playwright-cli

Playwright CLI Adapter

Defer to the ui-browser-testing discipline skill for when/whether to run UI verification and how to record evidence. This skill covers how on the CLI surface.

Config commands

From [recipes.playwright-ui-flow.config]:

• Prefer ui_smoke_command for pre-merge smoke.
• Use ui_test_command for the full suite.
• If unset, discover (e.g. package.json scripts named test:e2e / test:ui / playwright) and propose config — do not hardcode forever.

Run exactly the configured string in the project root (or documented cwd).

Common CLI operations

# Suite / smoke (prefer configured commands above)
npx playwright test
npx playwright test --grep @smoke

# Browsers (once per machine/CI image)
npx playwright install

# Authoring helpers
npx playwright codegen <url>
npx playwright show-report

If playwright_config is set, pass it through (--config <path>) when the configured command does not already include it.

Pitfalls

• Prefer project-local Playwright (npx / package scripts) over a global binary so versions match the repo.
• Headless is normal for suite/smoke; headed/debug is for local investigation.
• Do not replace a failing suite run with ad-hoc MCP clicks when claiming verification — fix tests or record the failure.