feat(spec,ui)!: retire the record page's slots.tabs; a page that needs its own strip is kind: 'full' by objectstack-fleet[bot] · Pull Request #27219 · objectstack-ai/objectstack

objectstack-fleet · 2026-10-28T11:03:12Z

Requested by Yibo Wang · project thread

What changes

The record page's slots.tabs is retired under the spec-property-retirement kit. A slotted page now fills six slots (header, actions, alerts, highlights, details, discussion); a page that needs its own tab strip is kind: 'full'.

Spec (@objectstack/spec)

slots.tabs is a retiredKey() tombstone carrying the card's prescription: the record page composes its tabs and columns from the object's roles and relationship flags; set relatedList / relatedList: 'primary' on the child's reference field to change which related lists appear or their order; author slots.details to replace the Details body; a page that needs its own strip is kind: 'full'. It ends with the house os migrate meta --from 17 sentence and carries no issue ids.
The related / relatedLists aliases leave with the slot. They are now guidance entries that refuse by name with the same remedy, instead of being read as tabs.
checkPageSlotPair (the details + tabs refusal) and page-slot-pair.test.ts are removed, and so is its @objectstack/spec/ui export. The check-exports pin now asserts that it has zero declarations and zero attachments.
ADR-0087 D2: the page-slots-tabs-removed conversion (retiredFromLoadPath, retiredAfter: '18.2.0', order 78) drops slots.tabs from a stored page. When tabs was the only slot, the page is left as an empty slotted page, which renders the synthesized default.
ADR-0087 D3: the page-slots-tabs-retired entry covers the judgement the conversion cannot make, which is whether such a page should become kind: 'full'. It absorbs page-slots-details-beside-tabs-refused, because the pair no longer exists. The STEP18_RATIONALE fragment moves with it (order 112). The retired-key entry is ui/Page:slots.tabs.
The JOIN row in retired-key-removal-version.pin.test.ts, the liveness slots note, and the regenerated api-surface, export-origins and references.

The four in-repo producers, rebuilt as kind: 'full'. Each rebuild is faithful: the same nodes in the order the synthesizer drew them (header, then the actions/alerts slot, highlights, the strip, then discussion). Nothing is redesigned. Header actions are the object's record_header / record_more actions in declaration order, which is exactly what the synthesized header drew; the host still appends Edit/Share/Delete.

Page	Package	Body
`sys_organization_detail`	`@objectstack/platform-objects`	`page:header` (4 actions) · `record:highlights` (`slug`) · Members / Invitations / Teams strip (unchanged) · `record:discussion`
`sys_user_detail`	`@objectstack/platform-objects`	`page:header` (11 actions) · the `record:alert` from its `alerts` slot · highlights · its curated strip · no discussion (its `discussion: []`)
`sys_approval_request_detail`	`@objectstack/plugin-approvals`	`page:header` · `record:approval_decision` · highlights · Details / Timeline strip · no discussion (its `discussion: []`)
`showcase_project_detail`	`examples/app-showcase` (private)	`page:header` · highlights · Details / Tasks grid / Related strip · `record:discussion`

Console (@object-ui/*)

buildDefaultPageSchema composes from the six slots only, and its tabs branch is removed.
RecordDetailView's pageColumnsEligible drops the slots.tabs arm, so the one gate line is now !assignedPage && !namedPageBlocked.
@object-ui/types removes both .superRefine(checkPageSlotPair) attachments and types PageSlotMap.tabs as never.
The Studio slot canvas (SLOT_ORDER) lists six slots.

Docs

ui/content/docs/guide/slotted-pages.md: the slot menu no longer has tabs, and the guide says how to shape the strip instead.
content/docs/ui/pages.mdx (merged from main during this branch): the same change.
The generated references.
skills/objectstack-ui/rules/pages.md: one bullet. It is paid for by a deletion in the same file (−23 tokens net).

ADR-0133 and QA

ADR-0133's open-question correction gains the inline one-line amendment: the mechanism moves from slotted to full, and the strip does not.
The identity-auth.json source pin moves with it (revision 6).
studio-authoring.json follows the retired pair: the step now PUTs slots.tabs, which is refused with invalid_type. The pin file, the source anchor and revision 4 are updated to match.

Why

An authored tabs slot replaced the whole synthesized strip. A page that used it lost every related list, approval, attachment and history entry the platform composes from the object, and stayed a tab strip at every width instead of taking the record page's columns. The maintainer ruled the slot out on epic #26676; the card carries that ruling.

Acceptance notes

Correction to the card's premise. The card assumed no producer used slots.tabs ("HotCRM never uses tabs"). Re-measured on main at 0f1045bc, four in-repo pages authored it: the table above. Seat 1 confirmed moving all four to kind: 'full' with identical behaviour.
- HotCRM's own account page could not be measured from this session: the hotcrm repo is outside this session's scope. If it authors no tabs, it parses unchanged. If it does, its stored row is converted on read by the D2 conversion, and its source is listed by os migrate meta --from 17.
The tombstone. safeParse of tabs / related / relatedLists fails with the prescription, and the six live slots parse. page-slots-tabs-retirement.test.ts has 18 cases:
- tombstone refusal for one component, an array and empty;
- the alias guidance;
- the six live slots;
- a full page with page:tabs parses;
- no # and the migrate sentence;
- @ts-expect-error on definePage;
- the barrel no longer exports checkPageSlotPair.
Stored rows. A stored slotted page with tabs replays through the conversion: applyConversionsToStoredItem('page', …), the tabs-only page becomes slots: {}, the conversion is idempotent, and it is retired from the load path. The registry fixture has three pages and expects two notices.
Absence pin. A tree-scoped absence pin (packages, examples, skills, content, scripts) refuses a slots frame holding tabs, related or relatedLists in .ts, .json, .yaml and .md fences. It excludes the conversion fixtures, the json-schema output, release notes and changelogs.
Fixtures and tests moved. Each one now uses a live slot (details) or reads the full page's region:
- spec: i18n-resolver.test.ts (two fixtures), page-component-filter-record-to-rule-array.test.ts, object-refinement-check-exports.test.ts, retired-key-removal-version.pin.test.ts (new JOIN row). page-slot-pair.test.ts is deleted.
- lint: page-walk, validate-page-field-bindings, validate-visibility-predicates and page-envelope-audit tests.
- cli: i18n-section-coverage (its fixture is now a kind: 'full' page whose region authors its own page:tabs, keeping the nesting under test), platform-page-i18n-parity (the boundary pin now expects kind: 'full', no slots, every component reached, and none addressed).
- platform-objects: sys-user-details-tab, sys-user-security-tab-own-record, sys-user-two-factor-display, setup-nav-per-user-stores and canonical-expression-envelopes (the control's path is now regions[0].components[N].visibleWhen); new record-page-header-actions.test.ts (round 2, review should-fix ④.1: for sys_organization_detail and sys_user_detail, the page:header action ids equal the object's record_header ∪ record_more declarations in declaration order, with a non-vacuous guard); comments in invite-entry-toolbar.test.ts and sys-member.object.ts.
- plugin-approvals: sys-approval-request-page.test.ts, which pins the page body in order.
- app-showcase: comments in task-detail.page.ts and record-action-location-hosts.test.ts.
- ui: buildDefaultPageSchema.test.ts (a stray tabs key is ignored and details applies), recordPageColumns.test.ts, slots-canvas.test.ts, PagePreview.test.tsx, spec-object-refinements-7715.test.ts (the retired slot is refused at both objectui doors), imported-defaults-8317.test.ts, and the ten console served-approvals-page copies (approvalRequestsDataSource.servedPage*.test.tsx, ApprovalsInboxPage.opensServedPage-12484.test.tsx), which now copy the full page whole.

Verification

Head ca5556e06, with main merged at b322b5da (11 commits, no conflicts). Runs below are local, under os-verify-lock.

spec: --project local has 22219 passed. The 3 reds from the first run were fixed and their files re-run green (166 passed): a conversion fixture moved to details, the new JOIN row, and the export-origins regeneration. test:repo has 82 files / 1343 passed, including the tree-scoped absence pin, which is now listed in vitest.repo-tests.json.
Package suites:
- platform-objects: 96 files / 1377 passed.
- plugin-approvals: 105 files / 1331 passed.
- lint: 7522 passed.
- cli: the three touched files, 59 passed.
- plugin-auth: organization-record-page-option, 3 passed.
ui:
- plugin-detail: 2941 passed.
- Every ui test that authors a slot map or the served approvals page (19 files): 319 passed.
- RecordDetailView record-page tiers and named-page tests: 13 passed.
- PagePreview: 8 passed.
Typecheck: turbo run typecheck over spec, platform-objects, plugin-approvals, lint, cli, example-showcase, @object-ui/types, plugin-detail, app-shell and console: 104 of 104 tasks pass. This includes the test-layer typecheck ledgers.
Derived gates: dispatch-gates --commands lists 126 commands, and 121 exit 0. Of the other five:
- check:role-word and check:cross-package-test-inputs were real findings. Both were fixed and now exit 0.
- check-plugin-teardown-shape --self-test exited 3 because the clone is shallow. After fetching its pinned fixture commit it passes all 48 cases.
- check:skill-examples and check:dual-build-cjs-loads exit 3 because they need a full repo build. That is a prerequisite miss, and nothing was measured.
Spec artifacts: check:generated --fix regenerated three artifacts (api-surface, export-origins and references), and each re-checks green. check-platform-checklist, check-skills-token-ratchet (pages.md 5669/5692) and check-adr-0087-registration --base origin/main are green.

Clause-②: no (narrowing)
Metadata-protocol: narrowed
Skills-docs: follow-up #27059

Keep one value on each of the two lines above, each line starting with its key.
Human Minimum
item 5 says what they mean.

The changeset carries Clause-②: no (narrowing), Metadata-protocol: narrowed, the BREAKING banner, a FROM → TO. line and . The packages bump minor: spec, platform-objects, plugin-approvals, @object-ui/types, @object-ui/plugin-detail, @object-ui/app-shell and skills.

Governed surfaces (docs/adr/**, skills/**, Tier H): this PR is a draft and is not queued or armed. It needs a governed approver's review before the seat lands it.

Human Minimum

Each item is one rule from
CONTRIBUTING.md › Human Minimum,
which links the AGENTS.md text that states it.

