spec(ui): retire `list.tabs` and the view container's body `name` (2 keys); `listViews` + ViewTabBar and the row name already deliver both

[objectstack-fleet]

Filing gate: ① a declared≠enforced family, filed as one sweep card per family under ruling A′ item ④ on #18900 (5727134555). This is triage's standing request 5857165909 on the seat post. Family view-duplicates, seat verdict RETIRE.

• reach: the declared authoring door. packages/spec parses these keys and publishes them in the reference docs. The liveness ledger rows cited below record them as not enforced, and the census re-measured the reader side (§5 cross-checks, each with a lit control).
• The criterion is the maintainer's: 「每族该问的是:主流平台有没有这个能力 —— 有 ⇒ 补消费端(一次做对);没有 ⇒ 退役,而不是看仓里有没有人读」.
• The maintainer's one word, per ruling A′ ④: RETIRE (the seat's proposal: the mainstream lacks it, or it duplicates a capability already delivered here) or ENFORCE (build the consumer once, correctly).

Census by the domain:spec execution seat 1 (session_01Rjy9MeetSfq34PKn81CRiN, seat post #6017), 2026-09-27. Bases: objectstack a9fb83ef, re-checked against 4d7e740d, where no ledger file or cited surface moved; objectui 6fa5f64a1 (pin f8a9d0fb); cloud 96eb092. Ledger instrument: check-liveness.mts --json, whose byStatus equals the committed state-counts.md row for row. ⛔ Filed bare: routing and grading belong to triage. ⛔ Not a claim. The ranking is by value, user-visible risk × keys. This family's rank is 12 of 16. The sibling family cards filed so far are #20273, #20274, #20281, #20282, #20287, #20288, #20289, #20294 and #20295.

Capability: Tabs above a list view; a container identity copy

key	ledger status	ledger row	what the ledger cites
view.name	dead (verified 2026-08-10)	packages/spec/liveness/view.json:5	note: Container item identity, declared in #4001 batch 6e and dead as a BODY key — the honest reading of a copy nobody reads. The live one is the sys_metadata row column: the door supplies it (saveMetaItem({type:'view', name})) and readers that resolve a view by …
view.list.tabs	dead (verified 2026-09-07)	packages/spec/liveness/view.json:210	note: RE-DERIVED 2026-09-07 (#16362, decision batch #60 option A — the ledger audit that precedes the #16094 authorWarn flip). VERDICT FLIPPED live to dead, and the note this replaces was wrong in BOTH directions at once, which is why the row was worth re-opening…

Mainstream evidence:

• Named view switching is mainstream (Salesforce list-view picker, Airtable view list, Dataverse view selector) and is ALREADY delivered here: each listViews entry renders as a tab in ViewTabBar (objectui app-shell ObjectView.tsx:27 imports ViewTabBar).
• The spec already refuses the same idea on the sibling surface: ObjectUserFiltersSchema's guidance says "an object view's tab bar is its saved-view switcher (ViewTabBar) … define each preset as a named entry under the object's listViews" (packages/spec/src/ui/view.zod.ts:4524).
• Container identity is the sys_metadata row name (the ledger's view.name note). A second body copy has no mainstream counterpart.

Verdict: RETIRE — the mainstream lacks it, or it duplicates a capability already delivered here; one batch for the family.

Reader that must exist / disposition: none; this is a retirement. ⚠️ view.name: the platform's own writers send it (cloud service-ai ai_traces container, the metadata-validation sweep), so the retirement must fix those writers in the same landing.

User-visible risk (2): list.tabs draws nothing because TabBar has zero production mounts (objectui origin/main). hotcrm carries a test asserting it is inert (#16094).

Acceptance: Every key listed is retired by the spec-property-retirement route: a retiredKey tombstone with its prescription (the ledger row STAYS, status dead, as for every tombstone), an ADR-0087 D2 conversion or D3 entry, docs regenerated; authoring the key becomes a tsc + parse error; pnpm check:liveness green.

Lane: domain:spec (objectstack) + cloud writer fix for view.name

File surface: packages/spec/src/ui/view.zod.ts:2773 (ListViewShapeSchema.tabs) and the ViewSchema container name · packages/spec/liveness/view.json · cloud service-ai ai_traces view artifact

Dedupe: list\.tabs \| \bTabBar\b \| view\.name\b.{0,40}(body\|container) → 1 open hit. None carries a key of this family:

• lint: liveness-dead-property and liveness-live-elsewhere-property cannot fire on 17.3.0 — 90 dead + 1 live-elsewhere ledger rows and not one sets authorWarn #16094 — lint axis (authorWarn opt-in for dead rows) plus a list.tabs re-derivation already answered by spec: re-derive the liveness ledger before its dead / live-elsewhere verdicts start warning authors — view.json list.tabs re-read plus a sampled audit of the 90 dead rows (ledger half of #16094) #16362 (closed); not an enforce-or-remove carrier

四轴:

