spec(object-timeline): `items` still says "the author owns the item shape" — objectui#6356 declared the timeline element (feed | gantt, judged by `variant`)

[objectstack-fleet]

Filing-gate category: ④ a cross-repo child — the objectstack half of objectui#6356, filed by the seat that accepted its objectui PR (cross-repo rule 3: an accepted PR whose product flows into another repo gets its follow-up in the consuming repo at once). Reader: the objectstack domain:spec seat, which dispatches it once triage grades it and objectstack's .objectui-sha pin covers the objectui merge commit (the second reading, by REST compare). Filed by domain:ui seat #1 @ objectui, session_01DuWo5bdP9SdVebamn99GGk. ⛔ Not graded here.

Blocked-by: objectstack-ai/objectui#10822

What landed in objectui (PR objectstack-ai/objectui#10822, Fixes objectstack-ai/objectui#6356, merged 2026-09-27 as objectui 3a5817f07)

Maintainer ruling on objectui#6356, Q1 = A and Q2 = C:

• @object-ui/types TimelineSchema.items is now an element union, TimelineFeedItem | TimelineGanttItem, in place of an element with an open string index signature.
  • The feed arm (variant absent, vertical or horizontal) declares the documented seven keys: time, title, description, variant, icon, content, className. title is required.
  • The gantt arm (variant: 'gantt') is a row: label is required, and items holds bars (title, startDate, endDate, variant).
• The zod mirror judges each element against the arm the node's variant selects, with a node-level superRefine. A gantt row on a feed timeline is refused, and so is a feed item on a gantt one. The published TimelineSchema stays a ZodObject.
• The strict authoring face refuses undeclared keys by name (meta, group, color, startDate, endDate, _data, date). The tolerant face keeps .passthrough().
• The five keys ObjectTimeline composes from records are typed renderer-internal inside @object-ui/plugin-timeline and are not exported.
• objectui's object-timeline renderer types its schema as TimelineSchema & { … } (packages/plugin-timeline/src/ObjectTimeline.tsx ObjectTimelineProps). So the authored items it passes straight to the renderer are that same element union.

What the spec says today (objectstack origin/main 1207baf0)

packages/spec/src/ui/component.zod.ts:4400, in ObjectTimelinePropsSchema:

• items: z.array(z.unknown()).optional()
• It is described as 「Static inline entries … (the renderer becomes a pass-through and the author owns the item shape)」.

After the objectui merge, 「the author owns the item shape」 is no longer true: the element has a declared vocabulary. An author, human or AI, who reads the spec's ComponentPropsMap['object-timeline'] gets no guidance on it, and no refusal for a gantt row under a feed variant.

Direction (sequencing only, ⛔ not a shape ruling)

Order of work, decided by the objectui seat in box 5857426543 on objectui#6356 and not overturned by the ruling 5857468846: objectui first, then the spec follows. The element's implementation, docs and only producer live in objectui.

Which spelling the spec takes is the spec seat's to derive, under the posture this file's header gives for a contract that lives in objectui (mapping stays z.unknown() for that reason). The candidates:

• declare the element by name (feed and gantt arms, judged by variant), or
• keep z.unknown() and correct the describe, so that it names @object-ui/types TimelineFeedItem / TimelineGanttItem as the owner.

Either way, the pinned read points (:546 if (schema.items) return schema.items; and :745 hasAuthoredItems) are re-read at the pin that covers the merge.

Dedupe

REST listing of the 1,000 most recently updated objectstack issues and PRs (open and closed, down to #19633), grepped locally:

• object-timeline near items → 1 hit, [finding] the tracked root sdui.manifest.json and the browser dump differ on the fields the parity gate consumes (18 of 57 components; dataSource missing on 15 in the root artefact) — the gate's verdict set is identical under both today, which is the configuration in which the drift is least visible #17735 (closed; the sdui manifest parity, not a carrier)
• TimelineFeedItem / TimelineGanttItem → 0
• objectui#6356 / objectui#10822 → 0
• author owns the item shape → 0
• Lit control, object-timeline → 7 hits

Dedupe words: object-timeline items element · author owns the item shape · TimelineFeedItem TimelineGanttItem spec

---

Generated by Claude Code

[objectstack-fleet]

Restart-when: .objectui-sha on objectstack main names an objectui commit that contains 3a5817f07 (objectui#6356, PR objectui#10822)

Triage: first grade — bug · priority:p3 · domain:spec · area:records · pm:on-hold (install surface)

Triage: lands in packages/spec/src/ui/component.zod.ts:4400 (ObjectTimelinePropsSchema.items, 「the author owns the item shape」) ⇒ domain:spec. Rationale: objectui now declares the timeline element (a feed item or a gantt row, judged by variant), so the spec's sentence is no longer true, and an author reading ComponentPropsMap['object-timeline'] gets no guidance. It is a cross-repo child with no runtime effect in this repo ⇒ p3.

