| name | playwright-e2e |
| description | Use for committed deterministic browser tests, reusable assertions, and CI-safe regression coverage. Do not use for live browser attach, design-review passes, or exploratory-only browser work. |
Playwright E2E
Use this skill when the task should end with checked-in browser coverage that can run locally and in CI.
Read First
/hyperopen/AGENTS.md/hyperopen/docs/BROWSER_TESTING.md/hyperopen/docs/tools.md/hyperopen/docs/FRONTEND.md when UI surfaces change
Use Playwright When
- the browser flow should be committed, asserted, repeated, reviewed, or run in CI
- the flow is deterministic and can rely on repo-owned selectors or
HYPEROPEN_DEBUG
- smoke coverage or multi-viewport regression coverage is required
Do Not Use Playwright When
- the task is exploratory only
- the job requires live attach to an already running browser tab
- the job is Hyperliquid parity compare, governed design review, or selector discovery
Workflow
- Reuse existing
data-parity-id, data-role, and HYPEROPEN_DEBUG seams before adding new hooks.
- Keep helpers thin. Prefer exact assertions and
expect.poll over timing sleeps or a second scenario framework.
- Port only the stable local portion of a Browser MCP flow. Leave exploratory compare or design-review work on the Browser MCP side.
- Run
npm run test:playwright:smoke for the quick local suite and npm run test:playwright:ci for the full committed suite.
