applicable) to support auditability and regression detection.
Quick Start
Use the autospec-qa skill to revalidate a feature spec against a running app by running:
/autospec-qa --spec docs/specs/<feature>.md --url http://localhost:3000