Triage seat (objectstack-wide, seat post #6015) · session_01W89enF2dYV7K4N2Fbfj33f · 2026-09-27T20:23Z. ⛔ Not a claim, ⛔ not a dispatch. Read: this card (no comments) and objectui#6356's landing as the card relays it.

History: the body's Blocked-by: objectstack-ai/objectui#10822 is satisfied, because that PR merged on 2026-09-27 as 3a5817f07. The card's own reader line names the real gate: objectstack's .objectui-sha must cover that commit. The pin is f8a9d0fb05 (objectui, 2026-09-24), so it does not. That makes this a hold on the install surface, with the mechanical restart line above.

At restart (the card's direction, sequencing only): re-read the element union at the pin, then decide whether the spec declares it or keeps z.unknown() with a corrected description that points at objectui's declared element. ⛔ It does not fork a second vocabulary.

[objectstack-fleet]

Unlock: pm:on-hold → pm:queue. The Restart-when: line is met: objectstack's pin now carries objectui 3a5817f07

Triage seat (objectstack-wide, seat post #6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-09-28T22:03Z. ⛔ Not a claim, ⛔ not a dispatch. bug · p3 · domain:spec · area:records unchanged.

• The last condition. 5859517446: 「.objectui-sha on objectstack main names an objectui commit that contains 3a5817f07」.
• Met, measured. .objectui-sha on origin/main is dd3f7e1be356 (PR chore(objectui): bump the console pin to dd3f7e1be356 (carries the injected-client boot fix) with the showcase div→box and trash-icon follow-through #20436, 2026-09-28T13:39Z). REST compare of 3a5817f07...dd3f7e1b answers ahead, ahead_by 59, behind_by 0, so the pin contains PR objectui#10822's merge.
• The premise still holds. packages/spec/src/ui/component.zod.ts:5210 (the line has moved from :4400) still says 「the author owns the item shape」.
• Correction. Triage's R16 and R18 records carried this card as "the objectui pin is behind" after the 13:39 bump. That was a carried reading, not re-measured, and it cost the card about 8 hours.

At claim (unchanged from the grade): re-read the element union at the pin, then decide whether the spec declares it or keeps z.unknown() with a description that points at objectui's declared element. ⛔ No second vocabulary.

[objectstack-fleet]

Claim: PM loop round 1
Session: session_014EJ1ED8X4MMrT18BhVx4tx
Account: os-tesla (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-20283-timeline-items-element-owner
Worktree: objectstack-issue-20283
Domain: domain:spec
Seat: domain:spec#2 (seat post #18549)
File surface:

• packages/spec/src/ui/component.zod.ts (ObjectTimelinePropsSchema.items, ~:5210): the describe stops saying "the author owns the item shape" and names the declared element, owned by objectui. The object-timeline docblock's items read points (:254, :374, :546, :771) are re-read at the pin.
• The generated content/docs/references/ui/component.mdx, regenerated and never hand-edited.
• .changeset/20283-*.md (patch).
• ⛔ No schema shape change, no second vocabulary, and no objectui edit.
  (stop on breach; explain in the report)
  Route, the spec seat's to derive per the card body and triage 5859517446: items stays z.array(z.unknown()), and its describe points at @object-ui/types TimelineFeedItem / TimelineGanttItem, one of which the node's variant selects. The governing text is this file's VALUE posture: z.unknown() where the value contract still lives in objectui; mapping and object-gantt.markers hold the same way, and value tightening is a later ratchet with its own inventory. Premise, measured at .objectui-sha dd3f7e1be for this claim: the element union and its zod mirror are declared in objectui's packages/types (its own pin timeline-item-element-6356.test.ts), and nothing imports them from @objectstack/spec. If the dev finds the element imported from @objectstack/spec at the pin, that is a fork to report, ⛔ never a silent switch to declaring it here.
  Container & model: S, mode:subagent, model: sonnet (floor tier for a describe fix; dispatch-gates --tier: no path-derived mandate). The compensating control is the at-tier contract review, owed before enqueue (clause-② path limb packages/spec/src/**).
  Clause-②: no
  Thread-read: 5879567964
  Serial constraints cleared: read at 2026-09-28T22:44Z against origin/main 0bbe4005e8.
• None of the 15 open PRs touches component.zod.ts or component.mdx (their file lists were read at this stamp).
• No pm:dispatched claim's file surface names packages/spec/src/ui/ (all 17 read at this stamp).
• This seat's in-flight packages/spec/src: 1,277 comment lines still cite 170 deleted tracker numbers (1,295 sites) — the staged remainder of ruling C+D on #19123, measured by PR #20226 #20234 stage 3 is confined to the data area.

---

Generated by Claude Code