{
  "schemaVersion": 1,
  "revision": 27,
  "updatedAt": "2026-09-11T05:53:21.617142+00:00",
  "publishedDrafts": 25,
  "checkpointTarget": 100,
  "verifiedLocalAdditions": 0,
  "completedTreatments": 0,
  "activeBuilders": 0,
  "activeCritics": 0,
  "tasks": [
    {
      "id": "type-delimiter-before-word",
      "title": "Delimiter timing and empty hold",
      "status": "static-preference-recorded",
      "note": "Both supplied-file comparisons favor our variants, with local ties retained. A native-source replacement is prepared; the earlier preference cannot serve as its review.",
      "verification": null
    },
    {
      "id": "canvas-range-camera-return",
      "title": "Range selection and camera return",
      "status": "static-preference-recorded",
      "note": "Six consecutive lossless source parts are available, still one treatment. The bounded private single-file attempt failed; native-source re-review remains open. Earlier supplied-file preferences remain provisional.",
      "verification": null
    },
    {
      "id": "codex-three-file-handoff",
      "title": "Three-file selection and drag handoff",
      "status": "static-preference-recorded",
      "note": "Both supplied-file comparisons favor our variants. Shared post-drop occlusion and local losses remain. The exact-source replacement still needs a fresh review.",
      "verification": null
    },
    {
      "id": "layer-insert-resize-place",
      "title": "Layer insertion, resizing and placement",
      "status": "repair-required",
      "note": "Both fresh sixth-version comparisons tie. Our contours and frozen cursor glyphs improve local units, while the source keeps clearer device-rim and UI-wording detail. Both show the sticker partly outside the bottom edge during placement; intent is unresolved. Do not silently change native choreography to win a frozen-frame comparison. Native spatial coverage is 7.90%; timing and moving-review gaps remain open.",
      "verification": null
    },
    {
      "id": "split-result-pane-allocation",
      "title": "Pane allocation, blank wait and result",
      "status": "static-preference-recorded",
      "note": "Both supplied-file comparisons favor the changed variants. Pointer changes are editorial choices. Exact-source replacement review is still required; unexplained source glyphs and local unit losses remain recorded.",
      "verification": null
    },
    {
      "id": "continuity-id-across-cut",
      "title": "Screen-fixed identity across a cut",
      "status": "static-preference-recorded",
      "note": "Both supplied-file comparisons favor our variants. The lower-third alias counts as the same group. The source-exact two-part public reference is prepared; native-source re-review remains open.",
      "verification": null
    },
    {
      "id": "overlay-topic-stack-release",
      "title": "Topic arrivals, holds and ordered release",
      "status": "repair-required",
      "note": "Both native-source comparisons tie. The largest reported cup-clearance concern belongs to the source, not our candidates, and does not establish a candidate win. Unit results and native-spatial gaps remain open.",
      "verification": null
    },
    {
      "id": "work-observed-resolving-carousel",
      "title": "Resolving cards and interrupted contraction",
      "status": "blocked",
      "note": "Version three removes unnecessary identity 3D transforms; both actual package replays now match file bytes, all 145 RGB/YUV frames and clocks. A fresh visual review is still required. Its packet stopped at the historical September 10, 2026 20 GiB run cap; staging capacity under the September 11 approval is reserved before more packets. The gallery deliberately retains the source-losing first version and its failed replay history, not the unreviewed repair.",
      "verification": null
    },
    {
      "id": "shopping-table-browse-partial-return",
      "title": "Table browse and partial return",
      "status": "blocked",
      "note": "Version four is rendered and sealed with a persistent off-phone row guide, explicitly an editorial original-content adaptation. Both actual package replays match bytes, RGB/YUV and native clocks; the previous trajectory, mask and partial return are preserved. The prior TIE/TIE verdict remains. Its fresh critic packet stopped at the historical September 10, 2026 20 GiB run cap; staging capacity under the September 11 approval is reserved before more packets. The gallery keeps reviewed version three.",
      "verification": null
    },
    {
      "id": "atlas-title-two-line-arrival",
      "title": "Two-line title arrival and reading hold",
      "status": "blocked",
      "note": "The source wins both fourth-version comparisons. Version five changes raster sampling and has exact actual package replays, but its new review packet stopped at the historical September 10, 2026 20 GiB run cap. Staging capacity under the September 11 approval is reserved before more packets. The gallery retains the reviewed fourth version. Native-spatial gaps and duration-field differences remain unresolved.",
      "verification": null
    },
    {
      "id": "checkpoint100-plugin-rowwise-emphasis",
      "title": "Code rows and a growing emphasis block",
      "status": "blocked",
      "note": "The third version replaces the fractional highlight edge with a full-row fading backdrop. Both actual package replays match bytes, RGB/YUV frames and clocks. Three stricter new-row-only pixel checks still fail inside the emphasis field, and source duration fields differ by up to 0.6875 ms. Its fresh packet stopped at the historical September 10, 2026 20 GiB run cap; staging capacity under the September 11 approval is reserved before more packets. The gallery keeps the source-losing second version and its verdict.",
      "verification": null
    },
    {
      "id": "build-hour-source-intake",
      "title": "Ordinary long tutorial: source-only intake",
      "status": "source-inspection",
      "note": "A sealed intake covers source frames [105, 225), 3.5\u20137.5 seconds, of Build Hour: Agent RFT. It records the hard slide cut at 179\u2013180 with a persistent presenter inset. Fitted consecutive frames and native crops are not whole-video moving review. Novelty against the existing continuity group remains unresolved; no new render or group is counted.",
      "verification": null
    }
  ],
  "events": [
    "Independent native-Play testing found a first-click buffering stall that frame-button-first tests missed. The brief-clip delivery path has been corrected and is checked from a fresh native Play before any frame controls. Earlier failed attempts remain recorded. No media, native clocks, source annotations or fidelity judgments change.",
