[pm:seat] domain:spec @ objectui — 🟢 session_01E7snar5mwF7qoXJazqKhys · R5 closed 12:06Z · 7/7 merged · in flight 0 · queue 0 · standby

[os-sam]

Seat registration post for the domain:spec @ objectstack-ai/objectui lane — the repo-split half created by the maintainer's 2026-08-23 roster ruling.

Body is authoritative; title and assignee are derived views. Single writer: the sitting seat PM. Comments are audit only and never carry state. Position description is versioned in references/lanes/spec.md; this post carries state only.

1. Current PM

• Session: session_01E7snar5mwF7qoXJazqKhys (GitHub identity os-sam)
• R5 closed 2026-08-23T12:06Z. R6 standby patrol ran 13:15Z — nothing moved, no dispatch. In flight 0 · queue 0 dispatchable · decision box 2 (Carrier for Phase 2 of the row-predicate deprecation: the stored-metadata survey that must size the removal window before any spelling retires #5741, objectstack#11284) · awaiting a human merge 0 · rework 0 · red CI merged 0 · re-runs issued 0.
• Posture: STANDBY. Patrol armed 14:20Z.
• R6 produced no work but discharged one standing debt: the merge-timing audit (comment 5386204916). See §4.

2. Landing ledger

R1–R5 — 7 cards / 7 PRs / 7 merged / 0 rework / 0 red CI merged / 0 re-runs / 1 transferred out / 2 to the decision box.

round	card	PR	merged	gates
R1	#5330 row-predicate canon	#5737	d1ab06f	19+3 full green
R1	#5667 CalendarViewSchema converge	#5739	065bba7	18+3+1 (Build Docs real failure, inherited)
R2	#5738 Phase 0 docs sweep	#5758	702c48a	17+3 (20-gate docs run)
R2	#2890 ObjectView drift guard	#5760	af3861f	19+3 full green
R3	#5740 retire 'agenda' from CalendarViewMode	#5782	b55a346	18+3+1 (Build Docs real failure, inherited)
R4	#5494 ReferenceRailEntry derive + retire icon	#5794	09:27:58Z	19+3 — merged BEFORE 2 shards reported; see §4
R5	#5784 retire 'agenda' from defaultView ×3 faces	#5806	12:05:47Z	19+3 full farm read by name BEFORE arming

Deliverables verified on main, not from PR state: CalendarViewMode = 'month'\|'week'\|'day', the zod enum matching, object-view-spec-parity.test.ts present, PER_VIEW_CONFIG_ALIASES in the normalizer, rowPredicateCanon.ts referenced from the evaluator barrel.

Transferred out: #5435 → objectstack#11195 (open, pm:queue, still unassigned/unclaimed at 13:16Z — the objectstack spec lane has not picked it up in ~11 h).
Findings filed by this lane's devs: #5738 ✅ · #5740 ✅ · #5741 · #5759 · #5783 · #5784 ✅ · #5792 · #5793 · objectstack#11195 · objectstack#11284.

3. Hot-file serial queue

EMPTY — packages/types released by #5806; packages/plugin-detail by #5794.

⛔ The old "blocked twice" constraint is RETIRED — #5716 closed 04:57Z via PR #5752, so #5435 is single-blocked on objectstack#11195 alone. Not over-read: nobody has measured that a pin refresh succeeds end-to-end.

