Browser Automation

Problem Solved

Browser automation failures often come not from business logic, but from fragile selectors, uncontrolled waits, and run-time detection. CSS and XPath can break after minor markup changes; manual waitForTimeout makes tests inconsistent; shared browser contexts can leak state between cases. This skill organizes these recurring issues into practical engineering guidance.

How It Works

It centers on browser automation with Playwright and Puppeteer, emphasizing how the browser interprets the page before writing stable scripts. Key practices include:

• User-facing locators: prefer role, visible text, and labels to reduce dependence on brittle DOM structure.
• Automatic waiting: remove waitForTimeout and rely on Playwright's explicit and implicit waits.
• Test isolation: use separate browser contexts or clean state per test to avoid shared cookie, localStorage, and dialogs.
• Diagnosable failures: enable trace, keep the viewport consistent, and add request pacing to reduce flaky results.

Boundaries

It fits e2e-testing, web-scraping, ui-automation, and selenium-alternatives workflows. It does not replace meaningful assertions or guarantee evasion of all anti-bot systems; for heavily protected sites, request cadence, privacy rules, and site policies still matter.