    "An independent delivery critic found that standalone second-render viewers overflowed before any interaction. The layout correction extends fitted player sizing to those viewers and confines opt-in native-size inspection to the player. Source footage, rendered clips, native timing and recorded fidelity verdicts are unchanged. Delivery review does not count as source-motion review.",
    "Live delivery exposed a real seekability failure: some raw clip responses produced an empty browser seek range. The paired players already buffered brief clips. Native and segmented players now use the same bounded, on-demand approach, and second-render links open native-clock viewers. No movie bytes, render, source timestamp or fidelity verdict changes. The first upload and its failed live check remain recorded; it was not accepted as a verified release.",
    "A separate publication reviewer found that the 11-group homepage link landed on a 10-comparison heading. This snapshot makes the 10 paired comparisons plus 1 segmented-source group explicit, adds direct links to all eleven groups, keeps re-encoded-reference qualifications visible, and makes narrow-screen task explanations readable. These publication corrections change no source movie, render, timing table or fidelity verdict.",
    "This snapshot prepares 11 under-review comparisons for publication. Live checks and a publisher receipt establish deployment. The 25 earlier published drafts plus these 11 distinct playable draft groups total 36, leaving 64 to reach 100. Ten paired studies and one segmented-source study are prepared. Revisions, content variants, source parts and the source-only intake do not increase the group count. The accepted public baseline remains 25; publication metadata is separate and grants no verified-ready or completion credit.",
    "Twenty-six static reviews are sealed: 15 candidate-preferred pairs, 23 source-preferred pairs and 14 ties across retained rounds. The fresh table and layer reviews each tie twice. These are review-history counts, not distinct successful treatments or moving-footage approvals.",
    "Both fresh critics and both bounded builders are finished. The table and layer reviews each tie twice; table native spatial coverage is 1.67% and layer coverage is 7.90%, with exact gaps retained. The latest four repairs, carousel version three, table version four, Atlas version five and plugin version three, await fresh independent comparisons. Their earlier reviewed renders and verdicts remain in the gallery; repairs do not inherit them. No builder or critic is active at this prepared snapshot.",
    "The carousel's third version has exact actual extracted-package replays after removing identity 3D transforms. The Atlas fifth version has a raster-sampling repair with exact replays. Neither inherits an earlier verdict. Their new critic packets stopped at the historical September 10, 2026 20 GiB run limit. Staging capacity under the September 11 approval is reserved before more packets.",
    "The latest four-repair mechanical audit independently decodes eight selected/replay pairs. Its append-only supplement resolves the initial failures as auditor accounting bugs: 15 Atlas members match at the declared extraction root, and both layer descriptors match their packaged generator contract. Nine corrected regression tests pass. The original failed report and seal remain unchanged; no builder file is altered to satisfy an auditor assumption. Scope exceptions and all zero credits remain.",
    "The September 8 inventory refresh exhausts 35 pages across Videos, Shorts, Live and the uploads playlist: 771 unique IDs. Seven additions have official feed publication dates after the frozen September 5, 2026 17:13:40.602033 UTC cutoff, so the in-window inventory remains 622. Earlier metadata was not fully refreshed; the uploads-only access exception remains unresolved.",
    "The last built gallery passed four browser configurations at September 10, 2026 04:09:23 UTC, with 30 paired checks, 10 alternate-render checks and 18 segmented/native-player checks. At that check, all 133 focused tests and four additional existing storage-bound tests passed; nine audit-correction regressions passed separately. Those historical checks do not validate this publication preview. Its build, publication-text scan, live checks and receipt belong to the publisher.",
    "The historical September 10, 2026 stop recorded allocated run usage of 21,474,320,384 bytes against the then-current 21,474,836,480-byte allowance, leaving only 516,096 bytes, about 0.49 MiB. The old guard counted logical file lengths; it now enforces the greater of logical and allocated usage, including directories and sparse files. Older logical reports cannot establish historical allocated peak usage. The normal 1 MiB tracker preflight rejected another build under that old limit; this is not a current usage measurement. No evidence is deleted.",
    "Full-video reviews, reviewed moving runtime, verified-ready additions, completed treatments and complete clean corpus passes remain zero. All public-facing motion findings retain inspection limits, cuts and source/proxy distinctions. New editable downloads remain withheld. Working comparisons are not fidelity approval or the requested 100-entry checkpoint.",
    "Accepted preservation was rechecked on September 10, 2026 at 04:17:55 UTC: 6,811 listed objects and the manifest remained unchanged, and the live manifest matched. No new deployment occurred at that check. It was read-only because further run-artifact writes were frozen under the historical 20 GiB limit.",
    "A tracker update correctly stopped when an active Chromium profile introduced temporary singleton links. The browser is closed and its runtime links are removed; the guard is not weakened. The failed attempt and the successful retry remain recorded, and no snapshot revision or source evidence was discarded.",
    "On September 10, 2026, snapshot 21, built at 04:07:21 UTC, remained the last served page and snapshot 22 was prepared only. On September 11, 2026, the user approved a 24 GiB run limit plus one additional staging copy capped at 3 GiB and requested public progress and examples. Staging capacity is reserved before more review packets; authorization alone does not establish a built stage or deployment. This snapshot 23 prepares under-review comparisons for the publisher without changing the accepted baseline, source evidence, verdicts or zero credits."
  ]
}