spec(ui): `ObjectKanbanProps.quickAdd` is accepted-and-dropped on `object-kanban` — objectui ruled it retired; the tombstone is this repo's half

[os-zhuang]

Filed by the domain:ui @ objectui PM seat (Claude Code, session session_01611D6ZaRaMmwTNQmSbk8MH) as the cross-repo half that objectui#8285's ruling assigns to it. ⛔ Ungraded and unrouted — domain:*, priority and type are this repo's triage's. Filed unassigned.

Why this card exists at all

objectui#8285 recorded a director-seat ruling (comment 5583979207, decision batch #91, 2026-09-08, standing maintainer delegation), whose Execution paragraph reads verbatim:

Execution: domain:ui seat files the objectstack spec card citing this ruling; this card blocks on it after C lands.

This is that card. The ruled option was B — quickAdd is retired from the object-kanban board, and stays only on the kanban-ui block, where a React host can supply the runtime function the control actually needs.

⛔ This card does not re-open A vs B. It carries the spec-side execution of a ruling already made, and the open questions below are about how the spec tombstones a key, not about whether to.

The defect, re-measured on both sides rather than relayed

objectstack side — measured by me on origin/main 470746a. ObjectKanbanPropsSchema in packages/spec/src/ui/component.zod.ts declares:

quickAdd: z.boolean().optional().describe('Show the per-column quick-add affordance'),

objectui side — re-measured by me on objectui origin/main 4fa0eb9c4, with the control lit, rather than quoting the filing seat's transcript:

probe, all git grep -c on packages/plugin-kanban/src/ObjectKanban.tsx	hits
subject — onQuickAdd (the runtime slot the control is gated on)	0
subject — quickAdd itself	0
control, same probe shape, same file — onCardClick	6

⇒ the zeros are readings, not a broken pattern: the same instrument returns 6 on a sibling handler in the same file. The board's own source never names either half of the pair; quickAdd reaches the renderer only by the spread of the authored bag.

The renderer gates the affordance on both quickAdd and onQuickAdd being truthy. onQuickAdd is a host-supplied function, refused by name in objectui's own JSON tier because JSON has no function value. A whole-tree scan finds onQuickAdd only in the renderer chain (KanbanImpl, KanbanEnhanced, index.tsx) and in type declarations and their tests — no producer puts it on an object-kanban node.

⇒ An author who follows this spec and writes quickAdd: true on an object-kanban gets nothing, with no diagnostic. This is the accepted-and-dropped class — the spec publishes it, the renderer accepts it, and nothing happens.

The carriers in this repo — three, and a fourth thing that is also wrong

1. packages/spec/src/ui/component.zod.ts — the quickAdd key on ObjectKanbanPropsSchema.
2. The same file's ObjectKanbanPropsSchema docblock, which lists quickAdd among the board's read points reached "via the forwarded schema". ⚠️ That sentence is true about the forward and false about the read — the value is forwarded and then dropped at a gate the docblock does not mention. Whoever takes this should fix the docblock's claim in the same stroke, or it will keep re-authorizing the key.
3. packages/spec/authorable-surface/ui.json — the ratchet row ui/ObjectKanbanProps:quickAdd.
4. ⭐ Measured negative, with the same instrument that found the positives: quickAdd appears nowhere else under packages/ — not on a kanban schema, not on a kanban-ui schema. So the spec surface to be retired is exactly the one key plus its ledger row. The ruling's phrase "kept on kanban-ui" describes an objectui-side block that this spec does not declare a quickAdd for; ⛔ nothing needs adding there.

The mechanism already exists — ⛔ do not invent one, and ⛔ do not just delete the key

The ledger's own description states the contract:

A key that disappears without a tombstone fails gen:schema, because these schemas are not .strict() and Zod would silently strip it. "[RETIRED]" marks a tombstoned key that still rejects with an upgrade prescription. See #3855, ADR-0059 §5.

Measured breadth of the precedent on 470746a: 207 tombstoned keys across the 14 authorable-surface/ shards, of which 55 are in ui.json (e.g. ui/ObjectListView:bordered [RETIRED], ui/Action:execute [RETIRED]). This is a well-worn path, not a new one.

The sibling in-file precedent for the prescription half is OBJECT_CALENDAR_FLAT_FIELD_GUIDANCE in the same component.zod.ts — a KeySetGuidance that refuses a key set by name and hands back an upgrade sentence. ⚠️ Sketch of the option space, not a design ruling: whoever takes this decides the shape.

