๐ QuizWizz Audits
The QuizWizz audit family is the evidence trail for one WordPress quiz plugin under audit on an isolated bench: what was measured, when, by what method, and what is still unresolved. Read it in two passes โ the themed groups tell you which page answers your question, and the dated archive underneath is the chronology, newest first. The current spine links standing authority and the latest dated evidence with their limits; a dated receipt is not planning authority or blanket acceptance. Every dated page remains a historical record of what was observed on its date, and any later correction is stated explicitly rather than silently rewriting the observation. Planning authority stays with the remediation masterplan. Nothing in this family authorises a release.
The cluster series
Four bench clusters landed on 2026-08-26 and 2026-08-27 and are packaged as one ordered, digest-enforced series. Cluster 3 was an audit that changed no file, which is why the applied series skips from 2 to 4.
flowchart LR C1["Cluster 1 ยท 2026-08-26<br/>play surface repair<br/>Wave 7 ยท 5 files ยท gate 1..51"] C2["Cluster 2 ยท 2026-08-26<br/>playable quiz repair<br/>Wave 8 ยท 3 files ยท gate 1..55"] C3["Cluster 3 ยท 2026-08-26<br/>performance audit<br/>changed no file, has no page of its own"] C4["Cluster 4 ยท 2026-08-27<br/>performance remediation<br/>Wave 9 ยท 14 files ยท gate 1..72"] C5["Cluster 5 ยท 2026-08-27<br/>not-found contract and token scope<br/>Wave 10 ยท 2 files ยท gate 1..76 at the time"] H["๐ฆ 2026-08-27 canonical handoff<br/>18 files ยท applies 1, 2, 4, 5 in that order"] C1 --> C2 --> C4 --> C5 --> H C2 -.->|"measured the baselines"| C3 C3 -.->|"implemented by"| C4
Ordering is a property of the cluster artifacts: the apply script runs clusters 1, 2, 4 and 5 in order (audits/qwizz/2026-08-27-apply-clusters.sh:56). The pushed HEAD 715474f, idle-bench shell 94/94, six bounded browser selectors at 47+9+13+8+7+7, sixteen mutation rows and recovered 95-row baseline belong to the amended Sep-2 receipt, not a fresh Sep-8 acceptance. The Sep-8 live punchlist receipt is the newest dated evidence: reconciled shell Wave 9 1..21 and all-wave 1..98 are green; original a-d and e pass; corrected f and g pass 5/5 with exact current-run cleanup; Fresh-Surfaces passes 9/9; and the corrected separate Editor lifecycle passes 8/8 for E03-08/10/11/12. Earlier red lifecycles and unknown capabilities remain explicit. The 19-file plugin packet and 39-file after-only harness supplement are noncanonical source artifacts, not release or publication authority.
At a glance
- plugin: QuizWizz (WordPress)
v1.10.0, DB schema1.6.8, on the WRDP bench - current decision: 2026-08-17 launch readiness โ public launch is
NO-GOuntil owner and canonical release inputs land - release boundary: bench only,
releaseEligible: false; no release, production publication or human acceptance is established by the Sep-8 receipt - dated open-work census: the amended Sep-2 open-work ledger records 154 rows โ 23 shipped + 21 bench-actionable + 50 owner-decidable + 49 terminal-blocked + 11 deferred. Its original Sep-2 additions were eight reviewer judgments, one owner-1 backup-cap hazard and two metric-reconciliation rows; the amendment adds shipped gate hardening
UIUX-GATE-01and open Wave 5 diagnosisUIUX-FLAKE-01. TOUR-19โs rendered-verification gap was closed without creating a new id. Sep-8โs 39 working identities are a separate census, including September S03 picks distinct from August canonical S03 findings; they are not added to or substituted for ledger, issue or cluster totals. - still open from the four clusters, and nothing else from them: passive embeds prompt before start, and removing that prompt would make a bare page view insert an attempt row, so it needs an operator decision. The zero-question draft population is closed as of 2026-08-27 โ it was measured precisely and deleted, and the previously published figure of about 109 was wrong; see the cleanup note in the changelog.
