Subtext UI Verification
Standardizes browser-based verification to provide visual proof and automated evidence for UI/UX changes.
The Subtext UI Verification skill acts as the central orchestrator for the Subtext framework within Claude, ensuring every UI-related task is accompanied by concrete evidence. It manages the invocation of specialized sub-skills to drive a real browser, capture session recordings, and attach visual proof like screenshots and viewer links to pull requests. By prioritizing verification over pure speed, it eliminates the uncertainty of visual regressions and provides reviewers with a clear 'live' look at changes, moving beyond simple code diffs to ensure what is shipped actually works as intended.
Características Principales
015 GitHub stars
02Automated browser-based verification for all UI/UX and frontend tasks
03Automatic generation of reviewer-facing evidence such as screenshots and session URLs
04Real-time polling of browser operator states to manage agent and human transitions
05Seamless orchestration of sub-skills including proof, live, review, and sightmap
06Intelligent subagent coordination to surface evidence back to the main orchestrator
Casos de Uso
01Validating frontend UI changes against a live application to prevent visual regressions
02Mapping application runtime models and component naming for new project onboarding
03Generating automated session recordings and evidence for complex pull request reviews