Click Path And Browser QA

Purpose

Make browser-facing behavior testable by mapping important user paths to controls, routes, expected state changes, and verification evidence.

Inspect First

• app routes, pages, components, forms, buttons, links, dialogs, and Playwright tests
• docs/development/ui.md, observability/debugging docs, and existing browser QA reports
• local runtime scripts and healthcheck commands

Procedure

1. Inventory click paths.

   • Run scripts/inventory_click_paths.py <repo> --json when available.
   • Identify routes, forms, buttons, links, and missing accessible names.
   • Prioritize create/edit/delete/filter or equivalent user lifecycles.
   • Add scripts/click-path-audit when UI surfaces are part of the repo validation contract.

2. Define browser evidence.

   • Use assets/templates/click-path-audit.md.tmpl or browser-qa-report.md.tmpl.
   • Map each path to expected state, data persistence, error handling, and validation command.

3. Validate with the right tool.

   • Prefer Playwright for repeatable lifecycle checks.
   • Use $chrome-devtools-runtime-debugger for console, network, DOM, performance, or accessibility investigation.
   • Use $tdd-playwright-workflow when fixing a bug with a failing browser reproduction first.

4. Record gaps.

   • Missing e2e coverage, inaccessible icon-only controls, untested destructive actions, and console errors should become review findings or follow-up tasks.
   • When trajectory tracing exists, record UI and browser-evidence risk surfaces so $trajectory-risk-supervisor can distinguish tested from untested UI work.

Validation

• Run the narrow Playwright/browser command when available.
• Run ./scripts/click-path-audit when the wrapper exists.
• Confirm key controls have accessible names.
• Confirm browser findings include route, action, expected result, actual result, and evidence.

Completion Criteria

• Critical user paths are inventoried and have browser or Playwright evidence.
• UI workflow gaps are documented with concrete follow-up commands.