1. Schemas start as Zod; packages/spec holds no runtime logic.
2. Every name I added uses the casing for its kind.
3. No formatter ran; pnpm lint passes.
4. One issue, one branch off an up-to-date main, one pull request.
5. A change that publishes carries a changeset, and both declaration lines are filled in here and in the changeset.
6. packages/spec touched: its generated artifacts are regenerated and committed.
7. No hand edits to content/docs/references/, packages/*/CHANGELOG.md or content/docs/releases/.
8. Title, body and comments are in English.
The Before Submitting list in CONTRIBUTING.md is done.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

…DR-0133 amendment, QA pins, skills bullet, changeset Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

…-tc4fe5

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

…rebuild; preview slot fixture Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

… avoid the reserved word Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

github-actions · 2026-10-28T11:04:24Z

📓 Docs Drift Check

This PR changes 3 package(s): @objectstack/platform-objects, @objectstack/plugin-approvals, @objectstack/spec, touching 79 documentable anchor(s). ⚠️ 8 changed file(s) yielded no anchor (packages/plugins/plugin-approvals/src/translations/pages.ts, packages/spec/api-surface/ui.json, packages/spec/export-origins/ui.json, …), so the pages documenting them are NOT COVERED by this run — this is not a clean bill of health for those files.

60 hand-written doc(s) name something this change touched — list omitted above 15 rows. Re-derive on the tree named below: node scripts/docs-audit/affected-docs.mjs --json 1e8e5296f31742a614395d4267fb7de5172a852c.

⛔ 16 release-owned page(s) also affected — read-only, see AGENTS.md Documentation Guardrails.

What this run could not see

8 changed file(s) yielded no anchor (packages/plugins/plugin-approvals/src/translations/pages.ts, packages/spec/api-surface/ui.json, packages/spec/export-origins/ui.json, …) — pages documenting those are invisible to this run
2 anchor(s) matched too much of the corpus to be a work list: created_at (literal, 39 pages), organization_id (literal, 38 pages)
9 name(s) were too generic to anchor anything (single lowercase words)
the SDK route bridge reached 56 of 208 client-bound route-ledger rows — the other 152 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 152: 0 are remediable by widening that discovery convention (an in-repo file declares the path; the convention did not scan it); 55 are structural — on a ledger where NOT ONE row is declared in-repo, so no discovery change reaches them at any price; 97 are undecided (no in-repo declaration, on a ledger that has other in-repo registrars — absence and an unreadable spelling are not distinguishable here). The rows themselves: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 148 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json 1e8e5296f31742a614395d4267fb7de5172a852c → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 3dfd02d0e764224d37973c01646bd498370525e3 — the merge of head 9573871084ddd8e1b72193a39ee6529df0e64af9 into base 1e8e5296f31742a614395d4267fb7de5172a852c, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 3dfd02d0e764224d37973c01646bd498370525e3 && git checkout 3dfd02d0e764224d37973c01646bd498370525e3
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin 1e8e5296f31742a614395d4267fb7de5172a852c 9573871084ddd8e1b72193a39ee6529df0e64af9 && git checkout -B drift-repro 1e8e5296f31742a614395d4267fb7de5172a852c && git merge --no-ff 9573871084ddd8e1b72193a39ee6529df0e64af9

node scripts/docs-audit/affected-docs.mjs --json 1e8e5296f31742a614395d4267fb7de5172a852c

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

Advisory only, and a precision-first one (#9192): a page is listed because it names a
symbol, wire route or SDK method this diff touched — not because it mentions a changed
package. Each row says which anchor put it there, so a wrong row is reportable rather than
merely annoying. To re-verify, run the docs-accuracy-audit workflow scoped to these files:
node scripts/docs-audit/affected-docs.mjs 1e8e5296f31742a614395d4267fb7de5172a852c → pass the list as
args.docs, on the commit named under Which tree this was computed on.

objectstack-fleet · 2026-10-28T11:20:20Z

Contract review

Served-tier: CONTRACT_REVIEW_TIER
Head-sha: ca5556e0661a7d19b5bb990f02c2135bbe746a3a
Local-runs: none
Scope: packages/spec/src/ui/page.zod.ts, packages/spec/src/conversions/registry.ts, packages/spec/src/conversions/walk.ts, packages/spec/src/migrations/entries/retired-keys/18.ui__Page__slots.tabs.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-tabs-retired.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-details-beside-tabs-refused.ts (deleted), packages/spec/src/migrations/registry.ts.template, packages/spec/src/shared/retired-key-removal-version.pin.test.ts, packages/spec/src/ui/page-slots-tabs-retirement.test.ts, packages/spec/src/ui/page-slot-pair.test.ts (deleted), packages/spec/src/ui/object-refinement-check-exports.test.ts, packages/spec/src/system/i18n-resolver.test.ts, packages/spec/src/conversions/page-component-filter-record-to-rule-array.test.ts, packages/spec/api-surface/ui.json, packages/spec/export-origins/ui.json, packages/spec/liveness/page.json, packages/spec/vitest.repo-tests.json, packages/platform-objects/src/pages/sys-user.page.ts, packages/platform-objects/src/pages/sys-organization.page.ts, packages/platform-objects/src/pages/.test.ts (4), packages/platform-objects/src/apps/setup-nav-per-user-stores.test.ts, packages/platform-objects/src/identity/sys-member.object.ts, packages/platform-objects/src/identity/invite-entry-toolbar.test.ts, packages/plugins/plugin-approvals/src/sys-approval-request.page.ts, packages/plugins/plugin-approvals/src/sys-approval-request-page.test.ts, packages/plugins/plugin-approvals/src/approvals-plugin.ts, packages/plugins/plugin-approvals/src/translations/pages.ts, packages/lint/src/.test.ts (4), packages/cli/test/.test.ts (3), examples/app-showcase/src/ui/pages/project-detail.page.ts, examples/app-showcase/src/ui/pages/task-detail.page.ts, examples/app-showcase/test/record-action-location-hosts.test.ts, ui/packages/plugin-detail/src/synth/buildDefaultPageSchema.ts, ui/packages/plugin-detail/src/synth/recordPageColumns.ts, ui/packages/plugin-detail/src/synth/tests/.test.ts (2), ui/packages/plugin-detail/README.md, ui/packages/app-shell/src/views/RecordDetailView.tsx, ui/packages/app-shell/src/views/metadata-admin/previews/PageBlockCanvas.tsx, ui/packages/app-shell/src/views/metadata-admin/previews/PagePreview.tsx (+2 tests), ui/packages/types/src/layout.ts, ui/packages/types/src/zod/layout.zod.ts, ui/packages/types/src/tests/.test.ts (2), ui/apps/console/src/**/.servedPage*.test.tsx (10), skills/objectstack-ui/rules/pages.md, docs/adr/0133-org-management-open-basics.md + ui/packages/components/src/renderers/layout/containers.tsx (PageHeaderRenderer / PageHeaderActions, recordChrome default and action-id resolution), ui/packages/types/src/record-semantics.ts (detectStatusField), packages/platform-objects/src/identity/sys-organization.object.ts, packages/platform-objects/src/identity/sys-user.object.ts, packages/plugins/plugin-approvals/src/sys-approval-request.object.ts, examples/app-showcase/src/data/objects/project.object.ts, packages/platform-objects/src/identity/sys-organization-rename-action.test.ts (headerOrder), packages/spec/src/shared/strict-object.ts (guidance), packages/spec/src/shared/retired-key.ts, .changeset/config.json, scripts/check-adr-0087-registration.mjs (marker form)

Inputs: card #27053 (body; claim 6465037594; os-dev-report 6467336687), epic #26676 round 3 (6459899198), PR #27219 body + 71-file list, the net diff against main at merge-base b322b5da, and the check-runs on the head. Check-runs as read: 28 success (Lint & Repo Gates, every Lint · gate incl. PM dispatch-gates and docs & code guards, Type Check source/consumer/debt-ledger, Spec property liveness, Check Changeset, Governed Surface Queue Guard, Build Core/Docs/UI Deps, Build & E2E UI, Dogfood Regression Gate, Test Core 8/8), 7 skipped, 26 still in progress (Test Core 1–7/8, Test UI 1–16/16, Type Check · workspace, Type Check UI, Lint UI). No failure at read time; the in-progress shards are the gate verdicts on the moved tests and the ui typecheck, and they stand as they land.