- evidence JSON on this host:
/home/loca/dev/wrdp/evidence/2026-08-17-option-f/verify.json
By theme
Every Markdown page in audits/qwizz/ appears in a themed group below; dated pages also appear in the chronological archive. The design/ child family follows them.
Current spine โ standing authority and latest evidence
- iOS restyle โ phase G, the newest dated bench change: the Settings panel restyled to one Apple inset-grouped visual language, shipped as vision round 6 plus a CSS-only focus amendment after the loop hit its six-round cap with
qualifiedRounds: [], so it is explicitly not converged. Six owned files โ five sheets plusincludes/Admin/views/settings-page.phpcarrying class attributes only, withassets/js/qwizz-admin-settings.jsbyte-identical to its pre-round copy, so both JavaScript residuals are recorded rather than half-made. Gates: shell1..98/0 on an idle bench and four bounded browser chunks at 47/47, 7/7, 7/7 and 10/10; the keyboard walk cleared 3 missing indicators and 8 raw clipped rings on desktop Settings and 3 missing on mobile, leaving two mobile strip clips at 25.95px and 34.61px overhang and an overall walk verdict of FAIL. A scroll-snap candidate was installed, measured ineffective and reverted. Eight new ledger rowsUIUX-SET-01throughUIUX-SET-08; tracked-modified rows 49 โ 51; mirror HEAD unchanged at715474f. No human acceptance;releaseEligible: false. - UX repair round โ round F, the newest dated repair receipt: six reviewer rows installed on the bench,
UIUX-REV-02andUIUX-REV-06closed on measurement as already resolved, bothUIUX-METRICtargets re-scoped to a plugin-owned counter set, and one measured regression (the focusable+Nchip at 28x21, L1-desktop owned small targets 8 โ 17) found and repaired inside the round at 32x32. Gates: shell1..98/0 on an idle bench, eight bounded browser chunks and a post-chip re-gate at 13/13 and 7/7. Identical-driver owned metrics improved on eleven rows with two disclosed measurement artifacts; the re-run vision critique moved 172 โ 170 findings with 0 new P1 and 6 new P2, added as seven owner-decidable visual rows. Tracked-modified rows 41 โ 49; mirror HEAD unchanged at715474f. No human acceptance;releaseEligible: false. - 2026-09-14 backup exposure cleanup โ dated bench hygiene receipt: 58 untracked
.bak/.bak-20260904-cutoverdroppings preserved privately with SHA-256 parity, then removed from the web-served plugin tree; public vhost droppings deny made suffix-aware; before/after URL matrix, 7/7 health, mirror 0 untracked. No source, database or Sep-8 boundary change;releaseEligible: false. - Wave 5 residue diagnostics โ UIUX-FLAKE-01 closed as diagnostics: Wave 5 assertion 12 prints its six residue components (mutation-proved), 18 idle/loaded runs 0 failures (
not reproduced at 4ร/8ร loops; hypothesis remains unverified), and the all-wave gate reconciled to the operatorโs 2026-09-09 Settings (play_page_idpinned to 0 for Waves 7-9 with byte-exact restore; tree digest re-pinned after a two-leaf proof) โ1..98/0. Harness only;releaseEligible: false. - mirror decision sheet โ decision inputs for the blocked Sep-8 mirror commit: 3-way census of the 41 tracked-modified plugin files against mirror HEAD
715474f(13 packet-on-drifted-base, 6 packet-clean-base, 5 inherited-unpacketed, 17 inherited catalog artifacts, 0 unattributed), a fresh canonical apply-check of the final packet (exit 1, the same seven paths as v3), and per-class options for the owner. Stages and commits nothing;releaseEligible: false. - Sep-8 live punchlist receipt โ newest dated audit-as-of record, not planning authority or acceptance. All 39 working identities remain bounded; E03-01 and several wording/taste decisions remain unresolved. Reconciled shell gates pass Wave 9
1..21and all-wave1..98. Final original a-d passes 7/7 and e passes 5/5. Corrected response-observer f and corrected g each pass 5/5 with exact current-run capability cleanup. Fresh-Surfaces passes 9/9 for S03-02/03/05/06. The first authorized Editor lifecycle remains red because of a policy-blind Theme oracle; the corrected separate Editor lifecycle passes 8/8 for E03-08/10/11/12 across desktop/mobile. Original f and the separate Editor observation remain non-substitutable, earlier unknown capabilities unresolved, and no human acceptance follows.- Database comparison: independent replay proves all 94 full seven-column backup rows, including 10 operator backups, exactly match the pre-task dump. The current local manifest inventories 50 public-safe assetsโ28 JSON and 22 PNGโwith exact digest/length/mode agreement, parsed JSON and unique PNG stored-byte/decoded-RGBA digests. The two Editor frames are verified paint-masked viewport geometry only. The final 19-file plugin packet remains a dirty-baseline, noncanonical artifact. The sealed 39-file after-only harness supplement has 39/39 copy/current parity but excludes runtime outputs/capabilities/config/PDFs and is not runtime or release authority. Mirror commit, production publication and human acceptance remain not performed.
