What problem does it solve? Verifying that a web application actually works after shipping code requires tedious manual clicking, form filling, and console checking. This Skill automates that exploratory testing and delivers a structured bug report with screenshots, repro steps, and a health score, while deliberately never modifying any code. ## Core Features & Use Cases - Report-only QA testing: Systematically explores pages, clicks interactive elements, fills forms, checks console errors, and documents issues with screenshot evidence, but never applies fixes. - Diff-aware mode: When run on a feature branch without a URL, analyzes the git diff to identify affected pages and routes, then tests exactly what changed against a locally running app. - Multiple modes and scoring: Supports full, quick smoke-test, and regression modes (diffing against a saved baseline.json), plus a weighted health score across console, links, visual, functional, UX, performance, content, and accessibility categories. - Use Case: After merging a branch that touches the checkout flow, run the skill to automatically detect the local dev server, test the affected pages end-to-end, and receive a report listing 5-10 evidenced issues with severity ratings and a Top 3 Things to Fix summary. ## Quick Start Ask the AI to run a report-only QA pass on your web app, for example by saying: run /qa-only on http://localhost:3000 and give me a bug report without changing any code.