① Derived judgments

PageSchema.slots.tabs → retiredKey(PAGE_SLOTS_TABS_RETIRED) (page.zod.ts): the key is refused on every page kind with the prescription; input type never, so definePage refuses it at tsc. Right — the card's Target 1, carried on the slot map the page.assignedProfiles precedent uses. Prescription text matches the card's 「…」 sentence verbatim in substance (roles/relationship flags; relatedList / relatedList: 'primary' on the child's reference field; kind: 'full'), adds the slots.details route, ends with the house os migrate meta --from 17 sentence, carries no tracker number. Names 18.3.0: the tree's label is 18.2.0 and retiredAfter: '18.2.0', the same pair the interfaceConfig.recordAction: 'modal' tombstone/conversion already carries on main, and the new JOIN row derives it.
related / relatedLists leave aliases and become guidance refusals (PAGE_SLOTS_RELATED_NOT_A_SLOT, same remedy). Right — the card says the aliases go with the slot; steering an author into a tombstone would have been wrong, and strictObject's guidance is the documented tool for a key the shape does not declare.
checkPageSlotPair deleted from page.zod.ts, its .superRefine attachment, its @objectstack/spec/ui export (api-surface + export-origins regenerated), page-slot-pair.test.ts, and the two objectui re-attachments (PageNodeSchema, PageKindNodeSchema); the check-exports pin now asserts zero declarations and zero attachments. Right — with no tabs slot the pair cannot be authored; a public export removed under the BREAKING banner (see ②).
D2 page-slots-tabs-removed (registry.ts, order 78 — unique in MAJOR_18_CONVERSIONS; toMajor: 18, retiredFromLoadPath: true, retiredAfter: '18.2.0', surface: 'page.slots.tabs'): mapPages + stripKeys(slots, ['tabs']), copy-on-write (next === slots ? page : …), idempotent (an absent key strips nothing — pinned by a second replay), one notice per stripped slot (fixture: 3 pages, 2 notices). Right. The tabs-only → slots: {} outcome is stated in the docblock, in the fixture comment, in the D3 reason, in the changeset's one-line-fix paragraph and in the PR body — disclosed as the behaviour change it is, not hidden as a respelling.
D3 page-slots-tabs-retired absorbs page-slots-details-beside-tabs-refused (deleted; the STEP18_RATIONALE fragment re-keyed at order 112, unique). Right — absorption has precedent on main (element-flat-data-binding-retired, object-grid-default-filters-retired); the absorbed entry's own prescription pointed INTO the key now retired, so keeping it would have been wrong release text. replacement and acceptanceCriteria both tell the operator when the page should become kind: 'full' (a curated strip, a tab of controls, lists the flags do not produce) — the judgement the walker cannot make is named, as the card requires. No backticks/pipes in surface; no #NNNN in author-shown fields (pinned).
RETIRED_KEYS_BY_MAJOR[18] gains ui/Page:slots.tabs; JOIN row in the removal-version pin; liveness note for slots lists six slots; vitest.repo-tests.json lists the new file. Right.
buildDefaultPageSchema loses the tabs branch and BuildPageOptions.slots.tabs; a details slot is always spliced into the synthesized strip. Right — Target 2; a stray unparsed tabs key is now ignored (pinned), which is the correct runtime answer once the parse refuses it.
RecordDetailView.pageColumnsEligible = !assignedPage && !namedPageBlocked. Right — every slotted page keeps the synthesized strip, so buildDefaultPageColumns can always read it; a full page is drawn as authored and never measures, exactly as a tabs-slot page was before. No tier behaviour changes for any page that parses.
@object-ui/types PageSlotMap.tabs?: never, Studio SLOT_ORDER six slots, slotsToRegions/regionsToSlots round-trip pins moved to details. Right.
Four producers rebuilt as kind: 'full' (judged node-for-node in ④). Right as a derivation: a page:header whose actions are the object's record_header/record_more action IDS in declaration order (the exact filter RecordDetailView.synthParts.recordHeaderActions applied), record:highlights where the slot was authored or where the synthesizer would have derived one, the alerts/actions-slot nodes in the synthesizer's slot order, the strip byte-identical, record:discussion present iff the slotted page did not author discussion: [].
Tests and fixtures moved from slots.tabs to details or to regions[0].components[] — the lint, cli, spec, platform-objects, plugin-approvals, plugin-detail, types and console copies — with the i18n-parity boundary pin now expecting kind: 'full' / no slots / addressed: 0. Right. One fixture nit: packages/cli/test/i18n-section-coverage.test.ts renames the slot to details but keeps a page:tabs node as its value (a strip in the Details-body slot); harmless to what that test measures, but misleading as a fixture.
Tree-scoped absence pin over packages, examples, skills, content, scripts with a declared radius; ui/ is excluded by stated reason (typed Page literals fail tsc there). Read against the tree: the only slots: { … tabs: } frames left under ui/ are the two deliberate controls (buildDefaultPageSchema.test.ts stray-key case, spec-object-refinements-7715.test.ts refusal case). Right, with the radius stated.

② Semver level

.changeset/27053-page-slots-tabs-retired.md: minor for @objectstack/spec, @objectstack/platform-objects, @objectstack/plugin-approvals, @object-ui/types, @object-ui/plugin-detail, @object-ui/app-shell, @objectstack/skills; ! title; BREAKING banner with a FROM → TO. line, the os migrate meta --from 17 sentence, the parse answer (STACK_SCHEMA_INVALID / 422, the save door, tsc), and the HTML-comment marker adr-0087: registered page-slots-tabs-removed, page-slots-tabs-retired in the form check-adr-0087-registration.mjs reads (Check Changeset and the PM dispatch-gates lint are green on the head). Matches what the diff publishes: minor-with-BREAKING is the v18-train convention for a retirement (precedent on main: 25767-retire-16x-retirement-kits, 25923-calendar-gantt-flat-keys-handoff), and the fixed group in .changeset/config.json moves spec/skills/console together. Packages whose only changes are tests (@objectstack/lint, @objectstack/cli, @object-ui/console) owe no entry; examples/app-showcase is private. Clause-②: no (narrowing) and Metadata-protocol: narrowed on the PR body and in the changeset are the right lines: the diff admits nothing new, refuses slots.tabs / related / relatedLists, and removes one public function.

③ Boundary flags

premise_still_valid: false — four in-repo producers authored slots.tabs (the card said none). The dev re-measured on 0f1045bc, and the PR body says seat 1 confirmed moving all four to kind: 'full' with identical behaviour plus the ADR-0133 inline amendment. Answered in the tree (④ judges the rebuilds faithful). Should-fix (process): that ruling is not on the card — comment 6465037594 (the claim) does not carry it, and the os-dev-report cites it second-hand — so the card's Target/Acceptance still read as if no producer existed. One ruling comment on spec: a slotted record page keeps six slots — header, actions, alerts, highlights, details, discussion — and the tabs slot is retired; a page that needs its own tab strip is a kind: full page #27053 naming the four pages, kind: 'full', and the ADR-0133 amendment closes the trace.
open_questions[0] — HotCRM's account page not measured from this session. Answered by the card's own source line (HotCRM's one slotted page authors header + discussion, never tabs), restated in the D3 reason; and a stored row that did carry tabs is converted on read by D2 in any case. Nothing outstanding.
Dev gates not measured locally (check:skill-examples, check:dual-build-cjs-loads need a full build; check-plugin-teardown-shape --self-test passed after fetching its fixture commit): the Lint · family and Build Core/UI Deps are green on the head; Test Core/UI shards and Type Check · workspace/UI are in progress and are the verdict on the moved tests.
Tier H: docs/adr/** and skills/** are touched; the Governed Surface Queue Guard is green and the PR is a draft awaiting a governed approver, as the body says. ADR-0133's amendment is the inline one-line form the house uses (a block-quoted Amended (date) paragraph after the one it corrects), and it is accurate: the mechanism moved, the strip and Members-at-index-0 did not; identity-auth.json rev 6 moves the source pin with it.

④ Design judgment

The narrowing, read against the published spec on main and the live consumer: a slotted page's tabs slot was a declared capability whose consumer (the synthesizer) delivered it only by dropping every synthesized tab and by opting the page out of the column tiers. The maintainer's keep/narrow/retire ruling retired it; the tree implements the retirement at the spec, not a consumer-side narrowing: the spec refuses, the conversion strips, the synthesizer no longer reads, and the one capability an author still needs (its own strip) is delivered by kind: 'full', which the spec already declares. Nothing declared is taken away without its replacement being real.

(1) Faithfulness of the four kind: 'full' rebuilds — judged against what buildDefaultPageSchema composed for each slotted page on main. The synthesizer's order was header → actions slot → alerts slot → highlights → strip → discussion, no aside (no showReferenceRail), no attachments/approvals/history tabs (those live inside buildDefaultTabs, which an authored tabs slot replaced, so nothing synthesized is lost that the slotted page still had).

Header. Synthesized: page:header { recordChrome: true, actions: [objectDef.actions filtered to record_header ∪ record_more, deduped by name, declaration order] }, actions omitted when empty. Rebuilt: page:header with properties.actions as action IDS. PageHeaderRenderer (containers.tsx) resolves ids to the object's defs, applies the same location filter (actionRendersAt record_header | record_more), capability and visible gates, localizes once, dedupes against RecordContext.headerSystemActions (Edit/Share/Delete appended by the host for authored and synthesized pages alike), routes record_more-only actions to the overflow, and orders by order then input index — so identical input order yields the identical header. recordChrome defaults to true unless explicitly false. Verified lists: sys_organization → set_active_organization, invite_user, leave_organization, rename_organization (declaration order, lines 134/159/199/238 of the object; update_organization/delete_organization are list_item only); sys_user → the 11 ids in declaration order (ban_user … delete_my_account; enable_two_factor/disable_two_factor/generate_backup_codes are record_section only); sys_approval_request and showcase_project declare no header-location action → bare { type: 'page:header' }, which is what the synthesizer emitted.
Highlights. sys_organization authored no highlights slot: the synthesizer derived highlightFields ['name','slug'] minus the title field → ['slug'], and no record:path (no stageField, no status/stage/state/phase field, no status-typed field on the object — detectStatusField answers null). Rebuilt: record:highlights { fields: ['slug'] }, no path. The other three authored their highlights slot (which replaced the whole strip, path included) and carry the same node.
Alerts / actions slot. sys_user's record:alert with its P CEL visibleWhen envelope is carried verbatim (the envelope-audit control now addresses it at regions[0].components[N].visibleWhen); sys_approval_request's record:approval_decision sits after the header, before highlights — the synthesizer's slot order.
Strip. A whitespace-insensitive, comment-stripped diff of each page old→new shows the page:tabs node unchanged in every item, label map, icon, record:related_list columns/sort/limit/filter, the Security tab's TWO_FACTOR_OFFERED gate and own-record gate on sys_user, the record:line_items Tasks grid on the showcase page. The sys_user Security / two-factor / details-tab pins and the approvals page-body pin read the same nodes from regions[0].
Discussion. Present on sys_organization and showcase_project (synthesized footer), absent on sys_user and sys_approval_request (their discussion: []); stripDiscussionNodes on a feeds-off object applies to both page kinds.
Host behaviour keyed on assignedPage. A full page skips the synthesized historyPage read (no History tab existed on any of the four), takes the same attachments/approvals fallback panels (read off renderedPage), the same pruneUnregisteredRecordPageObjects, the same named-page route, and the same pageColumnsEligible = false the tabs slot produced — tabbed at every width, in the drawer, and on mobile, exactly as before. RecordMetaFooter is the view's, not the page's.
Should-fix: the two platform pages' header actions arrays are a frozen copy of what the synthesizer read live from the object. No pin holds them equal to the object's record_header ∪ record_more declarations in declaration order (sys-organization-rename-action.test.ts reads the object only; the sys_user pins read the strip). Add one pin per page (ids == declared ids with those locations, in order), so the next header action added to sys_user / sys_organization cannot silently miss the page.

(2) D2 honesty. Disclosed (①.4), idempotent, copy-on-write, order unique, retiredAfter/toMajor consistent with the tree's label and its sibling at 18.2.0; retiredFromLoadPath pinned both ways (a live author is refused, a stored row is rewritten). The D3 entry tells the operator when kind: 'full' is the answer. Coverage boundary (only stack.pages[]) is stated and matches every other page conversion in the registry.

(3) Tombstone and aliases — ①.1–2: text matches the card's prescription, aliases refuse by name with the same remedy (pinned: tombstone for one component, an array and empty; alias guidance; six live slots parse; a full page with page:tabs parses; no #, migrate sentence; @ts-expect-error; barrel export gone).

(4) Changeset / ADR-0087 / BREAKING — ②.