- 2026-09-02 QuizWizz UI and UX perfection receipt โ amended โ historical predecessor, not current installed-byte or final-gate authority: ACCEPT-WITH-WAIVERS, zero blockers at that close. Pushed plugin HEAD
715474f; shell 94/94 on an idle bench; browser 47+9+13 plusuiux-perfectiononly asa-d,e-g,h-jat 8+7+7; mutations 16/16; recovered 95-row baseline. The amendment supersedes the original false SET-07 closure and records Wave 5 issue #15.releaseEligible: false. - ๐งญ Program masterplan โ the retained status SSOT: current state, evidence gaps, release-owner blockers.
- ๐ฃ๏ธ QuizWizz shared language โ canonical QuizWizz product and WRDP bench vocabulary; work state and evidence stay in the masterplan and probe harness.
- ๐ค Handover โ read this first when picking the bench up: current state, the seven non-optional rules, the bench-actionable rows in pickup order and verification commands. The durable copy under
assets/evidence/2026-09-02-uiux/belongs to the dated predecessor; the newest surviving corpus and its limits are linked from the Sep-8 live punchlist receipt. - ๐ Open-work ledger โ the single list of everything open, with policy state and implementation state kept apart.
- ๐งช Probe harness โ Wave 1โ11 shell contract on an idle bench plus six serial browser selectors, including three separately bounded UI/UX-perfection chunks, sixteen mutation rows, a non-vacuous Settings icon-gap denominator, disposable-identity attribution and the owner-1 no-write rule.
- ๐ฆ 2026-08-17 launch readiness โ the standing decision matrix: operational, runtime and rendered gates green, PW-02 closed, public launch
NO-GO. - ๐จ Owner decision intake โ the returnable packet: PW-01, legal and operator identity, log governance, alerting, off-host backup, image policy, release-owner inputs.
- ๐ฏ Answer-visibility policy โ the standing D005 ruling on who decides whether answers are shown, with its code citations and falsifier.
- ๐งญ Admin information architecture strategy โ the object-first plan for the settings surface, with quantified per-pane budgets. Still unimplemented, so its targets remain live.
- โ๏ธ QuizWizz operator defaults โ the standing operations contract for
wp-content/quizwizz-defaults.json: allowed leaves, nested precedence, failure behavior, portability boundary and the WRDP stage-and-install path. - ๐๏ธ Settings tiers โ the standing shipped Tier 1/Expert/search behavior and its browser/mutation evidence.
- ๐จ QuizWizz design board โ the standing 50/50 mess-archive reuse map, 8/8 S9 preflight, exact eight-site execution boundary and deferred deep-audit line.
Remediation clusters โ what was repaired on the bench
- ๐งฏ 2026-08-31 HOLD remediation โ the dated F1โF8 receipt at
87a0e61: five mutation-proved contracts, final 17/17, 93/93 and 45/45 gates, and the then-accepted zero-residue reading; superseded for current totals and residue truth by the 2026-09-02 receipt. - ๐งฐ 2026-08-31 corpus repair โ the completion driveโs accepted fixes, then-fixture-limited TOUR-19/20 source repairs, red-first contracts, per-slice mutation pairs, final
1..93shell gate and then-current 37/37 authenticated browser smoke. TOUR-19โs rendered gap closed on 2026-09-02; TOUR-20 remains fixture-limited. Mirror pushed throughf79feb6; bench-only andreleaseEligible: false. - ๐ 2026-08-30 choice-card prepend fix โ issue #6: the Settings Prefix and Print columns fieldsets painted native radios because their shared class had no base rule; one CSS-only commit and the mutation pair that proves it.