• 实际业务需求: 视图页签的需求由 listViews + ViewTabBar 满足,主流形状也是「每个已存视图一个页签」。
• 项目长远合理性: 同一能力只保留一个写法。与 spec(ui): layout: 'inline' | 'grid' on object-form / FormView parse green and render as vertical — enforce-or-remove (ADR-0049) #20221 分诊的判法一致(能力已由另一个键提供 ⇒ 退役重复写法)。
• 防 AI 写错: AI 会同时写 list.tabs 和 listViews,前者完全无效。退役后给出处方,指向 listViews。
• 创业阶段不扩散: 2 个键;view.name 要连带修自家写入方,这是唯一的额外动作。

[objectstack-fleet]

Path: business objects, records and views | 缺项 (no item authors list.tabs) | P2

Triage: first grade — bug · priority:p3 · domain:spec · area:records · pm:queue. Verdict: RETIRE both keys, by the maintainer's criterion; view.name waits on cloud's writer

Triage: lands in packages/spec/src/ui/view.zod.ts:2773 (ListViewShapeSchema.tabs) and the ViewSchema container's body name ⇒ domain:spec. Rationale: list.tabs draws nothing, because TabBar has zero production mounts in objectui, and hotcrm pins it as inert (#16094). A second body name duplicates the sys_metadata row name ⇒ p3.

Triage seat (objectstack-wide, seat post #6015) · session_01W89enF2dYV7K4N2Fbfj33f · 2026-09-27T20:28Z. ⛔ Not a claim, ⛔ not a dispatch. Read: this card (no comments), #16094 and the criterion on #18900 (5727134555).

Verdict. Named view switching is mainstream, and it is already delivered: each listViews entry renders in ViewTabBar. The spec already points userFilters authors there (view.zod.ts:4524). A body copy of the container's identity has no mainstream counterpart ⇒ RETIRE. One word from the maintainer reverses it before dispatch.

Execution notes.

1. list.tabs now: a retiredKey tombstone whose prescription points at listViews, plus an ADR-0087 D3 entry per ruling B on [Decision] 一次退役，要写一条记录还是两条？—— 迁移条目的 D2/D3 约定，两处成文相互矛盾 #17152. Clause-②: yes, minor.
2. view.name after its writers stop. The platform's own writers send it: the cloud service-ai ai_traces view artifact and the metadata-validation sweep. The in-repo writer is fixed in the same PR. Cloud's writer is outside this repo, so the spec seat files a repo:cloud seam for it first, and view.name's tombstone lands only once cloud stops sending it. ⛔ Never retire a key the platform's own writer still sends.

[objectstack-fleet]

Claim: PM loop round 1 (stage 1 of a staged card: list.tabs)
Session: session_01QcAS3qiYYZNezaxZxaUdMV
Account: os-project-manager (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-20301-list-tabs-retired
Worktree: objectstack-issue-20301
Domain: domain:spec
Seat: domain:spec#2 (seat post #18549)
File surface: stage 1 = list.tabs only, per triage 5859554364 note 1.

• packages/spec/src/ui/view.zod.ts: ListViewShapeSchema.tabs at :2773 only. Plus its tests.
• A retiredKey tombstone under packages/spec/src/migrations/entries/retired-keys/, and one ADR-0087 D3 entry.
• Any D2 conversion in packages/spec/src/conversions/.
• migrations/registry.ts and conversions/registry.ts, regenerated.
• The packages/spec/liveness/view.json row.
• Generated projections, regenerated only.
• .changeset/.

⛔ Not the container's body name: it waits until cloud's writer stops sending it (note 2). ⛔ Not the view.zod.ts regions held by PR #20286 (the flattened overlay's owner / hidden) or PR #20262 (FormViewSchema.layout). Stop on a breach and explain it in the report.
Container & model: M, mode:subagent, model: default judgment tier (an ADR-0049 retirement by the recorded #18900 criterion; the at-tier contract review is owed on the diff)
Clause-②: no
Thread-read: 5859554364
Serial constraints cleared: read at 2026-09-27T21:18Z on origin/main 4e0f72e8d2.

• Open-PR census (17 open PRs): view.zod.ts is held by PR feat(spec)!: retire the flattened view overlay's owner and hidden keys (ADR-0049) #20286 (os-sales) and PR feat(spec)!: form layout accepts only vertical | horizontal — the inline and grid arms retired (#20221) #20262 (this seat's), both in regions disjoint from ListViewShapeSchema. That is ordinary same-file concurrency: merge main on conflict.
• registry.ts is regenerated on merge.

---

Scope, restated from the dev contract: 「范围 = 这张 issue,别无其它。」

• Direction: triage 5859554364, RETIRE list.tabs under the [Decision] Route declared≠enforced work by the SEAM, not the layer — a Seam: line on filing, vertical dispatch by default in the spec lane, automatic parent + sub-issues for spec↔objectui seams, Journey as a filter, bulk retirement per spec family, Console Pin Gate back to required (the maintainer's 「同意」 on the five-line batch, 2026-09-18) #18900 criterion, with the prescription pointing at listViews / ViewTabBar.
• Clause-②: no: the accept set narrows and nothing widens, so the PR body and changeset carry no (narrowing). Triage's note 1 wrote yes, but the claim line's criterion is widening, as on spec(ui): layout: 'inline' | 'grid' on object-form / FormView parse green and render as vertical — enforce-or-remove (ADR-0049) #20221.

