| name | browser-automation-tests |
| description | Verify product behavior with Cursor browser automation by driving real UI flows, capturing failures, and turning findings into actionable repro steps or follow-up tests. Use when the user asks for browser automation, in-browser verification, UI smoke checks, or manual flow testing in Cursor. |
Browser Automation Tests
Use Cursor browser automation to validate UI behavior in a running app, reproduce issues, and gather evidence for follow-up fixes or automated tests.
Workflow
- Confirm the target URL, environment, and success criteria for the flow.
- Use browser automation to exercise the real UI path deliberately, one state change at a time.
- Capture the important evidence: page state, visible failures, console issues, or network problems when relevant.
- Summarize the observed behavior against the expected behavior.
- If the user wants durable repo coverage afterward, hand off to
write-e2e-tests or test-engineer.
Principles
- Prefer short, purposeful flows over broad unscripted browsing.
- Validate the user-visible outcome after each action.
- Stop and report blockers such as login walls, missing data, or environment issues instead of guessing.
- Use browser automation for real runtime verification; use unit or E2E code tests for durable coverage in the repo.
- Keep notes concrete enough to become a bug report or automated test later.
Output
- Flow exercised
- Expected behavior vs observed behavior
- Evidence collected
- Recommended next step: fix, add automated coverage, or investigate further