4. Platform facts measured by this seat

• ⛔ Test (shard n/4) IS NOT A REQUIRED CHECK, AND A DIRECT MERGE IS REFUSED. Two halves, both measured: (1) on fix(plugin-detail): derive ReferenceRailEntry from @objectstack/spec and retire the icon key #5794 auto-merge fired at 09:27:58Z with shards 2/4 and 4/4 still running; (2) merge_pull_request is rejected outright — 405 … Changes must be made through the merge queue — so auto-merge IS the enqueue mechanism. The discipline is therefore when you arm it: read the full farm by name, then arm. "Green = 19 success + 3 skipped = 22" describes the FARM, not the MERGE GATE. Never tell a card auto-merge is holding for a check that is not required.
• ⛔ THE MERGE-TIMING AUDIT IS DONE, AND ITS RESULT IS A NEAR MISS, NOT AN ALL-CLEAR (comment 5386204916, R6). All five R1–R3 PRs merged after their four shards reported, so fix(plugin-detail): derive ReferenceRailEntry from @objectstack/spec and retire the icon key #5794 is the only PR that landed on incomplete shard evidence and nothing on main needs revisiting. But feat(core,react): declare record.* the row-predicate canon and warn on the two deprecated spellings #5737 cleared by 45 s and Add the ObjectViewSchema drift guard, and fold the four per-view-type config aliases #5760 by 35 s — margins produced by merge-queue latency exceeding the test-farm tail, not by any decision this seat made. Arming timestamps for those five were not recoverable, so "the practice was identical in all six and only the timing differed" is inference, not measurement — flagged as such. Do not read the audit as licence to arm early.
• ⛔ A GREEN WHOSE DURATION IS IMPLAUSIBLE FOR THE WORK IS NOT EVIDENCE — AND THIS IS NOT SPECIFIC TO Build Docs. Mis-diagnosed feat(types,plugin-calendar): converge CalendarViewSchema on the calendar-view renderer's read set #5739's Build Docs failure because a sibling "passed" — an 11-second cache hit against a real 3m50s build. Reproduced on fix(plugin-detail): derive ReferenceRailEntry from @objectstack/spec and retire the icon key #5794/feat(types): retire 'agenda' from defaultView on all three declaration faces #5806 (10 s each). Test shards cache too: on docs(plugin-form): Phase 0 of the row-predicate deprecation — stop teaching the bare shorthand (#5738) #5758 (docs-only) all four completed in 11–13 s. Legitimate there, but it means any job's green can be a cache hit. Compare durations before using check X on PR A as a control for X on PR B. Contrast: Type Check at 9m21s on feat(types): retire 'agenda' from defaultView on all three declaration faces #5806 is a plausible real run.
• ⛔ Build Docs is red on main (Build Docs is red on main: fs reaches the browser bundle via @objectstack/spec → pg-connection-string #5668) and does NOT gate the queue. fs via pg-connection-string ← @objectstack/spec ← @objectstack/formula ← packages/core/dist/evaluator/fieldRules.js ← packages/components ← InteractiveDemo.tsx. Real (uncached) failures measured on feat(types,plugin-calendar): converge CalendarViewSchema on the calendar-view renderer's read set #5739 (3m52s) and feat(types): retire 'agenda' from CalendarViewMode and the zod enum #5782 (3m43s) — both merged anyway. The MDX route in the trace varies between runs, so the failure is in the shared client graph, not any doc page. Exclude with cause; never call such a PR "green".
• ⛔ check_suite.completed IS NOT A GATE READING. It fired on feat(types): retire 'agenda' from defaultView on all three declaration faces #5806 at 12:04:56Z; the farm was read by name anyway. GitHub's own notice excludes cancelled suites, suites with no runs, and legacy statuses. Read jobs BY NAME.
• ⛔ YOUR OWN CHECKOUT GOES STALE, AND A GREP ON A STALE TREE IS NOT EVIDENCE. During feat(types): retire 'agenda' from defaultView on all three declaration faces #5806's review the shared checkout sat at 9a5d669c1, pre-feat(types): retire 'agenda' from CalendarViewMode and the zod enum #5782 — a "which faces still carry 'agenda'" grep on it would have been worthless. Re-grep at the PR's actual base sha.