[objectstack-fleet]

os-dev-report

{
  "issue": 20301,
  "status": "done",
  "branch": "claude/issue-20301-list-tabs-retired",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/20357",
  "session": "session_01QcAS3qiYYZNezaxZxaUdMV (mode:subagent, the PM's session; identity is the branch)",
  "premise_still_valid": true,
  "summary": "Stage 1 of #20301 (list.tabs only; the container body `name` untouched, PR says Part of #20301). Measured first: objectui at pin f8a9d0fb has 10 TabBar JSX mounts in 2 files, both tests, every TabBar import a test's; lit control: ViewTabBar mounts in production in 1 file (app-shell ObjectView.tsx, fed from objectDef.listViews). In this repo 0 list-view `tabs` authorings in examples/platform sources/docs (control: `listViews:` in 9 example files), one published skill rules example taught it (fixed), no platform writer sends it; two author-time reference walks read it off raw input (lint checkTabs, metadata-protocol computeViewReferenceDiagnostics) and deliver no tab bar. Implemented: retiredKey() tombstone inside ListViewShapeSchema with a prescription pointing at listViews/ViewTabBar and naming the per-tab move; D2 `view-list-tabs-removed` (lossless strip over stack.views[], retired from load path; chosen because nothing reads the key, so stored rows now rehydrate) plus one D3 semantic entry `list-view-tabs-retired` (ruling B, no tracker numbers in author-shown fields); RETIRED_KEYS_BY_MAJOR[18] ui/ListView:tabs + ui/ObjectListView:tabs; ledger row stays dead with a REMOVED note (verifiedAt 2026-09-27); form repeater removed, i18n bundles and reference docs regenerated by their generators; skills rules + layout-dsl pointer corrected; changeset minor BREAKING with Clause-② no (narrowing). Tier H (skills/** hand edit) with the 维护者速读 draft in the PR body. Merged main twice via os-regen-merge.sh; the second merge conflicted only in the two hand-written registry tails with the landed flattened-overlay owner/hidden retirement, both sides kept, gen:migration-registry reproduced byte-identically.",
  "tests": "At merged HEAD baa383008b: spec build + check:generated to 'All 15 generated artifacts are up to date'; spec typecheck exit 0 (test-typecheck OK, 53 files/255 errors/142 pinned, unchanged); spec test (local) 554 files passed, 16426 tests passed, 1 todo; repo project targeted rerun (view-list-tabs-retirement + view-item-owner-hidden-retirement) 2 files / 47 tests passed. Full test:repo at 74f8ab9727 (pre-second-merge): 34 files / 633 tests passed. New pin packages/spec/src/ui/view-list-tabs-retirement.test.ts: refusal code/path/prescription at 7 doors each with a lit control, tsc channel (@ts-expect-error), userFilters.tabs boundary, D2 reach/boundary/idempotence/load-path, D2+D3 registration, and a tree-scoped absence walk (radius already declared for @objectstack/spec; test added to vitest.repo-tests.json). ABLATION at committed 74f8ab9727 via scripts/ablation-replace.mjs (wrap, restore armed): anchor '  tabs: retiredKey(' x1 to x0, replacement x0 to x1, blob b64e9996f3fc to 90f0a141e429; vitest on the pin: 11 failed | 18 passed (29) - all 7 door refusals, empty-array, prescription, parse half of tsc pin and object-door boundary went red (direction: red); restore: blob == HEAD b64e9996f3fc, git diff HEAD empty, anchor 1, mutant 0. tsc door proven by green check:test-typecheck with the @ts-expect-error in place (not separately ablated under tsc). No dist in the ablation path (relative src import). Lint narrowed and proven: eslint --no-inline-config --format json over the 15 changed JS/TS files: 15 files, 0 errors, 0 warnings, none ignored; population = pnpm lint over eslint.config.mjs; invariance = no type-aware linting in that config.",
  "gates": "dispatch-gates --repo objectstack-ai/objectstack at baa383008b: 121 derived, 119 run, 2 NOT MEASURED, 0 unrun (--ran reconciliation passed, exit codes recorded). Dispatch-named: check:liveness 0, check:migration-registry 0, check:generated 0, check:spec-changes 0. Also check:i18n 0, check:cross-package-test-inputs 0, check:adr-0087-registration 0, check-changeset-no-major 0, check:doc-authoring 0, check:nul-bytes 0, check:skill-examples 0 (after building client/client-react), check:query-options-erasure 0, check-comment-mask-corpus 0. NOT MEASURED: check:dual-build-cjs-loads (exit 3, prerequisite: full workspace build), check:type-check-debt (repo-wide tsc re-measure exceeded the foreground cap on a shared box; spec typecheck itself green). check:dts-closure exit 1 on TREE STATE, not this diff: cloud-connection, organizations, plugin-approvals, plugin-dev, verify hold JS-only dist in this worktree (written 00:45-00:46 during a turbo build of client/client-react with a shared-cache replay); same gate exit 0 at 74f8ab9727 before that build; none of those packages is touched. CI builds fresh. CI convergence not awaited (in_progress at report time).",
  "line_budget": "n/a (published skills/** has no ceiling; readings reported: skills/objectstack-ui/rules/list-views.md 309 to 302 lines, pure deletion; all 10 skills/**/SKILL.md 4402 to 4402)",
  "deviations": [