- ๐ฆ 2026-08-27 canonical handoff โ Clusters 1, 2, 4 and 5 as one ordered series for the release owner: 18 files, four patches, eight manifests, a transactional applier.
- ๐ง 2026-08-27 not-found contract and empty-card token scope โ Cluster 5: all three not-found conditions serve a real 404, and the unavailable card finally paints on the standalone route.
- โก 2026-08-27 performance remediation โ Cluster 4: the Cluster 3 audit implemented across 14 files, plus two proven-missing indexes at schema
1.6.8. - ๐ฎ 2026-08-26 playable quiz repair โ Cluster 2: dashboard Play links and the empty-render play route, both gated on a quiz actually being playable.
- โถ๏ธ 2026-08-26 play surface repair โ Cluster 1: play-route autostart, a full-bleed themed surface, and correct play-route document identity.
- โก 2026-08-25 settings memoization โ the per-request settings re-normalization that every translated string paid for, and the core option-hook trap that made the first attempt silently stale.
Corpus โ questions, subjects and capacity
Every count in this group is a dated pre-replacement measurement. Present-tense corpus, taxonomy-content and census work is deferred โ corpus replacement inbound in the open-work ledger.
- ๐ Question corpus โ the pre-replacement corpus history: how it grew across four intakes, why the old 9,069-question state still yielded few playable subject cards, which page answered which question, and why counts over different populations must never be swapped.
- ๐ง Question corpus strategy โ the corpus contract and the fusion pipeline.
- ๐ฌ 2026-08-30 corpus x-ray and question SSOT โ fresh 9,073-row live export anchoring one canonical row per published question, cross-layer reconciliation by normalized stem hash, the frozen 5,247-question bundle measured as pending, and the deterministic prompt/reveal TTS queue.
- ๐ฌ Question-corpus forensics โ the per-bank census of all 11 source banks: 12,271 rows holding 1,841 distinct stems.
- ๐๏ธ Subject sufficiency census โ every pool-count gate in the plugin, and the
can_satisfycutover that stopped the builder advertising subjects it cannot deliver. - ๐ 2026-08-19 question-corpus audit โ the inventory, validator and duplicate census, plus the local-corpus fingerprint count.
- ๐๏ธ 2026-08-19 legacy WXR import โ 8,401 nodes mined as a data source, 8,337 accepted, taking the corpus to 9,069.
- ๐ฆ 2026-08-19 corpus import โ 60 ledger candidates, 53 accepted, 7 rejected at the importer boundary, and the slash-handling defect the run uncovered.
- ๐ 2026-07-28 World Quiz Dataset import โ the first bulk intake: 229 rows accepted across 58 origins and 89 subjects.
UI and visual โ pixels and rendered behaviour
- ๐ 2026-08-31 corpus triage โ 25 measured verdicts, the FL-001 refutation, fixture-limited rows, corpus deferrals, and the vaultโs single batched owner-decision sheet.
- ๐ธ 2026-08-25 UI corpus โ the current visual corpus, superseding the 2026-08-20 gallery for every route it covers, and the 25 findings it raised.
- ๐งช 2026-08-25 TNDM UI review โ the resolution record for 36 verbatim operator findings, and the only place that operator wording is preserved.
- ๐๏ธ 2026-08-20 UI and UX feedback tour โ the largest rendered gallery in the vault, retained for the feedback dialogue; its pixels are superseded.
- ๐ 2026-08-16 PL-05 cross-surface discovery sweep โ 26 persona routes and 53 captures with zero navigation, console or overflow failures; PW-02 and PW-03 closed, PW-01 decision-required.
- ๐ฉบ 2026-08-15 visual and IA remediation โ the visual pass whose IA rework was deliberately not implemented.
- โ 2026-07-29 post-remediation visual audit โ 33 embedded shots verifying the July repairs.
