browser-automation

To prevent stale references in browser automation, use a snapshot-before-act workflow that captures page state before interacting, maintains consistent targetId usage, and retries once before reporting UI state changes as blockers.

Handle login and profile selection in web automation by running doctor, status, and profile checks before acting, then using snapshot-based reads and narrow act steps to navigate authentication while explicitly reporting manual blockers like 2FA prompts.

Browser automation scripts fail after modal changes because DOM references become stale, requiring a snapshot-retry mechanism to recover once before pivoting to explicit blocker reporting when the UI state changes unexpectedly.

The best way to manage browser tabs in multi-step web interactions is to reuse labeled tabs or stable targetIds, manage duplicates proactively, and avoid brittle tab addressing methods that break during state changes.

Recover from timeouts in multi-step web automation by implementing a snapshot-retry loop that re-reads page state before retrying the action, then explicitly reporting the timeout as a manual blocker if recovery fails.

Browser automation works with Playwright by stabilizing multi-step web interactions through snapshot-based reads, consistent targetId usage, and optional aria refs, preventing brittle clicking and stale references during navigation.