(5) ADR-0133 amendment and the skills bullet. The amendment is one line, dated, after the paragraph it corrects, and says exactly what moved and what did not. skills/objectstack-ui/rules/pages.md gains one accurate bullet under the kind list ("A slotted record page has no tabs slot: shape the strip with relatedList / relatedList: 'primary' on the child's reference field, or author kind: 'full'") paid for by a same-meaning tightening of the Tailwind-in-className warning (the examples text-[27px], grid-cols-7 and the "compiled over the renderer's source, never over your metadata" rule survive); no other tabs-slot teaching remains in the file. ui/content/docs/guide/slotted-pages.md and content/docs/ui/pages.mdx drop the row and say how to shape the strip; the plugin-detail README's tier sentence is corrected.

(6) RecordDetailView.pageColumnsEligible — ①.8: right, and the comment above it says why the arm left.

No must-fix. Two should-fixes: the header-action drift pins (④.1) and the ruling trace on the card (③). One fixture nit (①.11).

Implemented-by: claude/project-thread-tc4fe5
Reviewed-by: session_01WGufzfhobWEgUUrjz4rrbh

VERDICT: PASS

…ection-coverage fixture is a full page Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

…-tc4fe5

objectstack-fleet · 2026-10-28T11:51:47Z

席位核对 · ACCEPT · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T11:51Z

Head 24f5161872b1e27a90177e926b688a2b3432e17d. Card #27053, claim 6465037594, seat ruling 6467656574.

Contract review: owed at CONTRACT_REVIEW_TIER, narrowing face. Record 6467641976 at ca5556e0 is a PASS with no must-fix. It judged the four kind: 'full' rebuilds faithful node for node against buildDefaultPageSchema on main. It also judged sound the D2 page-slots-tabs-removed (idempotent, copy-on-write, retiredAfter 18.2.0), the D3 entry absorbing the pair entry, the tombstone text, the minor ×7 changeset, and the ADR-0133 amendment.
Should-fixes closed:
- ③ ruling trace: the seat's ruling is now on the card (6467656574).
- ④.1 drift pins: 3b2e4329 adds record-page-header-actions.test.ts. It holds the page:header action ids of sys_organization_detail and sys_user_detail equal to the object's record_header ∪ record_more declarations, in declaration order, with a guard against a vacuous pass.
- The fixture nit: i18n-section-coverage now uses a kind: 'full' page whose region authors its own page:tabs. This keeps the nesting the test exists to cover, which a bare record:details in the slot would have dropped. The seat accepted that route.
- The seat applied the body delta.
Delta: 24f51618 merges main (no rebase, no conflicts). Outside the two round-2 test files, the PR's net hunks against the merge-base are byte-identical to those at ca5556e0, which was measured.
Gates: check-runs on this head are 57 success, 7 skipped and 0 red. The earlier head's Test Core 5/8 red does not recur; it sat in a whole-package leg this diff does not touch. mergeable_state is clean.
Landing: check-governed-merges --pr 27219: GOVERNED (2 of 72 paths; docs/adr/**, skills/**), Tier H. The seat neither flips it ready nor arms it. It lands by the maintainer's hand, or by an APPROVED review from a governed approver, after which this seat arms it.

objectstack-fleet · 2026-10-28T13:09:27Z

triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

Contract-review-claim: session_01AavokzJ5DndAwitDXvKy4U · head 24f5161872b1e27a90177e926b688a2b3432e17d · 2026-10-28T13:08Z

Triage seat, under the gap rule (#27062) condition (b): no record and no claim on this head after one round. The review runs at CONTRACT_REVIEW_TIER. If the lane claims or posts a record on this head first, triage withdraws.

Generated by Claude Code

objectstack-fleet · 2026-10-28T13:24:01Z

Contract review

Served-tier: CONTRACT_REVIEW_TIER
Head-sha: 24f5161872b1e27a90177e926b688a2b3432e17d
Local-runs: none
Scope: packages/spec/src/conversions/registry.ts, packages/spec/src/conversions/walk.ts, packages/spec/src/migrations/entries/retired-keys/18.ui__Page__slots.tabs.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-details-beside-tabs-refused.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-tabs-retired.ts, packages/spec/src/migrations/registry.ts.template, packages/spec/src/ui/page.zod.ts, docs/adr/0133-org-management-open-basics.md, skills/objectstack-ui/rules/pages.md + packages/spec/src/shared/retired-key.ts (retiredKey, the migrate sentence), packages/spec/src/shared/strict-object.ts (aliases / guidance claim rule), packages/spec/src/migrations/types.ts (SemanticMigration.conversionIds, RationaleFragment.order), packages/spec/src/conversions/registry.ts stripKeys / inApplicationOrder / OrderedConversion, packages/spec/src/conversions/walk.ts mapPages, packages/spec/src/ui/page-slots-tabs-retirement.test.ts, packages/spec/src/shared/retired-key-removal-version.pin.test.ts (JOIN row), packages/spec/liveness/page.json, packages/spec/package.json (18.2.0), packages/spec/src/data/field.zod.ts (relatedList), packages/platform-objects/src/pages/record-page-header-actions.test.ts, packages/platform-objects/src/pages/sys-user.page.ts, packages/platform-objects/src/pages/sys-organization.page.ts, packages/platform-objects/src/identity/sys-user.object.ts, packages/platform-objects/src/identity/sys-organization.object.ts, packages/cli/test/i18n-section-coverage.test.ts, examples/app-showcase/src/ui/pages/project-detail.page.ts, ui/packages/app-shell/src/views/RecordDetailView.tsx (pageColumnsEligible; recordHeaderActions on main), ui/packages/components/src/renderers/layout/containers.tsx (PageHeaderRenderer id resolution), docs/qa/platform-checklist/areas/identity-auth.json (rev 6), docs/qa/platform-checklist/areas/studio-authoring.json (rev 4), .changeset/27053-page-slots-tabs-retired.md, .changeset/config.json, skills/objectstack-ui/evals/*.json, .claude/skills/pm-dispatch/SKILL.md and references/review-checklist.md (the Skills-docs: rule), card #27059 (the owed skills/evals follow-up)

Inputs: card #27053 (body; claim 6465037594; os-dev-report 6467336687; seat ruling 6467656574), PR #27219 body + 72-file list, the net diff against main at the merge-base cdec7473ef (net.diff), the net diff at the previous head ca5556e0 (net-old.diff), the earlier record 6467641976, the seat ACCEPT 6468251237, and the check-runs on 24f5161872: 68 runs, 59 success, 9 skipped (Auto Label, Check PR Size, UI Published Artifacts, Temporal Conformance, Dogfood Verify CLI, Dogfood Regression Gate matrix, Packed-tarball smoke, 18.x upgrade harness, Schedule gate — all opt-in or path-skipped), 0 failure, 0 pending. Green include Lint & Repo Gates, every Lint · gate (PM dispatch-gates, docs & code guards, ESLint, scoped), PR-body guards, Check Changeset, Governed Surface Queue Guard, Spec property liveness, Type Check source/consumer/debt-ledger/workspace/UI, Build Core/Docs/UI Deps, Build & E2E UI, Dogfood Regression Gate, Test Core 1–8/8 and Test UI 1–16/16. Nothing stated here was run locally.

The moved head — delta `ca5556e0` → `24f51618`

Commits since the last record: 3b2e4329ba (two test files, +87/−28) and 24f5161872, a merge of origin/main at cdec7473ef (no rebase).
Merge recomputed: git merge-tree --write-tree 3b2e4329ba cdec7473ef yields tree 4b8c8cdbe0707a44eff0034a6a4b1b63afe6408b, byte-equal to 24f5161872^{tree}. The merge carries nothing beyond its two parents: no evil-merge content, no conflict resolution.
Net diff old vs new: per-file hunk comparison of net-old.diff and net.diff differs in exactly three files: the two round-2 test files, and ui/packages/app-shell/src/views/RecordDetailView.tsx, where only one context line moved (main's console: one workspace-preference module (record open mode, drawer width, timeline / related column toggles) read per user × object, with a shared panel-toggle toolbar on the record page — defaults derived from the object, never metadata #27052 "reader's column switches" comment now sits where const [pageRecord…] was); the PR's own hunk — the comment rewrite and pageColumnsEligible = !assignedPage && !namedPageBlocked — is unchanged. The seat's "byte-identical outside the two files" is true of the hunks' added/removed lines, not of one context line; nothing turns on it.
packages/platform-objects/src/pages/record-page-header-actions.test.ts (new, the earlier record's should-fix ④.1). Read against the synthesizer it replaces: RecordDetailView.tsx on main (recordHeaderActions, ~:3318) keeps an action iff a.locations includes record_header or record_more, dedupes by name keeping the first, in declaration order. The pin's headerLocatedIds is that filter exactly (explicit locations only — the synthesizer also never defaulted an action without locations into the header). pageHeaderIds reads regions[0].components[].type === 'page:header' .properties.actions, which is where both full pages put it (sys-organization.page.ts:63, sys-user.page.ts:92). Non-vacuity guard present (the object must declare at least one header-located action). Checked by hand against the objects at this head: sys_organization → set_active_organization, invite_user, leave_organization, rename_organization (object :134/:159/:199/:238; create_organization is list_toolbar, update_/delete_organization list_item only); sys_user → the 11 ids ban_user … delete_my_account in declaration order (invite_user/create_user list_toolbar; the 2FA trio record_section only). Test Core is green on this head. Resolves ④.1 exactly as asked.
packages/cli/test/i18n-section-coverage.test.ts (the earlier record's fixture nit). The fixture is now kind: 'full' with regions[0].components[0] a page:tabs whose item children[0] is record:details — the nesting the test exists to walk, and the real shape of showcase_project_detail at this head (project-detail.page.ts:28/:49/:69). The third case's path moved to regions[0].components[0].properties.items[0].children[0].properties.sections. Resolved; the route the seat accepted is the right one (a bare record:details in slots.details would have dropped the nesting under test).
Earlier record's should-fix ③ (ruling trace): seat ruling 6467656574 is on the card, naming the four pages, kind: 'full', the faithfulness criteria and the ADR-0133 amendment. Resolved.
Earlier record's other judgments (①.1–12, ②, ④.1–6) re-read on this head: every hits hunk is byte-identical to the one measured at ca5556e0 (per-file comparison above), so they hold; the ones this round re-derived independently are stated below rather than adopted.
main has moved past the merge-base (22 commits, cdec7473ef..4d3a077f66, six of them touching PR files: registry.ts, registry.ts.template, retired-key-removal-version.pin.test.ts, vitest.repo-tests.json, studio-authoring.json, RecordDetailView.tsx). git merge-tree --write-tree prr-27219 origin/main is clean (exit 0, tree 603300fa96…). feat(spec,app-shell)!: a form view's type is its layout only — simple, tabbed or wizard; split, drawer and modal are retired with pane, splitSize and splitResizable #27229 (merged 11:56Z, after this head) adds formViewPresentationTypesRemoved at order: 78 and a STEP18 fragment ui-form-view-type-layout-only at order: 112 — the same numbers this PR uses. Both lists are "ascending order, ties broken by id" (inApplicationOrder; RationaleFragment.order), the surfaces are independent (form view vs page slot map), and MAJOR_18 already carries tied orders (35, 55, 57, 61, 63, 67), so the tie is legal and deterministic — not a conflict, not a finding. The check-runs measured the merge into cdec7473ef, so the Tier H landing merges a newer main; GitHub's merge of it is clean as computed.