- ๐ 2026-07-29 visual damage triage โ the 80-filename damage table behind that repair round.
- ๐ผ๏ธ 2026-07-28 Admin UI Atlas โ 42 embedded shots of the admin surface as it stood.
- ๐ UX Atlas audit โ the baseline visual ledger and its 2026-07-29 execution reconciliation. Archive despite the undated filename.
- ๐ฏ Play and print answers โ the dated evidence behind answer visibility: the anonymous 58-question browser receipt, the eight-row forced-policy matrix, and two retractions. The standing rule lives in the answer-visibility policy.
Release and provenance โ what a canonical owner would need
- ๐งฌ 2026-08-30 skill distillation โ why two bench-local skills carry QuizWizz vocabulary, gates, and the release boundary instead of installing the upstream set verbatim.
- ๐งพ 2026-08-17 installed-source preservation โ the 218-path installed-byte record. Bench evidence, not a release manifest.
- ๐ฆ 2026-08-16 W10 canonical-handoff packet โ the eight canonical inputs this bench cannot satisfy, tabled as why release is blocked rather than deferred.
- ๐ง 2026-08-16 post-W10 execution and verification โ the execution record behind that packet.
- ๐ฆ 2026-08-16 launch readiness โ superseded by the 2026-08-17 decision.
- ๐งพ 2026-08-16 installed-source preservation โ superseded by the 2026-08-17 record.
- ๐
ต 2026-08-02 F program โ the canonical recovery handoff, recorded at schema
1.6.6. - ๐ฆ 2026-07-30 release readiness โ the July release-readiness conclusion, evidence basis schema
1.6.5. - ๐งญ 2026-07-30 release readiness log โ the raw append-only question and decision chronology behind it.
- ๐ค 2026-07-30 canonical-port evidence packet โ the port packet as it stood.
- ๐ฉน 2026-07-28 settings remediation handoff โ obsolete patch handoff; do not use as a patch source.
- ๐ฉน Merge-ready patch spec โ obsolete code handoff; do not use as a patch source.
- ๐บ๏ธ Wave 6A classmap delta โ the bench-local stale-key table; canonical regeneration stays release-owner work.
Early archive and programme history
- ๐๏ธ 2026-08-19 masterplan narrative record โ the masterplanโs long-form narrative, moved out verbatim when the authority page was cut to 84 lines.
- ๐ Backlog history โ the dated sections preserved from the retired perfection-frontier backlog, including its explicitly superseded 2026-07-30 baseline.
- ๐ Tasklist completed log โ the append-only completed log preserved from the retired programme tasklist.
- 2026-07-31 W8 SSOT capability pass โ the pass that retracted the intermediate catalog figure and closed the German catalog at parity.
- 2026-07-31 W7 localization, cache and microcopy โ superseded by W8 on both the catalog count and the disputed translation.
- 2026-07-31 W6 regression sweep โ the W6 test matrix.
- ๐ฅ 2026-07-30 private service posture โ the July posture record; its open guest-AI frontier was authorised and delivered later.
- ๐งญ 2026-07-30 private service log โ the raw append-only chronology behind it.
- 2026-07-27 v1.10.0 full plugin audit โ the first whole-plugin audit, with a mermaid architecture diagram.
- 2026-07-27 architecture map โ the class and hook tables behind it.
- 2026-07-24 backend roast โ the opening backend measurement.
- 2026-07-24 frontend and UX audit โ the opening frontend measurement.
The design/ child family
Nine pages in audits/qwizz/design/, each naming its own replacement state.
- ๐จ Design surface audits โ the sub-vault index.
- Design masterplan ยท Design remediation patch spec โ the design programme and its obsolete patch handoff.
- Tokens and palette ยท Settings ยท Admin dashboard ยท Studio builder ยท Fine-tune editor ยท Player and library โ the six surface audits.
Archive โ dated records
The dated records in chronological order, newest first, plus the three undated pages that are historical rather than standing. Use the themed groups above to find a page and this list to place it in time. Frontmatter archive: true; each page stays where its links expect it, and none is planning authority.