• ⛔ A @ts-expect-error IN A FILE NO tsc READS IS A NO-OP THAT LOOKS LIKE A TEST. feat(types): retire 'agenda' from defaultView on all three declaration faces #5806's TS-face pin is real only because packages/types's type-check chains tsconfig.test.json — and an unused directive is itself an error, so a green type-check proves the rejection. Verify the tsconfig chain before crediting such a pin.
• ⛔ search_issues can silently return zero, control words included (objectstack#11193). Dedupe here uses two independent methods.
• ⛔ A RULING AGES IN BOTH DIRECTIONS, AND BOTH ARE WORTH REPORTING. ReferenceRailEntry is owned by @objectstack/spec as of 17.1.0 and objectui's local copy declares an extra icon key the spec's strict schema refuses #5494's central premise ("the renderer reads icon") measured false, and false at filing — the retirement was cheaper than priced. finding(types/plugin-calendar): defaultView declares 'agenda' in ObjectCalendarSchema and both zod twins, but ObjectCalendar reads it as 'month' | 'week' | 'day' #5784's measured stronger than briefed: @objectstack/spec's ObjectCalendarProps.defaultView is already three-valued, so objectui was the outlier and Option A closed a spec-alignment gap (commandment #0), not merely an ADR-0049 narrowing.
• ⛔ A RETIREMENT CARD SHOULD BE CHECKED FOR A LIVE PRODUCER, NOT JUST A CONSUMER. ReferenceRailEntry is owned by @objectstack/spec as of 17.1.0 and objectui's local copy declares an extra icon key the spec's strict schema refuses #5494 was written as "retire a key from a type"; objectui's own buildDefaultPageSchema was emitting that key into every synthesized rail page. Carried into finding(types/plugin-calendar): defaultView declares 'agenda' in ObjectCalendarSchema and both zod twins, but ObjectCalendar reads it as 'month' | 'week' | 'day' #5784's brief as a declared fork condition; there the two forwarders proved genuine pass-throughs. Ask "who writes this key?" every time.
• ⛔ A TITLE IS NOT A STATE. [Decision] Row-predicate spelling contract: does the canon narrow to record.*, or do the bare shorthand and data.* spellings stay? (the uncarried "#4075 step 3" question) #5330 read [Decision] in pm:queue and was not a mislabel.
• ⛔ TWO WAYS A GUARD GOES VACUOUS (both on Migrate the remaining ListView legacy vocabulary to spec-canonical keys, and audit ObjectView/DetailView (#2231 phases 4–5) #2890): the keyof exhaustiveness idiom collapses to never under an index signature — deleting a key left type-check green; and a loop-unwrap() walks z.array(z.string()) down to string in zod 4.
• ⛔ An ADR-0087 D2 tombstone stays IN the spec's .shape — a "is the key present" parity check reports a RETIRED key as live and passes.
• ⛔ A KILLED SUBAGENT IS NOT A DEAD CLAIM WHEN IT HAS PUSHED COMMITS. Both R3 devs died to the 5-hour usage wall (~05:50Z, reset 07:00Z); neither withdrawn (finding(types/plugin-calendar): view: 'agenda' is declared by CalendarViewMode and the zod enum but the calendar-view renderer silently falls back to 'month' #5740 had pushed 1f94b4c9); both resumed via SendMessage and completed. Check git ls-remote --heads origin 'refs/heads/claude/issue-*' before choosing resume vs withdraw.
• check:spec-symbols runs inside the Type Check job (ci.yml:238), under no name of its own. Its DEBT ledger is ratcheted (check-spec-symbol-derivation.mjs:1066). Sanctioned forms: export type { X } from '@objectstack/spec/…' and export type X = z.infer<typeof SpecX> (:907-926).
• Console eager-closure budget is not near its ceiling. 3918.3/3990.2 KB on fix(plugin-detail): derive ReferenceRailEntry from @objectstack/spec and retire the icon key #5794 (98.2%) but 3226.6/3990.2 KB on feat(types): retire 'agenda' from defaultView on all three declaration faces #5806 (80.9%) two hours later — ~692 KB came off main in between. A single budget reading is a snapshot, not a trend.
• Inherited priors from [pm:seat] domain:ui — objectui execution seat #5560 §5 holding across seven dispatches: no needs:contract-review in this repo (at-tier dispatch, 「直接入队」); dispatch-gates.mjs is the wrong tree; the sanitizer eats bare angle-bracketed tokens. Governed surfaces: AGENTS.md, CLAUDE.md, .claude/**, docs/adr/** — no PR touched one.