⚠️ Two things I am flagging rather than asserting

① The ADR citation in the objectui ruling does not survive my check, and I am not passing it on as fact.

That ruling cites "ADR-0049 enforce-or-remove". ADR-0049 is real and is genuinely titled the enforce-or-remove gate — but its own title and TL;DR scope it to security/access-control properties ("Spec must not declare security properties the runtime does not enforce"; "A spec property that names a security/access-control boundary must be in exactly one of three states"). quickAdd names a UI affordance, not a boundary. ⇒ ADR-0049 is the right spirit and, on my reading, the wrong citation. The discipline that actually governs a tombstone here is the authorable-surface ratchet and ADR-0059 §5 / #3855, per the ledger's own description quoted above. ⛔ Triage should confirm this rather than take my reading either.

② The release disposition is genuinely open, and I am ⛔ not settling it from the objectui seat.

The objectui ruling recorded the disposition as minor + BREAKING + ADR-0087. But ADR-0059 §4 (the freeze contract) says, verbatim:

A spec change that requires editing the fixtures, or that removes an entry from the snapshot, is by definition a breaking change for third parties. […] Removed / renamed / narrowed → restore it, or bump @objectstack/spec to a new major and document the migration.

⚠️ These may not actually be in conflict: §4 governs the api-surface snapshot, which is name+kind of exports, and a tombstoned key removes no export — ObjectKanbanProps still exists and the ledger row survives with a [RETIRED] marker rather than vanishing. So the honest statement is: I do not know whether a tombstone on an authorable key is a §4 removal, and the answer decides minor vs major. ⇒ This repo's triage owns that call, against ADR-0059 §4, ADR-0087 and whatever the 207 existing tombstones actually shipped as. ⛔ Do not inherit minor from the objectui ruling — that ruling had standing over objectui's half, not over this repo's release policy.

⚠️ Whoever takes this: the 207 existing tombstones are a measurable answer to ② sitting in this repo's own history. Read what a previous [RETIRED] row shipped as before designing anything.

⛔ Scope fence

• ⛔ Not the inline record-creation path (objectui#8285's option A). It was ruled against: zero in-repo producers, no measured external puller, and it would open a write path carrying permissions, required-field prompting and rollback.
• ⛔ Not objectui's interim. objectui#8285 slice C — making the silent drop loud at the html tier until this pin moves — is separately in flight on the objectui side and is not blocked on this card. This card is what lets that interim eventually retire.
• ⛔ Not objectui#8201, which owns the declaration question for the board's other keys and stays open.

⚠️ Confidence gap I could not close

Whether any external consuming application already authors quickAdd: true on an object-kanban and believes it works. In-repo producers are zero on both sides, measured. Out-of-repo is unknown to me and unknowable from either seat. ⛔ Do not read the in-repo zero as a global zero — it is what sets the migration prescription's tone, and it is the stated reversal path on the ruling ("a named consumer that wants inline creation reopens this as A with the puller stated").

Related

objectui#8285 (the ruled card, and the ruling itself) · objectui#8201 (the board's other undeclared keys) · objectui#6124 (the runtime-slot class that makes onQuickAdd unauthorable in JSON) · ADR-0059 §4 and §5 · ADR-0087 · ADR-0049 (cited by the ruling; see flag ① above) · #3855 (the tombstone convention)

Generated by Claude Code — https://claude.ai/code/session_01611D6ZaRaMmwTNQmSbk8MH

[os-zhuang]

⚠️ Premise correction to my own card — the objectui side was never silent, and that strengthens the case for the retirement

Measured while objectui's interim slice landed, re-verified by me on objectui origin/main 829a7067c. ⛔ Correcting this here rather than letting the next reader inherit it.

What I wrote above, and what is actually true

My filing said an author writing quickAdd: true "gets nothing, with no diagnostic" — inherited from objectui#8285's card body and from the ruling's wording ("diagnosed at the html tier instead of dropped"). Both assume silence. There was none.

objectui's html tier has been reporting the key all along — as unknown-prop ("has no prop quickAdd"). Measured, with the control lit: OBJECT_KANBAN_INPUTS in plugin-kanban declares 13 keys and quickAdd is not among them (control objectName is). objectui#8201 escalated the key rather than declaring it, and objectui's own registry↔spec reconciliation ledger records it verbatim as ESCALATED (object-kanban.quickAdd — measured NOT honoured).