- iOS restyle โ phase G: Apple inset-grouped Settings across six scored vision rounds plus a CSS-only focus amendment, class-attribute-only view change, JavaScript untouched, shell
1..98/0 and four bounded browser chunks, a 155-stop keyboard walk, a rejected scroll-snap experiment, eight newUIUX-SETrows, tracked-modified 49 โ 51. Not converged;releaseEligible: false. - UX repair round โ round F: eight reviewer rows and both metric rows closed, one measured regression repaired in the same round, shell
1..98/0 and eight bounded browser chunks plus a post-chip re-gate, owned metrics improved on eleven rows, vision critique 172 โ 170 with seven new owner-decidable visual rows, tracked-modified 41 โ 49.releaseEligible: false. - 2026-09-14 backup exposure cleanup โ 58 web-served backup droppings preserved and removed, nginx suffix bypass closed, Sep-8 boundary untouched.
releaseEligible: false. - Wave 5 residue diagnostics โ Wave 5 components printed and mutation-proved; 0/18 failures; all-wave reconciled to the Sep-9 operator Settings,
1..98/0.releaseEligible: false. - mirror decision sheet โ 41-row census vs
715474fwith classes, fresh final-packet apply-check and per-class owner options; nothing staged.releaseEligible: false. - Sep-8 live punchlist receipt โ newest dated audit-as-of record, all 39 working identities retained separately from canonical counts. Reconciled shell green; bounded browser/PDF outcomes; fresh f product geometry accepted but lifecycle red; final 19-file plugin packet offline-proved against exact dirty baselines with runtime and harness supplement incomplete. Mirror commit blocked. Not planning authority, release, publication, blanket cleanup or human acceptance;
releaseEligible: false. - 2026-09-02 QuizWizz UI and UX perfection round โ amended evidence receipt โ ACCEPT-WITH-WAIVERS, zero blockers, pushed mirror HEAD
715474f, shell 94/94 on an idle bench, six bounded browser selectors, mutations 16/16, 95-row recovered residue, two metric waivers and eight open reviewer judgments. The amendment visibly supersedes the original SET-07 closure and records the Wave 5 diagnostic issue #15.releaseEligible: false. - โจ 2026-09-01 QuizWizz mega-polish drive โ S0aโS10 evidence receipt โ ACCEPT with waived residuals; all four review P1s resolved. Nine drive commits through catalog close
d61d98c, accepted plugin fix at clean HEAD2da4b577, shell 94/94, bounded browser 60/60, mutations 2/10/11 and 95-row recovery. Residuals: #13, external-kill cleanup, benign tokens, content-identical saves. Correct for that dated close; superseded for current mirror, browser and residue authority by the 2026-09-02 receipt.releaseEligible: false. - ๐งฏ 2026-08-31 HOLD remediation โ F1โF8 evidence receipt โ dated final 17/17, 93/93 and 45/45 gates at
87a0e61, five mutation-proved contracts and the then-accepted zero-residue reading; retained verbatim as history, superseded first by the 2026-09-01 receipt and now by the 2026-09-02 receipt for current totals and residue. - ๐งฐ 2026-08-31 corpus repair โ completion-drive execution record โ accepted bench fixes, fixture-limited TOUR-19/20 source repairs, one rejected-and-superseded block-editor attempt, matched mutation pairs, the final
1..93shell gate and then-current 37/37 browser smoke, superseded as the standing browser total by HOLD remediation at 45/45, pushed mirrorf79feb6, and the explicit pre-publication receipt boundary. - ๐ 2026-08-31 corpus triage โ 25 verdicts and the owner-decision sheet โ one measured verdict per finding, design and site-content calls kept out of the fix set, FL-001 refuted, FL-003 unprobeable without an OpenRouter key, and corpus-dependent rows deferred.
- ๐ฌ 2026-08-30 corpus x-ray and question SSOT โ canonical ledger and TTS queue โ fresh full export reconciled against every surviving local corpus layer via the 16-hex normalized stem hash; one canonical row per published question, the frozen bundle measured 3 live / 5,244 pending and explicitly not imported, and an 18,146-asset deterministic prompt/reveal queue with a 24-clip voice audition.
