What makes agent-browser different from Selenium or Playwright?

While it uses Playwright internally, agent-browser provides an accessibility-tree snapshot and a simplified ref system designed specifically for AI agents to interact with web elements more reliably than traditional selectors.

Can I mock API responses with this skill?

Yes, the skill includes advanced networking features that allow you to intercept requests, block specific URLs (like analytics), and mock API responses for isolated testing.

Does it support authenticated sessions?

Absolutely. You can maintain login states using profile directories or environment variables, allowing you to run tests without re-authenticating every time.

Can I use this skill in CI/CD pipelines?

Yes, it is designed for integration with platforms like GitHub Actions and supports headless execution, automatic dependency setup, and artifact uploading for failed tests.

How do I debug failing tests?

The skill supports a 'headed' mode to see the browser in action, element highlighting, console log access, and full trace recording to analyze test failures.

AI-Driven E2E Browser Testing

Name: AI-Driven E2E Browser Testing
Author: jh941213

byjh941213

•

보안 및 테스팅

Automates stable end-to-end browser testing using accessibility snapshots and a reference-based interaction system.

The E2E Agent Browser skill empowers Claude Code to perform robust web automation and testing using the agent-browser CLI. Unlike traditional testing tools that rely on fragile CSS selectors, this skill utilizes accessibility trees and a unique 'ref' system to identify elements, making tests significantly more resilient to UI changes. It provides a comprehensive suite of capabilities including session management, network interception for API mocking, and multi-step flow automation for login, form submission, and complex UI interactions, making it an essential tool for modern QA and frontend development workflows.

주요 기능

0197 GitHub stars

02Accessibility-tree snapshots for reliable element identification

03Comprehensive UI automation including drag-and-drop and file uploads

04Headless and headed modes with built-in CI/CD integration patterns

05Advanced network control with request blocking and API mocking

06Stable interaction system using unique element references (refs)

사용 사례

01Running cross-page navigation and regression tests in CI/CD pipelines

02Validating frontend form logic and real-time validation messaging

03Automating complex multi-page user journeys like login and checkout flows

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jh941213/my-claude-code-asset e2e-agent-browser

For use in Claude.ai and ChatGPT

Download Skill