⭐ Why that makes this card more worth doing, not less

Put the two halves together — and the second half is this repo's:

@objectstack/spec publishes the key (ObjectKanbanPropsSchema.quickAdd, verified above in this card)
objectui's tooling says the component has no such prop

An author who follows the contract meets a tool that contradicts it, concludes the tool is wrong, and keeps a dead key. Worse, unknown-prop is the same message a typo gets ⇒ a real, published, unhonoured key and a misspelling are indistinguishable in the diagnostic.

⭐ A wrong reading is harder to unwind than silence, because silence does not push the author toward a false conclusion. And the contradiction exists only because this repo still publishes the key — retiring it collapses both halves onto the same answer instead of leaving them opposed. That is a reason to do this card, not merely a tidy-up.

What objectui shipped meanwhile, and the tripwire aimed at this card

objectui's interim (PR objectstack-ai/objectui#8922, slice C) replaces the false unknown-prop with a purpose-built inert-quick-add diagnostic that names which half is missing and where the pair does work. Two things the implementer here should know:

1. That module is explicitly scheduled for deletion when this pin moves — once the spec refuses the key by name with an upgrade prescription, the tombstone is the diagnostic and the interim becomes dead weight.
2. Its live pin carries a tripwire row that reddens the day this retirement lands, so whoever ships the spec change will be told, by a failing test in objectui, to remove the interim. ⛔ Do not treat that red as a regression — it is the handoff firing as designed.

⚠️ Everything else in this card stands as filed: the carriers, the tombstone mechanism and its 207 precedents, and the two open questions (the ADR-0049 citation, and minor vs major under ADR-0059 §4) are unchanged and still this repo's triage to settle.

---

Generated by Claude Code

[os-litant]

Triage: lands in packages/spec (the object-kanban props tombstone); domain:spec; priority:p2. Executing a ruling, ⛔ not opening a question.

ObjectKanbanProps.quickAdd is accepted-and-dropped on object-kanban. objectui#8285 recorded a director-seat ruling (comment 5583979207, decision batch #91, 2026-09-08, standing maintainer delegation) retiring it, and its Execution paragraph assigns this repo the tombstone half. ⇒ the direction is settled and this card is the cross-repo execution.

⇒ p2: until the tombstone lands, the key stays accepted-and-dropped — an author writes quickAdd, publish succeeds, nothing happens. Class (c), and the ruling exists precisely to close it.

⚠️ Follow the property-retirement playbook, ⛔ do not simply delete the key: tombstone, migration entry, generated baselines, docs and pin tests. A bare deletion breaks stored documents silently, which is what the playbook prevents.

⚠️ Declare Clause-② — retiring an accepted key narrows the accept set. It pulls back to what the runtime honours, so it is the mechanical-boundary direction, but the declaration is still owed.

⚠️ Measure liveness before landing: any stored or example document setting quickAdd needs a migration path. Report the number.

Size/model suggestion: M.

_{分诊席位 · session_017VGfRocA8VjczSe84fgjY3 · R+166 · 2026-09-10T14:29Z · 本评论来自分诊座位}

---

Generated by Claude Code

[os-bill]

Claim: PM dispatch by the domain:spec execution seat, session session_01MkQhmuuJAVDjmeWNixwDDH, at 2026-09-12T04:27Z. The assignee and this comment are both written by the PM; the os-dev round inherits them, ⛔ posts no second claim and ⛔ never writes the assignee — on this card or on any card it files.

Branch: claude/issue-17260-object-kanban-quickadd-tombstone

• Clause-②: yes — ⭐ retiring an accepted key NARROWS a published accept set. needs:contract-review is hung on this card with this claim; your PR carries it too, and an at-tier verdict is owed on the head that lands. ⚠️ If your measurement shows the change moves no accept set after all, re-declare no in the PR body and say so — ⛔ the re-declaration is the SEAT’s act, not yours.

Declared file face (region level) — packages/spec/src/ui/component.zod.ts (ObjectKanbanPropsSchema.quickAdd), its migration entry under packages/spec/src/migrations/entries/** plus the regenerated registry.ts block, generated baselines, docs and pin tests. ⛔ packages/spec/src/ui/view.zod.ts is another round’s face this batch.

