webapp-testing

The webapp-testing skill is a professional-grade toolkit designed for developers and QA engineers to automate, debug, and verify web interfaces within a local development environment. By leveraging the Playwright framework, this skill allows the agent to interact with dynamic web applications, capture browser screenshots, analyze console logs, and perform complex UI interactions programmatically. It simplifies the setup process by providing robust helper scripts to manage the lifecycle of local servers, ensuring that the backend and frontend are ready before automation tasks begin.

Key features and capabilities include:

• Native Playwright integration for high-fidelity browser automation and DOM manipulation.
• Multi-server lifecycle management via specialized helper scripts, allowing concurrent testing of complex stacks.
• Automated reconnaissance workflows that enable the agent to inspect rendered states, identify selectors through screenshot analysis, and wait for networkidle events.
• Support for debugging UI behavior, log capturing, and handling both static HTML and dynamic JavaScript-heavy applications.
• Pattern-based decision making to choose between direct HTML parsing or full browser-based interaction based on application complexity.

Usage notes and practical implementation details:

• Prioritize the use of the included helper scripts as black-box tools to manage server dependencies; use the --help flag to understand individual script capabilities.
• Always invoke wait_for_load_state('networkidle') for dynamic web applications to prevent DOM race conditions before inspection or interaction.
• For reconnaissance, utilize full-page screenshots and DOM inspection content to identify stable selectors (e.g., roles, labels, or IDs) rather than brittle positional coordinates.
• Follow the recommended decision tree: for static assets, parse directly; for dynamic apps, ensure servers are initialized using the provided utilities before running the automation script.
• Be mindful of context window limitations; treat large automation scripts as modular components and only ingest what is strictly necessary for the current task.