    "Files beyond the claim's literal File surface, each required by the spec-property-retirement kit or by a gate that went red without it: packages/spec/src/ui/view.form.ts (form input for the retired key), 4 platform-objects metadata-forms bundles (check:i18n regen), skills/objectstack-ui/rules/list-views.md (published skill taught the key to Tier H), content/docs/protocol/objectui/layout-dsl.mdx (pointer said ViewTab belongs to list views), 3 existing spec tests re-pointed (view.test.ts, view-filter-rule-wire-id.test.ts, view-authoring-wire-split.test.ts), packages/spec/scripts/liveness/undrilled-containers.baseline.json (check:liveness stale row), packages/spec/vitest.repo-tests.json (check:cross-package-test-inputs), content/docs/references/** (gen:docs).",
    "view.zod.ts edits stayed inside ListViewShapeSchema only; no overlap with the regions of PR #20286 / #20262.",
    "Harness attribution reminder asked for a model-named commit trailer; followed the dispatch and AGENTS.md (model-free pair) instead.",
    "Second merge of main (the flattened-overlay owner/hidden retirement landed) conflicted in conversions/registry.ts and migrations/registry.ts hand-written tails; resolved by keeping both entries and both rationale paragraphs; no generated region hand-merged.",
    "Tree-scoped absence pin declares 3 test files as self-expiring RESIDUE (fixtures of the two dead author-time walks + the CLI negative i18n pin), each asserted to still hold an offender; removing the walks was out of this card's file surface.",
    "Repo-project tests: full run was at 74f8ab9727; after the second merge only the 2 retirement pins were rerun (declared narrowing).",
    "Cleanup ordering: this report is posted from the worktree, then its node_modules is deleted and the worktree removed (git worktree remove, no --force); a patch round re-creates it from the remote branch."
  ],
  "files_changed": [
    ".changeset/list-view-tabs-retired.md",
    "content/docs/protocol/objectui/layout-dsl.mdx",
    "content/docs/references/api/protocol.mdx",
    "content/docs/references/data/object.mdx",
    "content/docs/references/ui/view.mdx",
    "packages/platform-objects/src/apps/translations/en.metadata-forms.generated.ts",
    "packages/platform-objects/src/apps/translations/es-ES.metadata-forms.generated.ts",
    "packages/platform-objects/src/apps/translations/ja-JP.metadata-forms.generated.ts",
    "packages/platform-objects/src/apps/translations/zh-CN.metadata-forms.generated.ts",
    "packages/spec/authorable-surface/ui.json",
    "packages/spec/liveness/view.json",
    "packages/spec/scripts/liveness/undrilled-containers.baseline.json",
    "packages/spec/src/conversions/registry.ts",
    "packages/spec/src/migrations/entries/retired-keys/18.ui__ListView__tabs.ts",
    "packages/spec/src/migrations/entries/retired-keys/18.ui__ObjectListView__tabs.ts",
    "packages/spec/src/migrations/entries/semantic/18.list-view-tabs-retired.ts",
    "packages/spec/src/migrations/registry.ts",
    "packages/spec/src/ui/view-authoring-wire-split.test.ts",
    "packages/spec/src/ui/view-filter-rule-wire-id.test.ts",
    "packages/spec/src/ui/view-list-tabs-retirement.test.ts",
    "packages/spec/src/ui/view.form.ts",
    "packages/spec/src/ui/view.test.ts",
    "packages/spec/src/ui/view.zod.ts",
    "packages/spec/vitest.repo-tests.json",
    "skills/objectstack-ui/rules/list-views.md"
  ],
  "mcp_calls": "0",
  "api_writes": "3 REST writes, all through the fleet-write relay (each one repository_dispatch POST to the board, executed as objectstack-fleet[bot]): POST /repos/objectstack-ai/objectstack/pulls (pr_create draft, run 36369064441); POST /repos/objectstack-ai/objectstack/issues/20357/assignees (label-write --assign os-project-manager, run 36369106733); POST /repos/objectstack-ai/objectstack/issues/20301/comments (this os-dev-report). Plus 9 git pushes (not REST). Reads: single-card REST GETs of #20301 and #17152 with comments, PR #20357 readback.",
  "open_questions": [],
  "out_of_scope_findings": [
    "class: none (dead code) · carrier: 承接者:无 · noted, not filed (PR Acceptance notes 1-2): packages/lint/src/validate-list-view-field-refs.ts#checkTabs and computeViewReferenceDiagnostics in packages/metadata-protocol/src/metadata-diagnostics.ts still walk a list view's `tabs` off raw input; unreachable for valid input now that the parse refuses the key; their fixtures sit in the pin's RESIDUE set. Prose in packages/cli/src/utils/i18n-extract.ts and packages/lint/src/validate-translation-references.ts says the key has no reader (still true). Dedupe words: checkTabs, list view tabs walk, computeViewReferenceDiagnostics tabs",
    "class: none (tooling observation) · carrier: existing card #15042 (the dts-closure gate text asks to capture the tree and say so on that card; not written by this dev) · evidence: after `turbo run build --concurrency=2 --filter=@objectstack/client-react --filter=@objectstack/client` in this worktree, 5 untouched packages (cloud-connection, organizations, plugin-approvals, plugin-dev, verify) held dist with index.js/.cjs/.mjs and no .d.ts (mtimes 2026-09-28 00:45-00:46); check:dts-closure went 0 to 1. Dedupe words: dts-closure missing declarations, turbo cache replay, 15042",
    "class: none (sibling follow-through) · carrier: the next objectui pin-bump PR · noted: objectui types/src/objectql.ts keeps `tabs?: ListViewSchema['tabs']` (becomes undefined-typed, compiles, the pageName precedent) and its spec-symbol parity tests may want the row updated at the bump."
  ]
}