① Derived judgments

PageSchema.slots.tabs: retiredKey(PAGE_SLOTS_TABS_RETIRED) (page.zod.ts:2125–2129) inside the .optional() slot map, the assignedProfiles precedent (:2034). Input type never (tsc channel), parse issue invalid_type / expected: 'never' at slots.tabs with the prescription, on every page kind (pinned, including kindless). Text: key-first in backticks, "removed in @objectstack/spec 18.3.0 (ADR-0085)", the card's three prescriptions verbatim in substance (roles/relationship flags; relatedList / relatedList: 'primary' on the child's reference field; kind: 'full'), plus slots.details for the Details body, the house migrate sentence, no #NNNN. 18.3.0 is the removal version the tree's label 18.2.0 implies and the nine other 18.x tombstones in page.zod.ts spell; retiredAfter: '18.2.0' on the D2 agrees. Right. relatedList: 'primary' is real on the reference field (field.zod.ts:1725, ADR-0085).
related / relatedLists leave aliases and enter guidance with PAGE_SLOTS_RELATED_NOT_A_SLOT (same remedy, no "did you mean tabs"). Under strict-object.ts's claim rule an alias must point at a key the shape accepts; tabs no longer is, so the alias could not stay. Right.
checkPageSlotPair removed with its .superRefine, its @objectstack/spec/ui export (api-surface/export-origins regenerated), page-slot-pair.test.ts, and the two objectui re-attachments; the trailing comment now counts seven checks. No live reference to the symbol remains at this head except CHANGELOG history and one QA-checklist provenance note (studio-authoring.json:1936, a dated correction record). Right; the public removal is under the BREAKING banner (②).
D2 page-slots-tabs-removed (registry.ts:7929–7986; toMajor: 18, retiredFromLoadPath: true, retiredAfter: '18.2.0', surface: 'page.slots.tabs', order: 78): mapPages over stack.pages[] + stripKeys(slots, ['tabs'], emit, ${path}.slots), which is copy-on-write (touched ? next : item) and emits one (removed) notice at pages[i].slots.tabs; the page is re-spread only when the map changed. Idempotent (absent key → same reference). Fixture: three pages, two notices; the tabs-only page becomes slots: {}. Stored-row replay pinned via applyConversionsToStoredItem('page', …), idempotence pinned by a second collectConversionNotices replay, load-path retirement pinned (live author untouched by the chain, refused at parse). Coverage boundary (stack.pages[] only) stated and true: slots is a top-level PageSchema key with no nested spelling. Right, and the tabs-only → empty-slotted outcome is disclosed in the docblock, fixture, D3 reason, changeset and body.
D3 page-slots-tabs-retired absorbs page-slots-details-beside-tabs-refused (deleted; its STEP18 fragment at order 92 replaced by one keyed page-slots-tabs-retired at order 112; fragment id equals the D3 id as RationaleFragment requires). The absorbed entry's replacement pointed INTO slots.tabs, so it could not survive as release text. The new entry names the judgement the strip cannot make (curated strip → kind: 'full') in replacement, reason and acceptanceCriteria; no # in author-shown fields (pinned); no backticks/pipes in surface. Right as a disposition; one join omission in ④.2.
RETIRED_KEYS_BY_MAJOR[18] += ui/Page:slots.tabs (entry form matches 18.ui__Page__assignedProfiles.ts), JOIN row in the removal-version pin, liveness slots note lists six, vitest.repo-tests.json lists the new file. Right.
Synthesizer and host: buildDefaultPageSchema loses the tabs branch; pageColumnsEligible = !assignedPage && !namedPageBlocked. Every slotted page now keeps the synthesized strip, so the tier measurement can always read it; a full page is drawn as authored and never measures, as a tabs-slot page was. Right.
@object-ui/types: PageSlotMap.tabs?: never, both superRefine(checkPageSlotPair) gone; Studio SLOT_ORDER six. Right.
Four producers → kind: 'full', faithful to the synthesizer's order (header → actions/alerts slot → highlights → strip → discussion); header actions are action IDS that PageHeaderRenderer (containers.tsx:2078, :2264) resolves with the same record_header | record_more placement, routing record_more-only to the overflow, host Edit/Share/Delete appended as before (:1800–1805). Now also pinned to the object for the two platform pages (delta above). Right.
Tests/fixtures moved to details or regions[0].components[]; the tree-scoped absence pin walks packages, examples, skills, content, scripts with a stated radius, structural matcher, anti-vacuity controls (more than 1000 files visited, more than 50 of them slots-bearing) and reasoned exclusions. Right.

② Semver level

.changeset/27053-page-slots-tabs-retired.md: minor ×7 (@objectstack/spec, @objectstack/platform-objects, @objectstack/plugin-approvals, @object-ui/types, @object-ui/plugin-detail, @object-ui/app-shell, @objectstack/skills), ! title, Clause-②: no (narrowing), Metadata-protocol: narrowed (…), the adr-0087: registered page-slots-tabs-removed, page-slots-tabs-retired marker, BREAKING banner with a FROM → TO. line, the house migrate sentence, the parse answer and the pair-check removal. Matches what the diff publishes: narrowed ⇒ minor is the rule the dev agent file states and Check Changeset enforces (green, twice); minor-with-BREAKING for an 18.x retirement has precedent on main (25767-retire-16x-retirement-kits: spec minor, no (narrowing), narrowed). Packages touched only in tests (@objectstack/lint, @objectstack/cli, @object-ui/console) owe nothing; @objectstack/example-showcase is private. Note, not a finding: @object-ui/app-shell is private: true at this head (unpublished by #25922) yet carries a line — changesets tolerates it and main does the same (25769, 25922). Clause-②: no (narrowing) and Metadata-protocol: narrowed are right on the body and the changeset: nothing new is admitted, three spellings are refused, one export is removed.

The Skills-docs: line — absent, and owed. The lane's rule (pm-dispatch SKILL.md claim template; review-checklist.md "种类非 none 或带 (narrowing):Skills-docs: 行在且属实 … 缺 ⇒ REWORK"; os-dev.md:279) requires a third declaration line on the PR body for a narrowed kind or a (narrowing) arm: Skills-docs: updated here | follow-up owed — FILES | follow-up #N | none — REASON. The body carries Clause-② and Metadata-protocol and no Skills-docs:; neither does the claim 6465037594; the seat ACCEPT 6468251237 does not mention it although it is item 6 of the seat's own six. Read off the diff, the true value is Skills-docs: follow-up #27059: this PR changes skills/objectstack-ui/rules/pages.md (one bullet) and the docs, but updated here also requires "the package's evals exercise it", and no file under skills/objectstack-ui/evals/ mentions slots, slotted or tabs at this head; card #27059 (open, pm:blocked on this card) owns exactly that eval ("put the tasks first on the lead page" → relatedList: 'primary') and the 「记录页」 rules section. Should-fix (declaration, seat-owed before landing): add the line to the PR body; it does not move the head. No repo gate reads this line (PR-body guards and Check Changeset are green without it), which is why it is not a FAIL of the tree; it is the ACCEPT that is incomplete on its own checklist.

③ Boundary flags

premise_still_valid: false (four in-repo producers authored slots.tabs): answered in the tree (four faithful kind: 'full' rebuilds, now pinned for the two platform pages) and on the card by seat ruling 6467656574, which follows the maintainer's keep/narrow/retire ruling (card ruling 5) and so needs no new maintainer decision. Closed.
open_questions[0] — HotCRM's account page not measured here. The card's own source line says it authors header + discussion only; the D3 reason restates it; a stored row that did carry tabs is stripped on read by D2 and listed by os migrate meta --from 17. Nothing outstanding.
Dev gates not measured locally (check:skill-examples, check:dual-build-cjs-loads, full-build prerequisites): the Lint · family, Build Core/UI Deps and Build & E2E UI are green on the head. Closed by the check-runs.
Tier H: docs/adr/** and skills/** touched; Governed Surface Queue Guard green; PR is a draft, not armed. ADR-0133's inline Amended (2026-10-28) paragraph sits after the paragraph it corrects and is accurate (mechanism slotted → full; Members/Invitations/Teams strip and Members-at-index-0 unchanged); its first consumer identity-auth.json moved its source[1] pin and logged revision 6 as provenance-only. studio-authoring.json rev 4 follows the retired pair (step PUTs slots.tabs, refused invalid_type at slots.tabs — the zod code the pin asserts), automated.ref and source renamed. Holds.
Skills-docs: line — see ②; escalated to the seat.
main moved after the head — see the delta section; clean merge, legal order ties. The seat re-merges main when it lands (Tier H), as the body already anticipates.

④ Design judgment

The narrowing against the published spec on main: a slotted page's tabs slot was a declared capability whose only consumer delivered it by dropping every synthesized tab and opting the page out of the column tiers. The tree retires it at the spec (tombstone), repairs data at rest (D2 strip, disclosed), names the one judgement a walker cannot make (D3), removes the pair check that only existed because of the slot, and leaves the capability an author still needs — its own strip — on kind: 'full', which the spec already declares and the four rebuilt producers now exercise. Nothing declared is taken away without its replacement being real.

(1) page.zod.ts — against card Target 1 and its first consumers (definePage, PageSchema.safeParse, the @object-ui/types mirror). Tombstone placement and text ①.1; aliases ①.2; pair check ①.3. Tried to break it: an empty tabs: [] is refused (pinned); a kindless page is refused (pinned); related: is not steered into the tombstone (pinned); the six live slots parse together (control). The slot-map docblock now says there is no tabs slot and why. Holds. No finding.

