# Feedback Lab demo gallery

34 actual browser screenshots captured on 11 November 2026 using Playwright Chromium.
All accounts and feedback shown are synthetic. The gallery contains static images;
it does not provide application login or access to the local demonstration backend.

## Exercise 1 — Web app
Login, empty list, required-field validation, creation, detail and status updates,
desktop/phone layouts, and the record created through the native feature.

## Exercise 2 — Native feature
Screens captured using Expo's browser renderer at phone width. These are not iOS
simulator or physical-device screenshots and do not establish native runtime validation.
The web and native feature used the same local backend/account. The CSV report created
and updated through Expo appeared alongside the Safari report in both clients.

## Exercise 3 — Static triage preview
Idle, queued, running, completed, failure, invalid output, accepted and dismissed states
at both widths, plus a prompt-injection fixture. Analyze/accept and invalid/retry/dismiss
were exercised. Focus handling and horizontal overflow were checked. All AI output is
canned synthetic data; no provider was invoked. The linked interactive preview is a
separate static artifact and needs no backend or login.

Desktop viewport: 1440 × 1000. Phone viewport: 390 × 844. Full-page PNGs may be taller.
Source: cumulative demo/03-robodev-artifacts; exercise-1 feature files are unchanged
across the three branches and the native feature matches branch 02. Exercises 4 and 5
were reviewed only.