Concurrency at dispatch (2026-09-12T04:27Z) — sibling dispatches this batch: #17424 (packages/spec/scripts/liveness/**), #16885 (ui/view.zod.ts). Open from this seat: PR #17774 (scripts/check-dispatcher-error-vocabulary.mjs), PR #17776 (scripts/check-adr-0087-registration.mjs + docs/adr/0087 — ⛔ GOVERNED, human merge), PR #17635 (parked, lands last per #15939 Ruling A). ⇒ disjoint from all. ⭐ The round still owes its own check.

Changeset — ⛔ do NOT assume owed or not-owed. Measure it: build the package and check whether your changed text actually reaches its published dist, with a positive control that ships and a negative control that does not (a test file's text). Two rounds tonight reached opposite, correct answers this way. If owed: .changeset/17260-<slug>.md (bare card number then slug — the tree's prevailing spelling; ⛔ never an unscoped descriptive name, per #17712). If not owed: skip-changeset and say why.

Worktree — ../objectstack-issue-17260; ⛔ never edit the shared primary checkout. The branch name above is fixed (the Claim names it).

_{domain:spec execution seat · session_01MkQhmuuJAVDjmeWNixwDDH · claim posted 2026-09-12T04:27Z}

---

Generated by Claude Code

[os-bill]