- ๐ 2026-08-30 choice-card prepend fix โ issue #6 closed on the bench โ
.qwizz-choice-cardshad no base CSS rule anywhere in the plugin at mirror baseline7ec58f4, only two co-tenant slots in shared restraint-pass rules, so both Settings fieldsets that use it rendered native radios and plain inline labels. Closed by one 80-insertion CSS-only commitcf07530defining the cards plus their selected, hover, focus and forced-colors states, with a matched mutation pair on the pristine issue state, wave 61..14, the standing1..76, a 30/30 browser smoke, and a review rejection of a first forced-colors focus outline that resolved white on white. Closes open-work row 3.2.7; first live pass of the bench fix-loop skill;releaseEligible: false. - ๐งฌ 2026-08-30 skill distillation โ two bench-local skills instead of an upstream install โ decision record for distilling
qwizz-shared-languageandqwizz-fix-loopfrom the MIT-licensed upstream set while keeping bench vocabulary, gates, andreleaseEligible: falseexplicit. - ๐ฆ 2026-08-27 canonical handoff โ four bench clusters as one ordered, digest-enforced series โ Clusters 1, 2, 4 and 5 packaged for the release owner: 18 files, four patches, eight manifests and a transactional apply script whose ordering is enforced by six shared-file digest links. Proved by reconstructing a pristine tree, applying the series and comparing all 18 files byte-for-byte against the live bench; exposed and fixed a Cluster 4 patch that could not have been applied at all.
- ๐ง 2026-08-27 not-found contract and empty-card token scope โ Cluster 5 evidence โ the play route stops answering a missing quiz with the whole home page: all three not-found conditions now serve a real 404 at 55,655 bytes instead of 121,285, and two of them stopped answering HTTP 200; the shared unavailable card finally paints on the standalone route, proved byte-identical against page 1289 including forced-colors. Standing gate
1..76, Wave 10 mutation-proved, plus the measured decision to leave the 37ms dashboard aggregate alone. - โก 2026-08-27 performance remediation โ Cluster 4 evidence and canonical handoff โ the Cluster 3 audit implemented across 14 bench-local files: settings pools 48q โ 2q, play start 33q โ 5q, dashboard TTL lapse 134.1ms โ 2.23ms behind a bounded stale mirror, player surfaces โ5,160 bytes of stylesheet, and two proven-missing indexes at schema
1.6.8; gate at the time1..72, superseded the same day by Cluster 5, with Wave 9 mutation-proved across four subsystems, three output digests byte-identical, three coverage gaps published, and two millisecond targets missed for a measured reason. Raised one new finding โ the play-route unavailable card had no token-scope ancestor and never painted its border โ closed the same day by Cluster 5, which declares.qwizz-library__emptya token root atassets/css/qwizz-tokens.css:153,:238and:376. - ๐ฎ 2026-08-26 playable quiz repair โ Cluster 2 evidence and canonical handoff โ dashboard Play links and the empty-render play route, both gated on a quiz actually being playable, across 3 bench-local files; gate at the time
1..55with Wave 8 mutation-proved on the query clause, admin browser evidence, the reversible source patch plus preimage/postimage digest sidecars, and the two findings Cluster 1 left open both closed since: the route soft-404 by Cluster 5 (includes/Frontend/QuizQueryViewer.php:64-82, called at:21and:30), and the Wave 7 doc drift by the 2026-08-27 vault correction. - โถ๏ธ 2026-08-26 play surface repair โ Cluster 1 evidence and canonical handoff โ play-route autostart, a full-bleed themed quiz surface, and correct play-route document identity across 5 bench-local files; gate at the time
1..51with Wave 7 mutation-proved, guest browser and all-three-entry-point evidence, the reversible source patch plus preimage/postimage digest sidecars, and 4 findings recorded as not fixed. - โก 2026-08-25 settings memoization โ bench performance fix โ per-request settings re-normalization on every translated string;
init1919.9ms โ 51.0ms,/?page_id=27.625s โ 0.355s, plus the core option-hook trap that made the first attempt silently stale. - ๐ธ 2026-08-25 UI corpus โ recapture after the TNDM remediation โ 294 shots over 58 routes with 52 causes recorded; supersedes the 2026-08-20 gallery, confirms the item-09/27/28 fixes and the mobile rail, and raises 25 new findings.
