Agent Browser Testing
Automates web browsers and executes end-to-end testing using Vercel's agent-browser CLI with ref-based element targeting.
Agent Browser Testing equips Claude Code with reliable browser automation and end-to-end testing workflows powered by Vercel's agent-browser CLI. By utilizing dynamic ref-based element targeting (@e1, @e2), the skill enables deterministic interaction with complex modern web pages without brittle CSS selectors or XPath queries. It provides full support for automated form filling, visual regression snapshots, session and cookie persistence, semantic accessibility locators, and automated web scraping, making it an essential companion for QA engineers, frontend developers, and AI agents needing direct browser manipulation.
주요 기능
01Comprehensive interaction suite including clicks, form filling, selects, hovering, and keyboard events
02Semantic locators for ARIA roles, visible text, and form labels to boost accessibility compliance
03Ref-based element targeting for reliable, dynamic UI interaction without complex selectors
04Automated snapshot generation with interactive element mapping and JSON exports
052 GitHub stars
06Session management to persist and restore authenticated states, cookies, and profiles
사용 사례
01Capturing screenshots and DOM snapshots for UI visual regression testing and debugging
02Executing automated end-to-end user journeys such as multi-step registrations and checkout flows
03Scraping dynamic, JavaScript-rendered web content and verifying web claims via live page interaction