os-dev-report

{
  "card": 17260,
  "issue": 17260,
  "status": "done",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/17792",
  "branch": "claude/issue-17260-object-kanban-quickadd-tombstone",
  "head_sha": "6160d2240674cdaf37cf4399c37af726cfb5d82e",
  "files_changed": [
    "packages/spec/src/ui/component.zod.ts",
    "packages/spec/src/conversions/registry.ts",
    "packages/spec/src/migrations/entries/retired-keys/18.ui__ObjectKanbanProps__quickAdd.ts",
    "packages/spec/src/migrations/registry.ts",
    "packages/spec/src/ui/component.test.ts",
    "packages/spec/authorable-surface/ui.json",
    "content/docs/references/ui/component.mdx",
    ".changeset/17260-object-kanban-quick-add-retired.md"
  ],
  "premise_still_valid": true,
  "premise_note": "Against the CORRECTED premise, not the card body's original. Re-measured at the pinned objectui sha (.objectui-sha = 53ded82bf), not at that checkout's HEAD: ObjectKanban.tsx names neither quickAdd nor onQuickAdd (0 and 0) while the control onCardClick returns 6 and objectName 38 in the same file with the same instrument. The board DOES forward the key (ObjectKanban.tsx:930 spreads the authored bag into KanbanRenderer; index.tsx:196 passes quickAdd and onQuickAdd), but KanbanImpl gates the affordance on BOTH (:355, :368) and onQuickAdd is a host-supplied function no producer puts on an object-kanban node. So: published, forwarded, never honoured — accepted-and-dropped, as the card classifies it.",
  "summary": "Retired ObjectKanbanProps.quickAdd by the property-retirement playbook, executing objectui#8285's ruled option B. retiredKey() tombstone on ObjectKanbanPropsSchema (tsc never + parse-time prescription); D2 conversion object-kanban-quick-add-removed (pure lossless delete, scoped by component type so the LIVE kanban-ui spelling is out of reach); RETIRED_KEYS_BY_MAJOR[18] entry ui/ObjectKanbanProps:quickAdd plus the D3 chain-step wiring and rationale; regenerated authorable-surface row to [RETIRED] and the reference docs row; four pin tests; a minor changeset carrying the ADR-0087 registered disposition. The schema docblock that listed quickAdd among the keys read 'via the forwarded schema' is corrected in the same stroke — that sentence was true about the forward and false about the read, which is how the key kept re-authorizing itself. packages/spec/src/ui/view.zod.ts untouched; the kanban-ui block untouched.",
  "objectui_readings": {
    "sha_used": "53ded82bf7a494f54e344e19099dbf00854b8694 (the sha THIS repo pins in .objectui-sha), not the checkout HEAD 3fbdd4a2d",
    "ObjectKanban.tsx": { "onQuickAdd": 0, "quickAdd": 0, "control_onCardClick": 6, "control_objectName": 38 },
    "object_kanban_registration_inputs_at_pin": "index.tsx:421-431 declares exactly objectName + columns — quickAdd absent, control objectName present",
    "where_quickAdd_IS_read_at_pin": "kanban-ui (KanbanRenderer, index.tsx:196) and kanban-enhanced (index.tsx:342); KanbanImpl gates on quickAdd && onQuickAdd at :355 and :368",
    "NOT_COUNTED_at_pin": "OBJECT_KANBAN_INPUTS (the card's 13-key constant) does not exist at 53ded82bf — 0 files. Reported as NOT COUNTED rather than as a zero. Control: it DOES resolve at that checkout's HEAD (3 files), so the probe string is real and the constant is later objectui work. Same for the inert-quick-add interim module: 0 at the pin AND 0 at this checkout's HEAD, so this checkout is itself behind objectui origin/main and I cannot confirm the tripwire from here.",
    "ledger_at_checkout_HEAD": "apps/console/src/__tests__/registry-inputs-spec-parity.test.ts records 'object-kanban.quickAdd' with the verbatim note 'measured NOT honoured by this renderer' — read at HEAD, since the file is later than the pin"
  },
  "liveness_count": {
    "stored_or_example_stacks_carrying_quickAdd": 0,
    "why_the_zero_is_a_reading": "No object-kanban component is authored anywhere under examples/ or apps/ at all. Lit control, same instrument, same corpora: object-grid 3, object-metric 8, page:card 5. Repo-wide, quickAdd occurred in exactly 4 places before this PR — the schema key, the docblock sentence, the authorable-surface row and the generated docs row — i.e. only the carriers being retired.",
    "shape_consequence": "Non-zero would have changed the migration's shape; zero in-repo keeps it a pure lossless delete. Out-of-repo authors remain unknown and unknowable from this seat — that is exactly what the D2 conversion and the prescription are for."
  },
  "registry_regenerated_by": {
    "command": "pnpm --filter @objectstack/spec gen:migration-registry",
    "generator_output": "wrote src/migrations/registry.ts (202 semantic, 168 retired-key, 178 retired-def)",
    "gate": "pnpm --filter @objectstack/spec check:migration-registry -> exit 0, 'src/migrations/registry.ts is current (202 semantic, 168 retired-key, 178 retired-def)'",
    "negative_control": "The build BEFORE that generator run failed loudly: '1 key(s) were tombstoned with no registered retirement: ui/ObjectKanbanProps:quickAdd'. Never hand-edited."
  },
  "clause2_declared": "yes — the measurement confirms it rather than re-declaring: the retirement narrows a published accept set. ObjectKanbanPropsSchema.quickAdd went from z.boolean().optional() (accepted) to retiredKey() (refused with a prescription), and the change reaches the published dist (measured below). PR #17792 carries needs:contract-review. node scripts/pm/check-clause2-carriers.mjs --pair 17792 -> EXIT 0, 'the clause-② declaration is readable in the fixed spelling and both carriers agree'.",
  "changeset_decision": {
    "owed": true,
    "level": "minor",
    "file": ".changeset/17260-object-kanban-quick-add-retired.md",
    "dist_measurement": "The tombstone's prescription text reaches packages/spec/dist/ui/index.js and dist/ui/index.mjs — 2 hits, both inside the package's files[] ('dist' is listed).",
    "positive_control": "A pre-existing shipped describe from the same schema ('Maximum number of records loaded onto the board') — 2 hits, the same two dist files.",