5. Next (R7) — STANDBY, queue empty

Dispatchable — 0. Re-verified at 13:16Z; lane holds 13 cards:

card	state	why not dispatchable
#2731	pm:queue	Design-first by its own header — an ADR gates it, every foundational step lands in @objectstack/spec + framework. objectui is the last step and a consumer. Belongs to the objectstack spec lane. Not relabelled — another author's card, and the label is not wrong, just not this seat's.
#4535	pm:queue	Assigned to yinlianghui (a human). Taken; never reassign.
#2890 · #5435 · #5362 · #2231 · #4356	pm:blocked	#2890 → objectstack#11284 · #5435 → objectstack#11195 · others carry their own blockers.
#5213 · #4631 · #4052 · #3834	pm:on-hold	Not dispatchable by definition.
#5741	needs-user-decision	Decision box, below.
#4043	tracking	Tracking card; its asked-for CI rule is the gate #5494 exercised.

Standby duties, in order:

1. Unlock sweep — objectstack#11195 (releases normalizeListViewSchema manufactures userActions.group / hideFields / rowColor — three keys ListViewSchema refuses BY NAME, so the fold's own output fails the view save gate #5435; unassigned/unclaimed ~11 h) · objectstack#11284 (releases Migrate the remaining ListView legacy vocabulary to spec-canonical keys, and audit ObjectView/DetailView (#2231 phases 4–5) #2890).
2. Triage grading watch — finding(i18n): calendar.agenda is a confirmed-dead key in all ten locale packs — no static reader, no dynamic calendar. head, and the enum value it labeled is retired #5783 (dead i18n calendar.agenda residue in all 10 locales, zero consumers; independently re-confirmed by finding(types/plugin-calendar): defaultView declares 'agenda' in ObjectCalendarSchema and both zod twins, but ObjectCalendar reads it as 'month' | 'week' | 'day' #5784's dev; still ungraded and unaddressed — the most likely next dispatch if it lands in this lane) · [finding] registry-inputs-spec-parity narrates the ReferenceRailEntry icon divergence as unsettled after #5494 resolved it #5792 (stale-by-landing — cites ReferenceRailEntry is owned by @objectstack/spec as of 17.1.0 and objectui's local copy declares an extra icon key the spec's strict schema refuses #5494 as unsettled; ReferenceRailEntry is owned by @objectstack/spec as of 17.1.0 and objectui's local copy declares an extra icon key the spec's strict schema refuses #5494 landed) · Packages whose published d.ts references spec-17.1.0-only symbols still declare "@objectstack/spec": "^17.0.0" — the range admits a spec that lacks the symbols #5793.
3. Re-read Carrier for Phase 2 of the row-predicate deprecation: the stored-metadata survey that must size the removal window before any spelling retires #5741 for the maintainer's A/B/C.
4. Merge-timing audit — DONE in R6, comment 5386204916. See §4.

⛔ Decision box — two cards, both awaiting the maintainer

#5741 — the stored-metadata channel. No new comment since this seat's ACCEPT at 08:06Z. Phase 2 of the row-predicate deprecation is gated on a survey, and no seat holds a channel to stored metadata. Four-facet block on the card (A: per-tenant REST token — ⛔ insufficient alone, it sizes one tenant and calls it the fleet; B: read-only replica/dump; C: offline export via Setup's existing read-only Export). Recommended C, fallback B.

⚠️ Three cards in a row have now hit this wall. objectstack#7917 hit it 2026-08-13 with ready-to-run queries; ten days on, no channel exists. #5494 and #5784 both had to ship on "no reachable usage" rather than "no usage" — the honest phrasing, and the standard for this lane, but it means every retirement here ships on an incomplete census. #7917 could proceed because its transform was mechanically lossless; the row-predicate case has no such escape hatch — data.* binds nothing at runtime, so rewriting a stored predicate needs authorial intent a machine cannot infer.

objectstack#11284 — the react-blocks.ts vocabulary question (releases #2890). Triage backfilled the four-facet block at 09:38Z and recommends 收敛 (converge) with deprecate-first discipline, viewType→type in the same stroke. ⚠️ Unchanged at 13:16Z — still needs-user-decision, maintainer has NOT answered. A recommendation is not a ruling; #2890 stays blocked.

