cesnimda
4.0 KiB
Dashboard defaults and discovery density
4 November 2026. Public-language receipt09904ecc is published and independently remotely verified. New visual correction is within approved IM10, not a new design concept or final acceptance.
Dashboard: current36-configuration baseline passed72 scans/captures and36 fixture layout-save/reload/onboarding journeys on runtimefa6bd5d9. Real desktop captures show excessive unused widget space compared with approved A3. Only the initial defaults for stage-age, source, salary, company and skill widgets change from8 rows to4. Initial total height reduces from54 to42 grid rows. All10 widgets remain visible. Saved geometry, explicit hidden widgets, order, revision, deliberate gaps and existing collision/drag/resize/keyboard controls are unchanged. Sparse new widgets retain content scrolling on wide screens; mobile remains flowing content. This is a default-density improvement, not complete dashboard visual parity.
Discovery: the pre-change320px pilot passed three scans and the existing saved-search journey but visibly nests the saved-search collection around saved-search objects and frames each vacancy. Remove only the collection border/padding and make vacancies open divided rows. Saved-search object outlines, source attribution, dates, rights/geography semantics, filters, saved-search actions and result actions remain. Result overflow remains visible to preserve keyboard outlines. Phone screenshots use actual viewports so fixed mobile navigation is not misleadingly composited midway through tall full-page captures. Preserve the old pilot separately; it is not a full baseline matrix.
PASSED before build:63 dashboard/layout/grid/hook/onboarding tests in six suites plus38 discovery/offer tests in two disjoint suites (101 distinct); TypeScript, full lint and verifier syntax checks. The six offer date regressions are included in38, not additive to it. RED density case measured54 rows against the42-row bound; saved visible/hidden geometry regression remains in the green suite.
PASSED on settled immutable runtime a7d726738b74d13fb805f7bd4f5780bdf19110d0: export build; six actual offer timezone journeys/scans with zero mutations;36 dashboard journeys /72 scans/captures and36 intercepted layout saves;36 discovery journeys /108 scans/captures and180 intercepted saved-search mutations. Zero page errors, unexpected writes or axe violations in these final matrices. Pilots/baselines and overlapping reruns are not additive. Prior public504-state evidence belongs to its previous runtime and scope.
Capture review: the first discovery run captured the top of phone pages, leaving vacancy rows below the screenshot. Retained its report and one named sample. The corrected final108-state run uses the existing Jump to results action before phone results/saved captures, verifies focus and captures the viewport. Manually inspected320px results show three readable divided rows with source attribution and View listing / Save to tracker actions.1280px NB dark saved state shows open filters/results columns and the retained saved-search object boundary. Final1440px dashboard capture is shorter with all widgets still present; the frozen A3 prototype remains more compact, so V14 is only partially resolved. Prototype data differs and cannot justify inventing interview/date facts or deleting existing widgets.
Read-only source review notes found no actionable issues in default-only sizing, saved geometry preservation or discovery boundary/action/focus retention. The review agent subsequently reached its usage limit; no tests, browser checks or formal acceptance came from that agent. Local executed evidence above is authoritative. Dense populated-widget scroll ergonomics, physical device/zoom, native language and final owner review remain open.
Next: seal source/harness/evidence/export identities and publish/read back; continue calendar, documents/mail and Career route review. Full34-route/57-group, connected workflows, native/device/human and final security/release/owner visual acceptance stay OPEN.