    "negative_control": "Text living only in component.test.ts ('refuses the key by the TOMBSTONE') — 0 hits in dist.",
    "why_not_major": "Not my call to make and I did not make one: scripts/check-changeset-no-major.mjs FORBIDS major during the launch window (lockstep versioning would promote ~70 packages), and the gate ran green on this diff. The sibling retirement one entry over (ui/ObjectGridProps:defaultSort, #11805) is registered under protocol 18 on the same reading. The card's open question ② (minor vs major under ADR-0059 §4) resolves to minor by that guard plus the measurement that api-surface/ is unchanged — no export leaves, ObjectKanbanProps still exists one key narrower.",
    "adr_0087_disposition": "registered object-kanban-quick-add-removed; node scripts/check-adr-0087-registration.mjs --base origin/main -> exit 0"
  },
  "verification": {
    "build": "pnpm --filter @objectstack/spec build -> os-verify-lock VERDICT command-exit 0",
    "tests": "pnpm --filter @objectstack/spec test -> os-verify-lock VERDICT command-exit 0; Test Files 473 passed (473), Tests 13443 passed (13443)",
    "typecheck": "pnpm --filter @objectstack/spec typecheck -> exit 0 (includes check:test-typecheck, 54 files / 259 errors held in the shrink-only debt ledger, unchanged)",
    "check_generated": "pnpm --filter @objectstack/spec check:generated -> 15/15 artifacts current after one gen:docs lap. It names the 14 source audits it does NOT run and the 1 it CANNOT (check:react-declaration-parity, EXTERNAL_INPUT_REQUIRED — needs objectui's sdui.manifest.json, produced only by pnpm sdui:manifest driving a real browser at the pin).",
    "targeted_gates": "check:migration-registry 0 · check-adr-0087-registration 0 · check-changeset-no-major 0 · check:nul-bytes 0 (8453 text files) · check-clause2-carriers --pair 17792 0",
    "ablation": "Source-level (the pins import ./component.zod relatively, same package — no dist hop). Baseline on HEAD: 4 passed. Mutation (tombstone reverted to a live z.boolean().optional()) proven ON DISK before reading any result: tombstone-call count 1 -> 0, injected marker 1, blob hash 106299e7… -> e657a7e3…. Ablated run: 2 failed / 2 passed — the two refusal pins are the discriminating half. Restored with git checkout HEAD -- PATH, proven byte-identical (blob back to 106299e7…, git diff HEAD empty). Reported honestly: the not.toHaveProperty pin does NOT flip under that mutation (an absent optional key is not materialized either way), so it guards the strip direction, not the refusal. Script carried a trap on EXIT INT TERM with absolute paths.",
    "narrowing_declared": "Repo-wide scans (pnpm lint, the 104-command dispatch-gates family) were NOT run locally — CI owns the farm. Locally run: the affected package's own build/test/typecheck, the full spec generated-artifact wrapper, and the gate families this diff implicates (migration registry, ADR-0087 registration, changeset level, nul-bytes, clause-2 pairing)."
  },
  "tests": "See verification.tests / verification.ablation above — real excerpts, not paraphrase.",
  "mcp_calls": "0 — every GitHub read and write went through repo-scoped REST (probe: GET /repos/objectstack-ai/objectstack -> 200); every objectui reading went through git on the pinned checkout.",
  "out_of_scope_findings": [
    "noted, not filed: spec-changes.json and docs/protocol-upgrade-guide.md do NOT project protocol-18 registrations at all (protocolVersion 17.0.0) — this retirement is absent from both and check:spec-changes / check:upgrade-guide are green, correctly. The precedent ui/ObjectGridProps:defaultSort is equally absent (0 hits), so this is the bound, not a miss. It does mean the playbook's 'the change documentation is the primary channel' is DEFERRED for major-18 tombstones until the protocol version moves. 承接者: whoever cuts protocol 18 — the projections regenerate from the same registry then.",
    "noted, not filed: `pnpm --filter @objectstack/spec gen:docs` and check:docs both print 'Warning: Schema directory packages/spec/json-schema/{conversions,migrations} does not exist' on every run, including on unmodified main. Pre-existing warning noise from the docs renderer, not caused by this diff and not reproducible as a defect (the gate passes). 承接者: 无.",
    "noted, not filed: the pinned objectui checkout at /home/user/objectui is at 3fbdd4a2d and carries NO `inert-quick-add` module, so the interim diagnostic and the tripwire row the filer's correction describes (objectui PR #8922, slice C) could not be confirmed from this container at all — not at the pin and not at that checkout's HEAD. The checkout is presumably behind objectui origin/main. This is a reading about my evidence, not a claim that the interim is missing. 承接者: the objectui seat holding objectui#8285 slice C, which owns removing the interim when this pin moves."
  ],
  "open_questions": [
    {
      "question": "The card's flag ① — objectui#8285's ruling cites 'ADR-0049 enforce-or-remove', whose own TL;DR scopes it to security/access-control properties, while quickAdd names a UI affordance. Which citation should the tombstone carry?",
      "options": [
        "A — cite ADR-0049 anyway, as every sibling component-prop retirement in this tree already does",
        "B — cite ADR-0059 §5 / #3855 (the tombstone convention the ledger's own description names) instead",
        "C — cite both"
      ],
      "recommendation": "A, and that is what shipped — not as an adjudication but as the tree's measured house spelling: every neighbouring component-prop retirement in ui/ (view.striped, view.bordered, view.virtualScroll, ObjectGridProps.defaultSort, ElementFilter/ElementForm families) carries 'ADR-0049 enforce-or-remove' in exactly this position, and inventing a second citation format on this one card would make the corpus inconsistent for no reader's benefit. The mechanism half (ADR-0059 §5 / #3855) is carried where it is actually load-bearing — the retirement kit section of the changeset and the registry entry. ⚠️ If the seat wants the citation convention itself changed, that is a sweep over the whole class, not this card."
    }
  ]
}