6. Notes

• Five dev corrections to this seat's briefs, all accepted publicly, plus three self-corrections: finding(types/plugin-calendar): CalendarViewSchema declares 9 of its 13 keys that the calendar-view renderer never reads — including events, its only required key, which the renderer deliberately drops #5667 described as narrowing when .passthrough() meant it widened; ReferenceRailEntry is owned by @objectstack/spec as of 17.1.0 and objectui's local copy declares an extra icon key the spec's strict schema refuses #5494's ACCEPT claiming auto-merge was holding for non-required checks; and the R4→R5 ~58-minute idle gap where a graded card sat undispatched. Plus one over-stated characterisation walked back to the user (the eager-closure budget called "a standing condition" when it moved 692 KB in two hours). A dispatch is a hypothesis; measurement outranks it — including measurement of the merge machinery and of this seat's own cadence.
• Two cards landed by ruling INHERITANCE under the sibling-branch meta-criterion: finding(types/plugin-calendar): view: 'agenda' is declared by CalendarViewMode and the zod enum but the calendar-view renderer silently falls back to 'month' #5740 from finding(types/plugin-calendar): CalendarViewSchema declares 9 of its 13 keys that the calendar-view renderer never reads — including events, its only required key, which the renderer deliberately drops #5667, finding(types/plugin-calendar): defaultView declares 'agenda' in ObjectCalendarSchema and both zod twins, but ObjectCalendar reads it as 'month' | 'week' | 'day' #5784 from finding(types/plugin-calendar): view: 'agenda' is declared by CalendarViewMode and the zod enum but the calendar-view renderer silently falls back to 'month' #5740. Both recorded as a veto window, not a permission gate; both re-measured before dispatch. Still overturnable.
• Cross-repo posture: files spec-side contract cards into objectstack's domain:spec queue, flipping its own card to pm:blocked + Blocked-by: in the same stroke. Does not scan domain:ui ([pm:seat] domain:ui — objectui execution seat #5560) or domain:devx (objectstack#10602).

[os-sam]

Round-open marker — R2, 2026-08-23T04:4xZ, session session_01E7snar5mwF7qoXJazqKhys. Audit only; §1 of the body still reads "R1 closed / standby" and is refreshed at the R2 boundary, per the write-side rule. Mid-round state is carried by the Claim: comments on the cards, not here.

Why the seat came out of standby: the 04:38Z standby sweep found the lane refilled. The maintainer directed a bulk triage of objectui's unlabelled backlog this session (verbatim, recorded on #5741: 「objectui 有 79 张卡在车道视图之外（占该仓 35%），你帮我分批处理」·「继续 批 2 批 3」), and it routed six cards into domain:spec — including all three R1 follow-ups.

Unlock sweep: no unlock. objectstack#11195 is still open (pm:queue, unassigned), so #5435 stays pm:blocked. Reminder for the next sweep: it is blocked twice — #11195 landing and the spec pin bump that #5716 holds at 17.1.0. Releasing one is not enough.

In flight (2, batch capped at 3, both mode:subagent, both clause-② no):

• Migrate the remaining ListView legacy vocabulary to spec-canonical keys, and audit ObjectView/DetailView (#2231 phases 4–5) #2890 — opus, branch claude/issue-2890-objectview-drift-guard. pm:blocking, so first by the lane's selection order. Scope is the triage-limited one (ObjectViewSchema drift guard + four phase-3 alias folds), not the card body's much larger project; A6 / viewType / conditionalFormatting / exportOptions are excluded as upstream packages/spec faces.
• Row-predicate deprecation is missing its Phase 0: docs and example apps may still teach the bare shorthand / data.* spellings the new warning now flags #5738 — opus, branch claude/issue-5738-row-predicate-phase0. Phase 0 of the row-predicate deprecation, the sibling of R1's [Decision] Row-predicate spelling contract: does the canon narrow to record.*, or do the bare shorthand and data.* spellings stay? (the uncarried "#4075 step 3" question) #5330.

Hot-file serial queue (R2):

surface	held by
packages/types/src/__tests__/ · packages/core/src/utils/normalize-list-view.ts	#2890
content/docs/** · example apps · schema-catalog · plugin READMEs	#5738

Serialised behind #2890 — packages/types is one package, and batch independence is per-package, not per-file: #5740 (view: 'agenda' residue) and #5494 (ReferenceRailEntry extra icon key). Both dispatchable the moment #2890's PR merges. fold-or-serial answered: serial — #5740 and #2890 fail gate ① of the five (different defect shapes: a value-level enum residue vs a declaration-drift guard), so they do not fold.

Also queued, not yet dispatched: #5741 (Phase 2 survey carrier). Triage graded it pm:queue with the note that establishing the data source is the first deliverable, and returning pm:blocked with the access channel named is the correct outcome if none exists — ⛔ explicitly not "guess a population and call it sized". That resolves the routing question this seat raised at R1 close; it does not need the maintainer.

#2890 assignee provenance, checked before claiming: cleared 02:36Z by the skills seat under maintainer ruling objectstack#11196 as assignee-field residue from the 07-28 filing session (shared agent identity, no Claim: comment, no branch). Not a live claim; claimable by protocol.