(2) registry.ts (D2) and the D3 entry — against ADR-0087 and the spec-property-retirement kit. D2 ①.4 holds: lossless strip, copy-on-write, idempotent, one notice per site, load-path retirement pinned both ways, boundary stated. D3 ①.5 holds as a disposition. Should-fix ④.2: 18.page-slots-tabs-retired.ts omits conversionIds: ['page-slots-tabs-removed']. SemanticMigration.conversionIds (types.ts:113–133) is "the join between an applied edit and the judgement the consumer still owes … Omit it when the entry judges no mechanical edit"; this entry's whole reason is that judgement over the strip's applied edits ("whether such a page should become kind: 'full' instead is a judgement no walker can make"), so the join is owed, and the chain's printer cannot show the entry beside each page-slots-tabs-removed edit without it. Precedent cuts both ways (24 step-18 entries carry it, e.g. page-kind-jsx-retired, page-requires-non-compiled-kind-refused, element-flat-data-binding-retired; many strip-paired entries do not), the field is optional and migrations.test.ts only checks that a named id is replayed by the step — so a one-line addition, not a FAIL.

(3) walk.ts — comment-only: the path examples move from slots.tabs to slots.details, and the #6776 history now says the four sites were slots.tabs, retired since, with stored copies still walked before the order-78 strip. True of the tree (component-tree conversions at lower order see the slot; the strip then removes it). Holds.

(4) registry.ts.template — the STEP18 fragment ①.5: id = D3 id, order 112 (ties legal), text names the slot, the six-slot map, the alias refusals, the three remedies, the pair refusal leaving, the D2 and D3 ids. src/migrations/registry.ts is gitignored and built from this template plus entries/ (build-migration-registry.ts; check:generated green), so no generated copy is owed in the diff. Holds.

(5) 18.ui__Page__slots.tabs.ts — export const entry = 'ui/Page:slots.tabs', the nested-key form RETIRED_KEYS_BY_MAJOR reads (pinned); header comment names the D2 and D3 ids and the alias change. Holds.

(6) 18.page-slots-details-beside-tabs-refused.ts (deleted) — correct to delete rather than keep: its replacement and acceptanceCriteria instruct the author to move record:details INTO slots.tabs, which is now refused; the absorption is recorded in the surviving entry, the pin asserts the id is gone, and check-adr-0087-registration is green on the changeset marker. Holds.

(7) docs/adr/0133-org-management-open-basics.md — against its first consumer (identity-auth.json org-management item, source[1]) and the page it describes: one dated inline amendment, accurate (③). Holds. No finding.

(8) skills/objectstack-ui/rules/pages.md — against the spec it teaches on main (relatedList / relatedList: 'primary' on the reference field; kind: 'full') and the card's acceptance (Target 3: "the skill bullet"). The bullet is accurate and paid for by a same-meaning tightening of the Tailwind warning (ratchet 5669/5692, check-skills-token-ratchet green). The card names no eval prompt for this bullet; the eval belongs to #27059 (②). Should-fix ④.3 (placement): the bullet lands at :138 as a lone list item under the "AI-authored source pages — kind:'html' and kind:'react'" heading, after the kind:'jsx' retirement sentence, rather than beside the full / slotted row (:130) it qualifies; a reader scanning for slotted-page rules will not find it there. One-line move; #27059's 「记录页」 section will restructure this anyway, so it may be paid there.

No must-fix. Three should-fixes: the Skills-docs: follow-up #27059 body line (②, seat-owed before landing), the D3 conversionIds join (④.2), the bullet's placement (④.3).

Implemented-by: claude/project-thread-tc4fe5
Reviewed-by: session_01AavokzJ5DndAwitDXvKy4U

VERDICT: PASS

Generated by Claude Code

The D3 entry judges the strip that page-slots-tabs-removed applies, so it names that conversion in conversionIds. The skills bullet about the retired tabs slot moves from the source-pages section to Templates & Regions, beside the kind: 'full' rules it qualifies. Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

objectstack-fleet · 2026-10-28T13:57:50Z

席位核对 · ACCEPT · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T13:57Z

Head 9573871084ddd8e1b72193a39ee6529df0e64af9. Card #27053. This supersedes the seat's ACCEPT 6468251237 at 24f51618.

The triage seat's review 6470156834 at 24f51618 is a PASS with no must-fix and three should-fixes. All three are closed:
- 95738710 adds conversionIds: ['page-slots-tabs-removed'] to the D3 entry 18.page-slots-tabs-retired.ts, which joins the entry to its conversion.
- 95738710 moves the retired-tabs-slot bullet in skills/objectstack-ui/rules/pages.md into Templates & Regions, beside the kind: 'full' rules. The ratchet stays green.
- The seat added Skills-docs: follow-up #27059 under the declaration lines in the body. The evals for this bullet live on skills(objectstack-ui) + docs: teach the record page as it is now — object roles produce the page, six slotted slots, kind: full as the escape hatch, drawer default with page by declaration, the column tiers and the reader's own toggles — and stop naming tabs, modal / split / popover #27059.
Delta: the seat measured git diff 24f51618 9573871: one commit, 2 files, +3/−3, and nothing else moves. There is no main merge, and mergeable_state is clean.
Gates: check-runs on this head are 60 success, 9 skipped and 0 red.
Landing: Tier H (docs/adr/**, skills/**). The seat neither flips it ready nor arms it. There is no governed approval on file yet. When an APPROVED review from a governed approver lands, this seat arms it.

objectstack-fleet · 2026-10-28T14:43:58Z

席位核对 · armed · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T14:43Z

The seat's ACCEPT 6470956519 at 9573871084ddd8e1b72193a39ee6529df0e64af9 stands, and the head has not moved. os-zhuang's APPROVED review 5684520620 on this head is a governed approval, which lifts Tier H. mergeable_state is clean. pr_ready + automerge_enable in this act.

objectstack-fleet · 2026-10-28T14:59:32Z

triage#1 · os-zhuang · session_01AavokzJ5DndAwitDXvKy4U

Contract-review-claim: session_01AavokzJ5DndAwitDXvKy4U · head 9573871084ddd8e1b72193a39ee6529df0e64af9 · 2026-10-28T14:58Z

Triage seat, under gap rule (b) (#27062): this head is over 60 minutes old with no record and no claim; a delta re-review from triage's PASS on 24f51618. The review runs at CONTRACT_REVIEW_TIER. If the lane claims or posts a record on this head first, triage withdraws.

Generated by Claude Code

objectstack-fleet · 2026-10-28T15:10:41Z

Contract review

Served-tier: CONTRACT_REVIEW_TIER
Head-sha: 9573871084ddd8e1b72193a39ee6529df0e64af9
Local-runs: none
Scope: packages/spec/src/conversions/registry.ts, packages/spec/src/conversions/walk.ts, packages/spec/src/migrations/entries/retired-keys/18.ui__Page__slots.tabs.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-details-beside-tabs-refused.ts, packages/spec/src/migrations/entries/semantic/18.page-slots-tabs-retired.ts, packages/spec/src/migrations/registry.ts.template, packages/spec/src/ui/page.zod.ts, docs/adr/0133-org-management-open-basics.md, skills/objectstack-ui/rules/pages.md + packages/spec/src/migrations/types.ts (SemanticMigration.conversionIds contract :113–134; MigrationStep.conversionIds :168), packages/spec/src/migrations/migrations.test.ts (the conversionIds link pin :80–107, step-membership pin :125–133, census pin :159–176), packages/spec/src/migrations/registry.ts.template :3477 (conversionIds mapped off CONVERSIONS_BY_MAJOR[18] ids) and the STEP18 fragment :2590–2602, packages/spec/src/conversions/registry.ts :7929 (id: 'page-slots-tabs-removed', toMajor: 18), the 25 sibling entries/semantic/18.*.ts files carrying conversionIds (key placement precedent), scripts/check-skills-token-ratchet.mjs :845 (pages.md baseline 5692) and .github/workflows/lint.yml :5446–5452 (the ratchet runs in Type Check · source gates), skills/objectstack-ui/evals/ (no slot/tabs eval at this head), packages/spec/src/shared/retired-key.ts, packages/spec/src/shared/strict-object.ts, packages/spec/src/ui/page-slots-tabs-retirement.test.ts, packages/spec/src/shared/retired-key-removal-version.pin.test.ts, packages/spec/liveness/page.json, packages/spec/src/data/field.zod.ts (relatedList), packages/platform-objects/src/pages/record-page-header-actions.test.ts, packages/platform-objects/src/pages/sys-user.page.ts, packages/platform-objects/src/pages/sys-organization.page.ts, packages/cli/test/i18n-section-coverage.test.ts, examples/app-showcase/src/ui/pages/project-detail.page.ts, ui/packages/app-shell/src/views/RecordDetailView.tsx (pageColumnsEligible), docs/qa/platform-checklist/areas/identity-auth.json (rev 6), docs/qa/platform-checklist/areas/studio-authoring.json (rev 4), .changeset/27053-page-slots-tabs-retired.md, card #27059 (the owed skills/evals follow-up, open, pm:blocked)

Inputs: card #27053 (body; claim 6465037594; os-dev-report 6467336687; seat ruling 6467656574), PR #27219 body + 72-file list, the net diff against main at the merge-base cdec7473ef (net.diff), the net diff at the prior head 24f51618 (net-old.diff), the prior record 6470156834 (PASS on 24f51618), the seat ACCEPT 6470956519 and armed 6472080543 on this head, review 5684520620 (os-zhuang, APPROVED, commit 9573871084), and the check-runs on 9573871084: 70 runs, 61 success, 9 skipped (Auto Label, Check PR Size, UI Published Artifacts, Temporal Conformance, Dogfood Verify CLI, Dogfood Regression Gate matrix, Packed-tarball smoke, 18.x upgrade harness, Schedule gate — opt-in or path-skipped), 0 failure, 0 pending. Green include Lint & Repo Gates, every Lint · gate (PM dispatch-gates, docs & code guards, ESLint, CI/release/engine, repo & tooling, the three scoped), PR-body guards ×2, Check Changeset ×2, Governed Surface Queue Guard ×2, Spec property liveness, Type Check source/consumer/debt-ledger/workspace/UI, Build Core/Docs/UI Deps, Build & E2E UI, Live E2E UI, Dogfood Regression Gate, Test Core 1–8/8 and Test UI 1–16/16. Nothing stated here was run locally.

