What problem does it solve? Static code analysis cannot catch runtime failures like broken routes, hydration errors, dead links, or unwired buttons. This Skill closes that gap by launching the app in a browser and verifying every screen and acceptance criterion from the PRD against the live application. ## Core Features & Use Cases - Multi-Agent Verification: Coordinates four specialized teammates covering page-by-page checks, acceptance criteria execution, AI vision screenshot review, and responsive/interaction testing. - Dynamic Scenario Generation: Builds a Feature Scenario Matrix from PRD sources and codebase analysis, covering feature flows, auth contexts, entity states, role permissions, and negative cases. - Dual Gate Decision: Produces a verification report with a PASS/CONDITIONAL PASS/FAIL gate combining percentage-based and fixability-based categorization, routing issues to remediation. - Use Case: After scaffolding and hardening a Next.js app, run Stage 6V to confirm every PRD screen loads, all acceptance criteria pass in the browser, and generate reusable Playwright regression tests. ## Quick Start Ask the agent to run visual verification of the running app against the PRD, optionally appending MODE: full, critical, or diff.