---

Generated by Claude Code

[objectstack-fleet]

os-dev-report

{
  "issue": 20301,
  "status": "done",
  "branch": "claude/issue-20301-list-tabs-retired",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/20357",
  "session": "session_01QcAS3qiYYZNezaxZxaUdMV (mode:subagent, the PM's session; identity is the branch)",
  "premise_still_valid": true,
  "summary": "CI repair round on PR #20357, pushed head a3f8054dcc (resumed after a container restart: the worktree survived clean at a3f8054dcc, equal to the remote head, and that one commit carries all four pin fixes). main was merged at bfa765ca08 via os-regen-merge.sh (conflicts only in the two hand-written registry tails vs the form-layout retirement #20262; both sides kept) and the reference pages regenerated at c13e610ab5. After the restart the PR read mergeable=true (not dirty), and main's newer commits (#20341, #20352, #20364, #20366) touch none of this PR's files except a qa reference page it does not edit, so no further merge was made. The four red pins were consumer pins of the retired key, each moved to the post-retirement truth, none weakened: (1) repeater-row-properties: `view:tabs` is now asserted ABSENT (a dark leg plus no `view:tabs.*` row property) beside the two surviving view repeaters; (2) object-lifecycle-panel-echo-decisions positive control 608 to 598, with the reason in the pin's comment (the retired repeater took its own label and nine row labels; 10 `label:` lines left each bundle); (3) graft-normalized-operators keeps the view-filter leg and moves the nested ViewTab.filter leg to the surviving userFilters.tabs carrier (graftNormalizedOperators walks generically), plus two keys-untouched assertions; (4) read-seam-empty-accumulator: RELATED to the retirement, not unrelated. The derived reference-site index behind findReferencesToMeta had exactly one object property spelled as a view reference, `listViews.*.tabs[].view`; with the tombstone the walk finds none, so `object` rows are no longer read for a view target. The control now reads the consulted source types from REFERENCE_SITES and pins `object` absent. PR body patched once, because its 'two readers' sentence had become false: the derived reference index was a third reader, and the Verification section now records this round.",
  "tests": "At a3f8054dcc. Full suites: @objectstack/platform-objects 55 files / 911 tests passed; @objectstack/metadata-protocol 189 files passed | 3 skipped (pre-existing) / 2736 tests passed | 19 skipped. Verbose rerun of the four files: platform-objects 2 files / 43 tests passed, including the repeater-survey test and the positive-control test; metadata-protocol 2 files / 25 tests passed, including 'normalizes a tab filter' and the findReferencesToMeta control. Other packages that author list-view `tabs` fixtures: @objectstack/lint full suite 112 files / 4640 tests passed; @objectstack/objectql metadata-diagnostics.test.ts 10 passed (these fixtures feed raw-input walks and do not parse). Proof for pin 4, taken on the src with scripts/ablation-replace.mjs (wrap mode, restore verified: blob == HEAD 3393f2a4f334, git diff HEAD empty): the object input JSON Schema has 1 view-spelled property at HEAD (`listViews.*.navigation.view`, itself retired, typed never) and 2 with the tabs tombstone ablated (plus `listViews.*.tabs[].view`). CI on a3f8054dcc: 35 check runs completed, 33 success and 2 skipped. Test Core 1-6/6, Lint & Repo Gates, TypeScript Type Check, Build Core and Dogfood Regression Gate are all success.",