- ๐งช 2026-08-25 TNDM UI review โ fixes โ 36 operator findings from tandem session 46937fb4893da860; direct bench fixes with per-item status; 3 reference-blocked items and the delivery-vehicle decision remain open.
- ๐๏ธ 2026-08-20 UI and UX feedback tour โ the complete UI/UX shot gallery for the 2026-08-20 feedback tour: 315 bench screenshots plus 5 evidence boards over 65 routes and four personas, with per-route feedback callouts and a living feedback log. Counted on disk 2026-08-27: 315 PNGs under
assets/qwizz-tour/2026-08-20/shotsand 5 underboards; the galleryโs own yamlsum records 315 on disk against a 320-entry shotlist, the 5 remaining entries being blocked shots kept with a cause each. - ๐ 2026-08-19 question-corpus audit
- ๐ฆ 2026-08-19 corpus import โ the 60 importable corpus survivors reduced to the 53 the deployed importer accepted across 5 packs: 7 multi-answer rows rejected because the validator allows exactly one
correct_idxand 2 duplicate source ids; published questions 679 โ 732 with index drift 0, and theqw_question_storedefect the run uncovered and fixed โupdate_post_meta()unslashed the string metas, so any question containing a backslash was unstorable behind an opaque import error. - ๐๏ธ 2026-08-19 legacy WXR import โ an 8,401-item legacy WordPress export mined purely as a data source and never fed to coreโs WXR importer: 8,379 cleaned en+de questions extracted, 22 rejected with named reasons, 42 already live by stem hash, and the remaining 8,337 imported through
QuestionPackImporterin 84 packs; published questions 732 โ 9,069 with index drift 0, gate1..46after the run. - ๐๏ธ 2026-08-19 masterplan narrative record
- 2026-08-16 PL-05 cross-surface discovery sweep
- 2026-08-16 post-W10 execution and verification
- ๐ฆ 2026-08-16 W10 canonical-handoff evidence packet โ the installed W10 source inventory, decisions and verification evidence frozen for a canonical owner before later bench changes blurred that boundary, with the eight canonical inputs this bench cannot satisfy (repository URL and immutable commit, locked build/sign tooling and dependency locks, provenance/SBOM/scope evidence, compatibility CI, signing context, compiled-block source and toolchain, owner review and regeneration, runtime quality gate) tabled as why release is blocked rather than deferred. Not a patch and not a release artifact.
- 2026-08-16 launch readiness (superseded)
- 2026-08-16 current installed-source preservation (superseded)
- ๐ฉบ 2026-08-15 visual and IA remediation
- 2026-08-02 F program โ canonical recovery handoff
- 2026-07-31 W8 SSOT capability pass
- 2026-07-31 W7 localization, cache, and microcopy
- 2026-07-31 W6 regression sweep
- 2026-07-30 canonical-port evidence packet
- 2026-07-30 ๐ฅ private service posture ยท ๐งญ log
- 2026-07-30 ๐ฆ release readiness ยท ๐งญ log
- โ 2026-07-29 post-remediation visual audit
- ๐ 2026-07-29 visual damage triage
- 2026-07-28 World Quiz Dataset import
- ๐ผ๏ธ 2026-07-28 Admin UI Atlas
- 2026-07-28 settings remediation handoff
- 2026-07-27 v1.10.0 full plugin audit
- 2026-07-27 architecture map
- 2026-07-24 backend roast ยท 2026-07-24 frontend & UX audit
- ๐ฉน Merge-Ready Patch Spec โ historical code handoff; not planning authority.
- ๐ฉน Design Remediation Patch Spec โ historical design handoff; not planning authority.
- ๐ฌ Question-corpus forensics โ measured per-bank census of all 11 question source banks across a 135-file, 6,520,043-byte walk: 12,271 rows holding only 1,841 distinct normalised stems, closing at 60 importable questions once the deployed validator floor and the dedupe policy are applied, plus stem-hash overlap, provenance verdicts, and the canonical schema against the deployed validator. Historical corpus record; not planning authority.