---

Generated by Claude Code

[os-sam]

⛔ CORRECTION to §3's standing lane constraint — the pin blocker has lifted. Audit comment; the body still carries the stale text and is corrected at the R2 boundary.

§3 currently reads: "any card whose fix is 'align objectui on the spec pin' is blocked twice — on the spec-side change landing in objectstack, and on the pin bump, which #5716 holds at 17.1.0."

The second half is no longer true. #5716 closed completed 04:57:58Z via merged PR #5752 (c9327c9), taking the Localize branch — objectui now owns Theme / ThemeMode / ColorPalette rather than re-exporting them from the retired @objectstack/spec/ui surface. #5716's own body stated the consequence it was gating: "any objectui spec-pin refresh past the retirement breaks @object-ui/types and @object-ui/providers builds." That is what localization removes.

Consequence for this lane: #5435 is now single-blocked, on objectstack#11195 alone (still open, pm:queue, unassigned). When #11195 lands, the next unlock sweep should find only one condition to satisfy, not two.

⚠️ Do not over-read this. What is established is that #5716's blocker is discharged — by its own closure and merged PR. I have not measured that a pin refresh now succeeds end-to-end, and the manifests still declare the range ^17.0.0, which 17.1.0 already satisfies. Whether a bump past 17.x is clean is a separate, unmeasured question, and #5716's body names related guards designed to trip at exactly that moment (the ThemeModeSchema deny-list row in spec-ui-schema-reexports.test.ts). The unlock sweep re-verifies on the merged ref rather than trusting this note.

R2 landing so far: #5738 → PR #5758 merged 702c48a (17 success + 3 skipped, zero red, zero pending). #2890 still in flight.

---

Generated by Claude Code

[os-sam]

R3 interrupted by the 5-hour usage wall, and resumed. Audit comment; §1 of the body still reads "R2 closed / ready for R3" and is refreshed at the R3 boundary.

What happened. Both R3 devs — #5740 and #5741, dispatched ~05:45Z — were terminated mid-task at ~05:50Z by an account-level session limit, with the reset time carried in the failure itself: 07:00Z. Per the seat's own discipline, the authoritative wall signal is the failure report, and the reset moment is available at that instant and not afterwards, so it is recorded here.

Neither card was withdrawn, and that was the deliberate call:

• finding(types/plugin-calendar): view: 'agenda' is declared by CalendarViewMode and the zod enum but the calendar-view renderer silently falls back to 'month' #5740 — its dev had pushed a branch before dying: claude/issue-5740-calendar-view-mode-agenda at 1f94b4c9, verified on the remote. The reclaim rule is explicit that a claim carrying a live branch with commits is never reclaimed, so withdrawing would have discarded real work.
• Carrier for Phase 2 of the row-predicate deprecation: the stored-metadata survey that must size the removal window before any spelling retires #5741 — no branch, but the interrupted-dev protocol is resume-first and withdraw only if resumption is unavailable.

Both were resumed via SendMessage at ~07:55Z, after the wall lifted. The original briefs stand; each resume message re-stated only the load-bearing constraints rather than re-issuing the brief.

A finding salvaged from #5740's dying words, verified before passing it back. Its last output was "More surface than the card cites — objectql.ts, i18n locales." Spot-checked on origin/main: CalendarViewMode appears in packages/types/src/complex.ts, packages/types/src/index.ts, packages/types/src/zod/complex.zod.ts and packages/types/src/zod/index.zod.ts — re-exported through both barrels, wider than the card's two cited files. The objectql.ts and i18n-locale legs are not confirmed and were handed back as the dev's to establish.

This matters because it feeds the card's load-bearing question directly: if CalendarViewMode has consumers beyond CalendarViewSchema.view, narrowing the shared enum is too wide a cut, and the correct fix may be narrowing CalendarViewSchema.view alone. That was already Zone 2 item 3 of the dispatch; the dev's own measurement is now evidence for it rather than a hypothetical.

Standing note for the next wall: a killed subagent is not a dead claim when it has pushed commits. Check git ls-remote --heads origin 'refs/heads/claude/issue-<n>*' before deciding between resume and withdraw — the two paths differ by whether work exists to lose.

---

Generated by Claude Code