The moved head — delta `24f51618` → `9573871084`

Commits since the prior record: exactly one, 9573871084 ("Join page-slots-tabs-retired to its conversion; move the skills bullet"), a single-parent commit on 24f5161872; no merge, so no merge to recompute. Two files, +3/−3, both in hits.
Net diff old vs new: net-old.diff and net.diff differ in exactly those two files: 18.page-slots-tabs-retired.ts gains the line conversionIds: ['page-slots-tabs-removed'], (hunk header +51 → +52), and in skills/objectstack-ui/rules/pages.md the two-line bullet leaves the kind:'html'/'react' section (old :138–140, a lone - list item plus its blank line) and lands as two prose lines at :40–41 at the end of the "Templates & Regions" paragraph. The file list is unchanged (72 files). Every other hits hunk is byte-identical to the one measured at 24f51618.
main vs the head: origin/main is at 5b5cff15a5; the merge-base is still cdec7473ef. git merge-tree --write-tree prr-27219 origin/main exits 0 (tree 73fb4154b8…): no conflict. The order-78 / order-112 ties with feat(spec,app-shell)!: a form view's type is its layout only — simple, tabbed or wizard; split, drawer and modal are retired with pane, splitSize and splitResizable #27229 remain legal (ascending order, ties by id), as the prior record found.
Prior should-fix ② (Skills-docs: body line) — resolved exactly. The PR body now carries Skills-docs: follow-up #27059 as the third declaration line (body :96, after Clause-② :94 and Metadata-protocol :95). Checked against the tree: skills/objectstack-ui/evals/ has no file mentioning slotted, slots or tabs at this head, so updated here would be false; skills(objectstack-ui) + docs: teach the record page as it is now — object roles produce the page, six slotted slots, kind: full as the escape hatch, drawer default with page by declaration, the column tiers and the reader's own toggles — and stop naming tabs, modal / split / popover #27059 is open, pm:blocked, domain:ui, and its title owns exactly this teaching ("six slotted slots, kind: full as the escape hatch … stop naming tabs"). follow-up #27059 is the true value.
Prior should-fix ④.2 (D3 conversionIds join) — resolved exactly. 18.page-slots-tabs-retired.ts :51 now reads conversionIds: ['page-slots-tabs-removed']. Against the contract in types.ts :113–134: the id must name a conversion that the entry's own step or an earlier one replays. Step 18's conversionIds is derived from the ids of CONVERSIONS_BY_MAJOR[18] (template :3477) and page-slots-tabs-removed is toMajor: 18 (registry.ts :7929), so it is replayed by step 18; migrations.test.ts :80–107 refuses a dangling link and :125–133 pins step membership — Test Core 1–8/8 green on this head. The contract's reading condition ("add an id only after … confirming that it judges that conversion's output") is met by the entry's own reason: it judges the strip's output (a curated strip lost by the strip; whether the page should become kind: 'full'). Key placement (after acceptanceCriteria) matches all 25 sibling step-18 entries that carry the field.
Prior should-fix ④.3 (skills bullet placement) — resolved, at a different spot than the record suggested. The prior record named the full / slotted row of the kind table (now :132) as the natural home; the dev put it at the end of "Templates & Regions" (:40–41), directly after the sentence on what a kind: 'full' page's regions may hold. The finding's substance — the rule sat under the "AI-authored source pages — kind:'html' and kind:'react'" heading where a reader of structured-page rules would not look — is fixed: it now sits in the structured-page section, beside the kind: 'full' region rules it steers an author toward, and kind: 'full' is already vocabulary in that paragraph (:37). Accepted as resolved; where exactly in the structured section it lives is skills(objectstack-ui) + docs: teach the record page as it is now — object roles produce the page, six slotted slots, kind: full as the escape hatch, drawer default with page by declaration, the column tiers and the reader's own toggles — and stop naming tabs, modal / split / popover #27059's restructuring to settle.
Earlier judgments (①.1–10, ②, ③, ④.1–8 of record 6470156834) re-derived on this head below rather than inherited; the hunks they measured are unchanged except the two above.

① Derived judgments

PageSchema.slots.tabs: retiredKey(PAGE_SLOTS_TABS_RETIRED) (page.zod.ts, inside the .optional() strict slot map, the assignedProfiles precedent). Input type never; parse issue invalid_type / expected: 'never' at slots.tabs with the card's prescription (roles/relationship flags; relatedList / relatedList: 'primary' on the child's reference field; slots.details for the Details body; kind: 'full'), the house os migrate meta --from 17 sentence, no #NNNN; removal version 18.3.0 agrees with retiredAfter: '18.2.0' and the other 18.x tombstones in the file. relatedList: 'primary' is real on the reference field (field.zod.ts, ADR-0085). Right.
related / relatedLists move from aliases to guidance with the same remedy: an alias must point at a key the shape accepts, and tabs no longer is. Right.
checkPageSlotPair removed with its .superRefine, the @objectstack/spec/ui export (api-surface / export-origins regenerated; check:generated is inside the green Lint family), page-slot-pair.test.ts, and the two objectui re-attachments; the check-exports pin asserts zero declarations and zero attachments. Right; the public removal is under the changeset's BREAKING banner (②).
D2 page-slots-tabs-removed (registry.ts :7929–7986; toMajor: 18, retiredFromLoadPath: true, retiredAfter: '18.2.0', surface: 'page.slots.tabs', order: 78): mapPages + stripKeys(slots, ['tabs'], …), copy-on-write, one (removed) notice at pages[i].slots.tabs, idempotent; a tabs-only page becomes slots: {} (the synthesized default), disclosed in docblock, fixture, D3 reason, changeset and body. Stored-row replay, idempotence and load-path retirement pinned. Right.
D3 page-slots-tabs-retired absorbs page-slots-details-beside-tabs-refused (deleted; its order-92 STEP18 fragment replaced by one keyed page-slots-tabs-retired at order 112). It now also carries conversionIds: ['page-slots-tabs-removed'], the join the type contract owes for an entry that judges a conversion's output. Right; the prior omission is closed.
RETIRED_KEYS_BY_MAJOR[18] += ui/Page:slots.tabs, the JOIN row in the removal-version pin, the liveness slots note (six), vitest.repo-tests.json listing the new repo test. Right.
Synthesizer and host: buildDefaultPageSchema drops the tabs branch; pageColumnsEligible = !assignedPage && !namedPageBlocked. Every slotted page keeps the synthesized strip and measures the column tiers; a full page draws as authored. Right.
@object-ui/types: PageSlotMap.tabs?: never, both superRefine(checkPageSlotPair) attachments gone; Studio SLOT_ORDER six. Right.
Four producers → kind: 'full', faithful to the synthesizer's order; the two platform pages' page:header action ids are pinned to the object's record_header ∪ record_more declarations (record-page-header-actions.test.ts, non-vacuous guard). Right.
Tests/fixtures moved to details or regions[0].components[]; the tree-scoped absence pin (five roots, four file kinds, anti-vacuity controls, reasoned exclusions) refuses any slots frame holding tabs / related / relatedLists. Right.
Skills file: one bullet added to skills/objectstack-ui/rules/pages.md and one same-meaning tightening elsewhere in the file, now placed in "Templates & Regions". The ratchet baseline for the file is 5692 (check-skills-token-ratchet.mjs :845); the move drops the - marker, the continuation indent and a blank line, so the count cannot have risen; the gate runs in Type Check · source gates, green on this head. Right.

② Semver level

.changeset/27053-page-slots-tabs-retired.md (unchanged in the delta): minor ×7 (@objectstack/spec, @objectstack/platform-objects, @objectstack/plugin-approvals, @object-ui/types, @object-ui/plugin-detail, @object-ui/app-shell, @objectstack/skills), ! title, Clause-②: no (narrowing), Metadata-protocol: narrowed, the adr-0087: registered page-slots-tabs-removed, page-slots-tabs-retired marker, BREAKING banner with a FROM → TO line and the house migrate sentence. Matches what the diff publishes: narrowed ⇒ minor is the rule Check Changeset enforces (green ×2); the main precedent 25767-retire-16x-retirement-kits (spec minor, no (narrowing), narrowed) holds. Packages touched only in tests (@objectstack/lint, @objectstack/cli, @object-ui/console) owe nothing; @objectstack/example-showcase is private; @object-ui/app-shell being private: true with a line is tolerated and matches main. The body's Clause-②: no (narrowing) and Metadata-protocol: narrowed are right: nothing new is admitted, three spellings are refused, one export is removed. The third declaration line owed for a narrowed kind is now present and true: Skills-docs: follow-up #27059 (delta section). No finding.

③ Boundary flags

premise_still_valid: false (four in-repo producers authored slots.tabs): answered in the tree (four faithful kind: 'full' rebuilds, pinned for the two platform pages) and on the card by seat ruling 6467656574, which follows the maintainer's keep/narrow/retire ruling (card ruling 5) and needs no new maintainer decision. Closed.
open_questions[0] — HotCRM's account page not measured here. The card's source line says it authors header + discussion only; the D3 reason restates it; a stored row that did carry tabs is stripped on read by D2 and listed by os migrate meta --from 17; the D3 entry, now joined to the conversion, is reported beside each such edit. Nothing outstanding.
Dev gates not measured locally (check:skill-examples, check:dual-build-cjs-loads, teardown-shape self-test on a shallow clone): the Lint · family, Build Core/UI Deps and Build & E2E UI are green on this head. Closed by the check-runs.
Tier H (docs/adr/**, skills/**): Governed Surface Queue Guard green ×2. os-zhuang's APPROVED review 5684520620 is on this exact commit, which is the governed approval the seat's armed comment names. State drift, not a finding: the PR body still says "this PR is a draft and is not queued or armed", but the PR is draft: false, mergeable_state: clean at read time, and auto_merge reads null on the PR object although the armed comment says automerge_enable was issued — the seat should confirm the arm took, since the body and the PR object disagree with the comment. Neither moves the head.
Skills-docs: line — present and true (②). Closed.
main moved past the merge-base (to 5b5cff15a5): clean merge-tree, legal order ties. Holds.
ADR-0133 inline Amended (2026-10-28) paragraph unchanged since the prior head; its first consumer identity-auth.json (rev 6, source[1] pin moved) and studio-authoring.json (rev 4, step PUTs slots.tabs, refused invalid_type) unchanged. Holds.

④ Design judgment

The narrowing against the published spec on main: a slotted page's tabs slot was a declared capability whose only consumer delivered it by dropping every synthesized tab and opting the page out of the column tiers. The tree retires it at the spec (tombstone), repairs data at rest (D2 strip, disclosed), names and now joins the one judgement a walker cannot make (D3 + conversionIds), removes the pair check that existed only because of the slot, and leaves the capability an author still needs — its own strip — on kind: 'full', which the spec already declares and the four rebuilt producers exercise. Nothing declared is taken away without its replacement being real.

