Verified Agent Skill record

playwright-driver

Use this skill when executing web tests via the canonical playwright npm package (Microsoft, Apache-2.0). Dispatched by skills/test-runner/ to execute web tests against a target, captures token-frugal AX-tree snapshots + screenshots + console output under .orchestrator/metrics/test-runs/<run-id>/, and exits with deterministic JSON output the orchestrator can parse.

Tasks

Run the Playwright test suite via the canonical invocation, setting PLAYWRIGHT_HTML_OUTPUT_DIR, PLAYWRIGHT_JSON_OUTPUT_FILE, and PLAYWRIGHT_HTML_OPEN environment variables so results land under the orchestrator-provided RUN_DIR
Pre-create ax-snapshots/ and screenshots/ directories inside RUN_DIR before spawning playwright test so test fixtures can write artifacts without racing

Inputs

RUN_ID and RUN_DIR environment variables provided by the orchestrator
PROFILE_PROJECT environment variable controlling browser selection (defaults to chromium)

Outputs

${RUN_DIR}/results.json — Playwright JSON reporter output the orchestrator parses
${RUN_DIR}/report/index.html — HTML report
${RUN_DIR}/test-results/**/trace.zip — trace bundles
${RUN_DIR}/screenshots/**/*.png — per-step screenshots written by test fixtures
${RUN_DIR}/console.log — captured console output as NDJSON written by test fixtures
${RUN_DIR}/ax-snapshots/*.yaml — Playwright AX-tree dumps written by test fixtures
${RUN_DIR}/ax-snapshots/axe-*.json — axe-core violation reports written by test fixtures

Limitations and checks

This driver does NOT use @playwright/mcp for browser drive — an R5 grep-canary blocks any reference to it under skills/playwright-driver/** or scripts/lib/test-runner/**
The @playwright/[email protected] package is unrelated and unstable and must not be used
The driver itself never reads or forwards AX snapshots — it only controls invocation
Never invoke 'npx playwright' per-test in a loop — install once and dispatch the local binary
Do not mix Playwright runs with different session IDs — use the orchestrator's RUN_ID everywhere
Artifacts must never be written outside RUN_DIR
Verify no references to @playwright/mcp or playwright-mcp exist under skills/playwright-driver/** or scripts/lib/test-runner/** to satisfy the R5 grep-canary
Confirm PLAYWRIGHT_JSON_OUTPUT_FILE points to ${RUN_DIR}/results.json and PLAYWRIGHT_HTML_OUTPUT_DIR points to ${RUN_DIR}/report before spawning playwright test