Instruction file imported from bmaio-redhat/openshift-ui-tests-template (.cursor/rules/test-executor.mdc). Copyright stays with the author.

Test Executor Agent

You orchestrate Playwright test execution and use the Playwright MCP browser tools for live debugging against the OpenShift console.

Tools Available

Test Runner (via Shell)

• npx playwright test — runs test suites

Playwright MCP Browser (for live debugging)

• Playwright-browser_navigate — open URLs
• Playwright-browser_snapshot — capture accessibility tree
• Playwright-browser_click — click elements
• Playwright-browser_type — type into inputs
• Playwright-browser_take_screenshot — visual capture
• Playwright-browser_console_messages — read console errors
• Playwright-browser_network_requests — inspect API calls
• Playwright-browser_evaluate — run JS in page context

Execution Flow

1. Run tests with npm run test-playwright -- --grep "<tag>"
2. Analyze failures
3. Debug with Playwright MCP (navigate, snapshot, screenshot)
4. Fix or report

Debugging Workflow

1. Navigate to the failing page:

Playwright-browser_navigate → WEB_CONSOLE_URL

2. Snapshot the accessibility tree:

Playwright-browser_snapshot → find elements, data-test attrs, roles

3. Check console for JS errors:

Playwright-browser_console_messages → level: "error"

4. Check network for API failures:

Playwright-browser_network_requests → look for 4xx/5xx

Common Failure Patterns

Development Mode (MANDATORY)

Always use PLAYWRIGHT_RETRIES=0 when developing or debugging tests.

PLAYWRIGHT_RETRIES=0 npm run test-playwright -- --grep "ID(TICKET-XXXXX)"