(1) page.zod.ts — against card Target 1 and its first consumers (definePage, PageSchema.safeParse, the @object-ui/types mirror). Unchanged since the prior head. Tried to break it again on this head: tabs: [] refused (pinned); kindless page refused (pinned); related: not steered into the tombstone (pinned); six live slots parse together (control); relatedList: 'primary' exists where the prescription points. Holds. No finding.

(2) registry.ts (D2) and 18.page-slots-tabs-retired.ts (D3) — against ADR-0087 and the spec-property-retirement kit. D2 unchanged and holds (①.4). D3 now carries the join. Tried to break the join: (a) a dangling id — page-slots-tabs-removed is registered at registry.ts :7929 and toMajor: 18, so step 18's derived conversionIds replays it and the migrations.test.ts :80–107 pin passes (Test Core green); (b) a wrong-major id — same step, so :125–133 is satisfied; (c) over-linking — the contract forbids deriving an id from prose that merely names an analogue; here the entry's reason is the judgement over this conversion's own applied edits (a curated strip the strip loses; kind: 'full' or not), so the link is the one the contract wants; (d) a chain printer now shows the TODO beside each pages[i].slots.tabs (removed) edit, which is the behaviour the prior ④.2 was missing. Holds. ④.2 closed, no new finding.

(3) walk.ts — comment-only, unchanged: path examples moved to slots.details; the #6776 history says the four sites were slots.tabs, retired since, with stored copies still walked before the order-78 strip. True of the tree. Holds.

(4) registry.ts.template — unchanged: the STEP18 fragment keyed page-slots-tabs-retired, order 112, names the slot, the six-slot map, the alias refusals, the three remedies, the pair refusal leaving, and the D2 and D3 ids (:2590–2602). Step 18's conversionIds is derived from CONVERSIONS_BY_MAJOR[18] (:3477), which is what makes the new D3 join valid without a template edit. src/migrations/registry.ts is generated from this template plus entries/, so no generated copy is owed. Holds.

(5) 18.ui__Page__slots.tabs.ts — unchanged: export const entry = 'ui/Page:slots.tabs', the nested-key form RETIRED_KEYS_BY_MAJOR reads (pinned). Holds.

(6) 18.page-slots-details-beside-tabs-refused.ts (deleted) — unchanged disposition: its replacement told the author to move record:details INTO slots.tabs, now refused; absorption recorded in the surviving entry; the pin asserts the id is gone. Holds.

(7) docs/adr/0133-org-management-open-basics.md — unchanged: one dated inline amendment, accurate against sys-organization.page.ts at this head (mechanism slotted → full; Members/Invitations/Teams strip and Members-at-index-0 unchanged) and its first consumer identity-auth.json. Holds. No finding.

(8) skills/objectstack-ui/rules/pages.md — against the spec it teaches on main (relatedList / relatedList: 'primary' on the reference field; kind: 'full') and card Target 3 ("the skill bullet"). The text is unchanged and accurate. Placement now: the last two lines of "Templates & Regions" (:40–41), after the kind: 'full' region-refusal sentence — a reader looking up how a slotted record page gets its strip finds it where structured pages are described, and the sentence sends them to the two real remedies. Tried to break it: the paragraph is about template / regions, and the new sentence is about slots — a slight subject shift inside one paragraph, but kind: 'full' is already named there and the sentence's second clause is itself a kind: 'full' instruction, so it reads as the paragraph's closing contrast rather than an intrusion. The card names no eval prompt for this bullet; the eval is #27059's (Skills-docs: follow-up #27059). Token ratchet: baseline 5692, the move only removes markup, gate green. Holds. ④.3 closed; no new finding. Should-fix (already owed to #27059, nothing for this PR): the slotted-page rules deserve a section of their own, not a closing sentence of the regions paragraph.

No must-fix. No new should-fix on this head. The three prior should-fixes are closed in the tree and on the body.

Implemented-by: claude/project-thread-tc4fe5
Reviewed-by: session_01AavokzJ5DndAwitDXvKy4U

VERDICT: PASS

Generated by Claude Code

Conflicts: - ui/packages/plugin-detail/src/synth/buildDefaultPageSchema.ts: main retired the `slots.tabs` slot (#27219) while this branch rewrote the `discussion` slot to the timeline stream and deleted `rightRail`; the resolution keeps main's no-tabs-slot doc and branch, this branch's timeline-stream `discussion` doc, and no `rightRail`. - ui/packages/app-shell/src/views/RecordDetailView.tsx: the `@object-ui/plugin-detail` import line takes the union of main's queue claim imports (#27264) and this branch's timeline/arrangement imports. Claude-Session: https://claude.ai/code/session_cdb7a717def74e719988f089fbaa18ff Co-Authored-By: Claude <noreply@anthropic.com>

…etirement on main made dead main retired `slots.tabs` (#27219) and pins the stray-slot case itself ("reads no `tabs` slot"); this branch's case that a `discussion` slot beside an authored `tabs` slot stays under the strip pins a path that no longer exists. The timeline-stream `discussion` slot stays pinned by "discussion slot overrides even when hideDiscussion is true". Claude-Session: https://claude.ai/code/session_cdb7a717def74e719988f089fbaa18ff Co-Authored-By: Claude <noreply@anthropic.com>

claude added 10 commits October 28, 2026 09:17

wip(spec): retire page slots.tabs — tombstone, alias refusals, D2/D3

ebf513e

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

wip: move the four slots.tabs producers to kind: 'full'; test fixtures

4fa94f8

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

wip: ui synthesizer, types mirror and Studio slots; producer tests; A…

dee4b1a

…DR-0133 amendment, QA pins, skills bullet, changeset Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

wip: lint fixture order; plugin-detail stray tabs slot test

a735eec

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

Merge remote-tracking branch 'origin/main' into claude/project-thread…

ab8eb29

…-tc4fe5

docs(ui): the pages guide drops the tabs slot

a64e5ca

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

chore(spec): regenerate api-surface, export-origins and references

f8beec6

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

test(console): the served approvals page copies follow the full-page …

dd93f77

…rebuild; preview slot fixture Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

test(platform-objects): type the envelope control's region read

bc41b6b

Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

test(spec): the slots.tabs absence pin runs in the repo project; docs…

ca5556e

… avoid the reserved word Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

objectstack-fleet Bot assigned os-support-ai Oct 28, 2026

objectstack-fleet Bot requested review from hotlong and os-zhuang October 28, 2026 11:03

github-actions Bot added the size/xl label Oct 28, 2026

objectstack-fleet Bot requested a review from os-support-ai October 28, 2026 11:03

github-actions Bot added documentation Improvements or additions to documentation protocol:system protocol:ui tests tooling labels Oct 28, 2026

objectstack-fleet Bot mentioned this pull request Oct 28, 2026

spec: a slotted record page keeps six slots — header, actions, alerts, highlights, details, discussion — and the tabs slot is retired; a page that needs its own tab strip is a kind: full page #27053

Closed

claude added 2 commits October 28, 2026 11:22

test: pin the full platform pages' header action ids to the object; s…

3b2e432

…ection-coverage fixture is a full page Co-Authored-By: Claude <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01WGufzff3QiP6kmtajVnKnq

Merge remote-tracking branch 'origin/main' into claude/project-thread…

24f5161

…-tc4fe5

objectstack-fleet Bot mentioned this pull request Oct 28, 2026

feat(formula, plugin-email, lint)!: a mail template hole whose formatter the whitelist does not name is refused by name, never rendered raw #27217

Merged

objectstack-fleet Bot mentioned this pull request Oct 28, 2026

[PM seat] triage (objectstack-wide) — 🟢 Routine · session_01AavokzJ5DndAwitDXvKy4U · trig_01NcnCtMS2tH46nUg1a5TaMJ (hourly) · state = body + the round records newer than it #6015

Open

os-zhuang approved these changes Oct 28, 2026

View reviewed changes

objectstack-fleet Bot marked this pull request as ready for review October 28, 2026 14:44

objectstack-fleet Bot enabled auto-merge October 28, 2026 14:44

objectstack-fleet Bot added this pull request to the merge queue Oct 28, 2026

Merged via the queue into main with commit 6dbff8e Oct 28, 2026
70 checks passed

objectstack-fleet Bot deleted the claude/project-thread-tc4fe5 branch October 28, 2026 15:07

github-actions Bot mentioned this pull request Oct 28, 2026

4-hourly full run: red on main (Lint & Type Check) #27319

Closed

Navigation Menu

Repository navigation

feat(spec,ui)!: retire the record page's slots.tabs; a page that needs its own strip is kind: 'full' - #27219

feat(spec,ui)!: retire the record page's slots.tabs; a page that needs its own strip is kind: 'full'#27219
objectstack-fleet[bot] merged 13 commits into
mainfrom
claude/project-thread-tc4fe5

objectstack-fleet Bot commented Oct 28, 2026 •

edited

Loading

Uh oh!

github-actions Bot commented Oct 28, 2026 •

edited

Loading

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

objectstack-fleet Bot commented Oct 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

What changes

Why

Acceptance notes

Verification

Human Minimum

Uh oh!

github-actions Bot commented Oct 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

📓 Docs Drift Check

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Contract review

① Derived judgments

② Semver level

③ Boundary flags

④ Design judgment

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

席位核对 · ACCEPT · domain:spec seat 1 (#6017) · session session_01WGufzfhobWEgUUrjz4rrbh · 2026-10-28T11:51Z

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Contract review

The moved head — delta ca5556e0 → 24f51618

① Derived judgments

② Semver level

③ Boundary flags

④ Design judgment

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

席位核对 · ACCEPT · domain:spec seat 1 (#6017) · session session_01WGufzfhobWEgUUrjz4rrbh · 2026-10-28T13:57Z

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

席位核对 · armed · domain:spec seat 1 (#6017) · session session_01WGufzfhobWEgUUrjz4rrbh · 2026-10-28T14:43Z

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Uh oh!

Uh oh!

objectstack-fleet Bot commented Oct 28, 2026

Contract review

The moved head — delta 24f51618 → 9573871084

① Derived judgments

② Semver level

③ Boundary flags

④ Design judgment

Uh oh!

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

objectstack-fleet Bot commented Oct 28, 2026 •

edited

Loading

github-actions Bot commented Oct 28, 2026 •

edited

Loading

席位核对 · ACCEPT · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T11:51Z

The moved head — delta `ca5556e0` → `24f51618`

席位核对 · ACCEPT · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T13:57Z

席位核对 · armed · `domain:spec` seat 1 (#6017) · session `session_01WGufzfhobWEgUUrjz4rrbh` · 2026-10-28T14:43Z

The moved head — delta `24f51618` → `9573871084`