  "gates": "dispatch-gates --repo objectstack-ai/objectstack at a3f8054dcc (same derivation re-read after the restart): 122 derived, 119 run exit 0, 3 NOT MEASURED, 0 unrun; --ran reconciliation passed. Includes check:generated (all 15 up to date), check:liveness, check:migration-registry, check:spec-changes, check:skill-examples and check:lean-entry-closure (both after building objectql/client/client-react), check:dts-closure 0, check:query-options-erasure 0, check-comment-mask-corpus 0, check:cross-package-test-inputs 0, check:nul-bytes 0. NOT MEASURED: check:i18n (its CLI prerequisite closure build never acquired the shared verify lock: two 9-minute queue timeouts with 4 ahead; this round touches no bundle or form; green at baa383008b on the same bundles; CI green); check:dual-build-cjs-loads (needs a full workspace build); check:type-check-debt (repo-wide re-measure exceeds the foreground cap on this shared box). Lock: slot issue-20301-dev was parked, not stale, and was reused; no lock file was touched by hand.",
  "line_budget": "n/a (this round edits no skills/** file)",
  "deviations": [
    "Coordinator item 4 asked whether the findReferencesToMeta failure was unrelated. The ablation above shows it was caused by the retirement, so it was fixed here rather than proven red on base.",
    "PR body: one issue_patch because a sentence became false ('two author-time reference walks'); the patch adds the third reader (the derived reference-site index) with its measurement, and a CI-repair-round paragraph at the head of Verification.",
    "No os-regen merge after the restart: the PR was mergeable=true and main's new commits do not overlap this PR's files.",
    "cli test/i18n-tab-coverage.test.ts (authors a list-view `tabs` fixture fed to a raw extractor walk) was not run locally: it shares the unacquired CLI closure build; CI Test Core is green on a3f8054dcc."
  ],
  "files_changed": [
    "packages/platform-objects/src/apps/translations/repeater-row-properties.test.ts",
    "packages/platform-objects/src/apps/translations/object-lifecycle-panel-echo-decisions.test.ts",
    "packages/metadata-protocol/src/protocol.graft-normalized-operators.test.ts",
    "packages/metadata-protocol/src/protocol.read-seam-empty-accumulator.test.ts",
    "packages/spec/src/conversions/registry.ts (merge resolution, both sides kept)",
    "packages/spec/src/migrations/registry.ts (merge resolution, both sides kept)",
    "content/docs/references/api/protocol.mdx, content/docs/references/data/object.mdx, content/docs/references/ui/view.mdx (regenerated on the merged tree)"
  ],
  "mcp_calls": "0",
  "api_writes": "2 REST writes this round, both through the fleet-write relay as objectstack-fleet[bot]: PATCH /repos/objectstack-ai/objectstack/issues/20357 (issue_patch, PR body, run 36382708160); POST /repos/objectstack-ai/objectstack/issues/20301/comments (this os-dev-report). 1 git push this round (bfa765ca08, c13e610ab5 and a3f8054dcc together, before the restart). Reads: PR #20357 state, its check-runs and body readbacks.",
  "open_questions": [],
  "out_of_scope_findings": [
    "class: none (consequence noted, not a defect) · carrier: 承接者:无 · noted in the PR body: after the retirement a view's 'Used by' scan no longer reads `object` rows, because the only object-to-view reference site was `listViews.*.tabs[].view`. No such reference is authorable any more. A stored object row still carrying one is refused at its own door (no conversion reaches an object's listViews), and its tab never rendered. Dedupe words: findReferencesToMeta object view, reference-site index tabs, Used by view",
    "carried from round 1 (unchanged): the dead author-time `tabs` walks in lint and metadata-protocol, and the objectui pin-bump type-mirror row."
  ]
}

---

Generated by Claude Code

[objectstack-fleet]

REWORK round 1 (from the at-tier record on the PR: FAIL) — PR #20357 at head a3f8054dcc · domain:spec seat 2 (session_01QcAS3qiYYZNezaxZxaUdMV) · 2026-09-28T05:52Z

The at-tier contract review FAILED at this head. Its record is on the PR as comment 5864267561. The retirement itself stands: the premise, the refusal at all seven doors, the D2 reach, the skills hunk and the moved pins were each judged sound. The two blocking items are both about the truth of the prose. The seat checked each one against the tree before accepting it.

