Collect shared browser evidence

Use this procedure to collect evidence efficiently for the selected Checks. The operating contract owns permissions, public-session boundaries; the check evaluation workflow owns decisions and records.

Plan one evidence matrix

Map selected Check IDs to required routes, states, viewports, inputs and evidence. Group Checks that can use the same browser observation. Use the shopping journeys to reach shared states; do not treat every possible journey step as an additional audit requirement.

Verify the effective viewport in the page. Browser emulation supplies ordinary mobile/desktop evidence; a physical device, named engine, on-screen keyboard or assistive technology is required only when the selected Check or explicit acceptance criteria names it. A Chromium result does not prove another engine's behavior. If resizing fails, retry another safe host browser path before blocking the affected decision.

Collect once, evaluate individually

1. Record the initial route, viewport, theme/revision when known, market, consent and retained state.
2. Exercise the actions and alternate states required by the mapped Checks. Observe visible feedback and underlying DOM, accessibility, console or network state where relevant.
3. Capture direct screenshots at useful decision points, reusing one capture for Checks sharing the same evidence. Follow the screenshot requirements.
4. Return exact observations to each Check's decision record. A shared browser session or screenshot never substitutes for completing each Check's procedure.

If capture fails, retain DOM, layout, interaction, accessibility, console, network or source evidence. Missing screenshots block only decisions that need the image. A broken path may directly fail its owning Check; only distinct downstream decisions needing the unreachable state are blocked. Cart/checkout testing is outside Core scope.