What problem does it solve? Manually verifying that a web application works as specified is slow and error-prone. This Skill drives a real browser through the UI like a user would, testing success criteria, crawling pages, and stress-testing inputs, then producing evidence-backed pass/fail reports. ## Core Features & Use Cases - Criteria-Based Testing (Quinn): Executes each success criterion from a criteria file, capturing screenshots and structured evidence for every PASS/FAIL result. - Discovery & Adversary Modes: Crawls the app to build a page inventory, then attacks inputs, buttons, navigation, and viewports with extreme values to surface defects. - Parallel Multi-Agent Runs: Splits criteria across multiple agents, each bound to a dedicated Playwright MCP browser instance, for large applications. - Use Case: After implementing a new feature, run the full mode against your staging URL with a criteria file to verify every acceptance flow end-to-end, then receive a merged report with screenshots and reproducible failure steps. ## Quick Start Ask the agent to run qa-test against your app URL with your success criteria file, for example by requesting a QA run on https://myapp.com using qa-criteria.md.