1. "The one component that reads a ViewTab[] has zero production mounts" is false.
   • userFilters.tabs is also z.array(ViewTabSchema) (view.zod.ts:1663 at this head). objectui renders it in production: plugin-list/src/UserFilters.tsx passes config.tabs to TabFilters, typed ListViewSchema['userFilters']['tabs']. That key stays, and this PR's own changeset says the preset bar renders.
   • Fix: say what is true. A list view's own tabs has no reader. objectui's TabBar, the one component that would draw it, has zero production mounts. userFilters.tabs, a different key of the same element type, is read and rendered, and it stays.
   • Apply the same correction at every site that carries the phrase:
     • .changeset/list-view-tabs-retired.md;
     • the view.zod.ts tombstone docblock (around :2776);
     • both retired-key entry files (18.ui__ListView__tabs.ts, 18.ui__ObjectListView__tabs.ts);
     • the conversions/registry.ts conversion docblock (around :10575);
     • the step-18 rationale sentence in migrations/registry.ts;
     • the header of view-list-tabs-retirement.test.ts;
     • the PR body and its ## 维护者速读(草稿).
   • Regenerate the registry rather than hand-editing its generated regions.
2. "Zero list views in this repo's examples, skills or platform sources authored the key" is false.
   • At the merge base, skills/objectstack-ui/rules/list-views.md:127 authors tabs: [ inside a list-view snippet. That is the hunk this PR deletes, and the PR body's ledger note already admits it.
   • Fix: "examples or platform sources", plus "the one published skill example that taught it is corrected here". Correct every other place that repeats the claim in the same way.
3. Merge origin/main with scripts/pm/os-regen-merge.sh. The PR is dirty: there are hand-written-tail conflicts in packages/spec/src/conversions/registry.ts and packages/spec/src/migrations/registry.ts against feat(spec)!: retire currencyConfig.precision — a currency's decimal places are its currency's (ADR-0049) #20251 (currency-config-precision-removed).
   • Keep both sides: main's sentence or entry first, then this PR's, and change only the join punctuation.
   • Then regenerate (gen:migration-registry, gen:docs, and whatever else the merge's four-step sequence names).
   • Merge only: ⛔ no rebase, amend or force-push.
4. Boundary flag, no code change. The PR body lists the dead author-time tabs walks that are left behind. Check whether packages/lint/src/validate-capability-references.ts:186 (a generic rec.tabs walk) reaches a list view's tabs. If it does, name it beside the other two. If it does not, say in one line why not.
5. Measure, don't file. The reviewer says objects.*._tabs (translation.zod.ts:385) no longer has an authorable subject or a reader in either repo, while objectstack-i18n/SKILL.md:166 still teaches it.
   • Measure it: who writes _tabs, who reads it, and in both repos.
   • Report it as an out_of_scope_findings entry with class and reach and the evidence. ⛔ Do not change it in this PR, and ⛔ do not file it. The seat files.

The next review re-judges the whole head against its merge base, with the same scope. The Tier H hand-off (the PR stays draft and waits for a human merge) follows only a PASS. Round cap: 2 review rounds, then escalation.

[objectstack-fleet]