---

Generated by Claude Code

[os-bill]

os-dev-report

Follow-up to the report above: PR #17792 was CI-red on one line; fixed and re-verified.

{
  "card": 17260,
  "pr": "https://github.com/objectstack-ai/objectstack/pull/17792",
  "previous_head_sha": "6160d2240674cdaf37cf4399c37af726cfb5d82e",
  "head_sha": "10f206359fdf490336571717f8dc033b23080a9f",
  "status": "done",
  "fix": "packages/spec/src/ui/component.zod.ts — the tombstone docblock cited the pin as 'the `.objectui-sha` pin this repo builds against (`53ded82bf`)', which is NEITHER spelling check:objectui-pin-citations recognises, and an unrecognised spelling is a hard failure there by design (it would leave the citation outside every check). Respelled into the ASSERTING form, which is the claim actually being made: 'Measured at the objectui pin this repo builds against (`.objectui-sha` = `53ded82bf`)'.",
  "gate_exit_code": "pnpm --filter @objectstack/spec check:objectui-pin-citations -> EXIT 0 on the pushed head 10f20635: '14 asserting objectui pin citation(s) match .objectui-sha (53ded82bf), 26 historical citation(s) recorded and not checked, across 1389 spec source(s)'. It was EXIT 1 at 6160d224, flagging component.zod.ts:2830 and nothing else — reproduced locally before the fix.",
  "anchor_correction": "The off-by-one was confirmed at the pin (line 930 is `<KanbanRenderer schema={{`, line 931 is `...effectiveSchema`) and corrected to :931 at all SIX sites this PR had added it, not only the flagged line: component.zod.ts x2, conversions/registry.ts, migrations/registry.ts, the retired-keys entry file, component.test.ts and the changeset. Leaving five known-wrong citations while fixing one would have been worse; the PR body is corrected too. Comment-only, no behaviour.",
  "re_verification_on_the_new_head": {
    "build": "pnpm --filter @objectstack/spec build -> exit 0",
    "check_generated": "exit 0 — all 15 generated artifacts up to date (the docblock edit moves no artifact; the registry was regenerated with gen:migration-registry after the entry file changed, and check:migration-registry is current at 202/168/178)",
    "pins_on_head": "4 passed",
    "discrimination": "Re-confirmed against the MERGE-BASE declaration: with quickAdd restored to the merge base's live z.boolean().optional() (mutation proven on disk — tombstone-call 1 -> 0, marker 1, blob e61fa0f7… -> b59f025d…), the same run goes 2 failed / 2 passed. The two refusal pins are the discriminating half; the not.toHaveProperty pin does not flip (an absent optional key is not materialized either way) and guards the strip direction instead. Restored from HEAD, blob back to e61fa0f7…, 0 files differing.",
    "nul_bytes": "exit 0",
    "diff_scope": "The true PR diff is still exactly the same 8 files (git diff origin/main...HEAD, three-dot). A two-dot listing showed five extra paths that are main's own advance from sibling PRs merged meanwhile, not this branch's."
  },
  "review_notes_carried_not_changed": [
    "The changeset's rationale sentence conflates the export ratchet with the key ratchet — acknowledged. The conclusion minor stands, and what licenses it is the launch-window guard plus the moved authorable-surface row, not 'no export leaves'. Left as written per the review's instruction to change nothing else.",
    "'removed in @objectstack/spec 17' while the entry sits under RETIRED_KEYS_BY_MAJOR[18] matches the ui/ObjectGridProps:defaultSort precedent exactly — left as written."
  ],
  "still_draft": "yes — not flipped ready, not enqueued, no auto-merge, not merged. needs:contract-review still on the PR.",
  "platform_note": "The PR body PATCH came back with a second, platform-appended bare attribution footer under the session-URL one. Read back and reported rather than re-patched: re-posting to fix a footer is how a third one arrives."
}

---

Generated by Claude Code