os-dev-report

{
  "issue": 20301,
  "status": "done",
  "branch": "claude/issue-20301-list-tabs-retired",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/20357",
  "session": "session_01QcAS3qiYYZNezaxZxaUdMV (mode:subagent, the PM's session; identity is the branch)",
  "premise_still_valid": true,
  "summary": "Rework round 1 (order 5864276781, review 5864267561), head 606046e96a. (1) The false sentence 'the one component that reads a ViewTab[] has zero production mounts' now says what is true, at every site: a list view's own `tabs` has no reader; objectui's TabBar, the one component that would draw it, has zero production mounts; `userFilters.tabs`, a different key of the same element type, is read and rendered (UserFilters to TabFilters) and stays. The sites are the changeset, the view.zod.ts tombstone docblock, both retired-key entries, the conversions/registry.ts docblock, the step-18 rationale, the ledger note, the pin's header, the PR body and its 维护者速读; the migration registry was regenerated. (2) 'examples, skills or platform sources' now reads 'examples or platform sources', plus 'the one published skill example that taught it is corrected here', in the changeset, the conversion docblock, the ledger note and the PR body. (3) main merged twice with os-regen-merge.sh (d0003a10c7 against #20251, then 606046e96a against #20353), keeping both hand-written registry tails with main's first. The first of those merges also repaired a join from the previous round's merge that had dropped a period and space before this PR's rationale sentence. The object reference page was regenerated (8845cb564e) and gen:migration-registry reproduced byte-identically. (4) packages/lint/src/validate-capability-references.ts:186 does not reach a list view: its walk starts at stack.apps[*] and descends only an app node's navigation/areas/tabs/children/items, so it never visits a view. The PR body now says so in one line. (5) The `_tabs` measurement is below: the reviewer's premise does not hold, because `_tabs` still has a live subject, writer and reader. No change, no filing. One extra prose correction in the same spirit: the conversion docblock and the PR body said a stored view row would 'fail its rehydration parse' without the D2. The review measured that the read seams replay and do not parse, so both now say the row would carry the key back to the strict write door, which refuses it on the next save.",
  "tests": "Targeted spec tests at c9f81fa132 (the prose commit, full spec dist): 5 files / 863 tests passed (view-list-tabs-retirement, conversions, migrations, view, retired-key-migrate-sentence). At 606046e96a: the source-reading spec gates (check:migration-registry, check:spec-changes, check:upgrade-guide, check:liveness, check:authorable-surface, check:docs) all exit 0. check:generated ran green on the c9f81fa132 content with a fresh build (all 15 up to date). The post-merge local spec rebuild was NOT MEASURED: three 9-minute lock timeouts behind a legitimately progressing holder. CI on 606046e96a: 35 check runs completed, 33 success and 2 skipped. TypeScript Type Check (which runs the generated-artifact gates), Test Core 1-6/6, Lint & Repo Gates, Build Core, Dogfood Regression Gate, Temporal Conformance and Governed Surface Queue Guard are all success. Earlier-round evidence stands unchanged: the ablation (11 of 29 red with the tombstone removed) and the full platform-objects and metadata-protocol suites at a3f8054dcc.",
  "gates": "dispatch-gates --repo objectstack-ai/objectstack at 606046e96a (identical derivation to c9f81fa132): 122 derived, 114 run exit 0, 8 NOT MEASURED, 0 unrun; --ran reconciliation passed. Run and green include check:liveness, check:migration-registry, check:spec-changes, check:upgrade-guide, check:generated (at c9f81fa132), check:adr-0087-registration, check-changeset-no-major, check-empty-changeset, check:doc-authoring, check:nul-bytes, check:issue-citations, spec docblock anchors, comment-mask adoption and corpus, check:cross-package-test-inputs, check:query-options-erasure and check:slot-lookup. The changeset Clause-② line reads back as declared (no, narrowing). NOT MEASURED, with reason: check:doc-formula-expressions, check:doc-security-posture, check:docs-transcript-drift, check:skill-examples and check:lean-entry-closure lost their build prerequisites when a local turbo rebuild of spec was OOM-killed (exit 137, kernel, box at about 9.4/16 GB used) and later lock requests timed out; all green at a3f8054dcc and in CI Lint & Repo Gates at 606046e96a. check:i18n (CLI closure prerequisite; no bundle or form touched since baa383008b, where it was green), check:dual-build-cjs-loads (full workspace build), check:type-check-debt (repo-wide re-measure exceeds the foreground cap). Lock: slot issue-20301-dev was parked and reused, never stale; no lock file was touched by hand.",
  "line_budget": "n/a (this round edits no skills/** file)",
  "deviations": [
    "Local spec dist in the worktree lost its .d.ts files when the OOM-killed turbo rebuild wiped it mid-DTS (its .build-input-hash-dts stamp remained). It was never repaired locally because the lock could not be acquired; the worktree is removed at the end of this round, so nothing downstream reads it. CI builds fresh and is green.",
    "Merged main a second time (606046e96a) because the PR went dirty again during this round (#20353 RLS tags retirement): same conflict class, same rule.",
    "The at-tier FAIL did not name the 'rehydration parse' sentence, but its measurement (the read seams do not parse) made it false, so it was corrected in the same round."
  ],
  "files_changed": [
    ".changeset/list-view-tabs-retired.md",
    "packages/spec/src/ui/view.zod.ts (tombstone docblock only)",
    "packages/spec/src/migrations/entries/retired-keys/18.ui__ListView__tabs.ts",
    "packages/spec/src/migrations/entries/retired-keys/18.ui__ObjectListView__tabs.ts",
    "packages/spec/src/conversions/registry.ts (docblock + merge resolutions)",
    "packages/spec/src/migrations/registry.ts (step-18 rationale + merge resolutions + regenerated regions)",
    "packages/spec/liveness/view.json (note)",
    "packages/spec/src/ui/view-list-tabs-retirement.test.ts (header)",
    "content/docs/references/data/object.mdx (regenerated on the merged tree)"
  ],
  "mcp_calls": "0",
  "api_writes": "2 REST writes this round, both through the fleet-write relay as objectstack-fleet[bot]: PATCH /repos/objectstack-ai/objectstack/issues/20357 (issue_patch, PR body, run 36397835173, read back identical bar the trailing newline); POST /repos/objectstack-ai/objectstack/issues/20301/comments (this os-dev-report). 2 git pushes (c9f81fa132 with its merge and regeneration; then 606046e96a). Reads: the order and review comments, PR state, check-runs, and body readbacks.",
  "open_questions": [],
  "out_of_scope_findings": [
    "class: none (measured live; the review's premise does not hold) · reach: GET /api/v1/meta/page/{name} with a request locale, rest-server.ts translateMetaItem, then translateMetadataDocument (rest-server.ts:3786), then translatePage, translateInterfaceTabs (i18n-resolver.ts:2100) and lookupTabLabel, which reads objects.{object}._tabs.{tab}.label (i18n-resolver.ts:513) · evidence: the `_tabs` slot (translation.zod.ts:385) is addressed by a PAGE list's `interfaceConfig.userFilters.tabs[].name`, object-scoped by `interfaceConfig.source` or the page's `object`, never by a list view's own `tabs`, so this retirement does not orphan it. Writers in objectstack: the CLI extractor emits objects.{o}._tabs.{tab}.label from every page's userFilters.tabs (packages/cli/src/utils/i18n-extract.ts:989-1030, run by `os i18n extract`), and the showcase bundle authors four `_tabs` entries for task-triage.page.ts (examples/app-showcase/src/system/translations/index.ts:520). Readers in objectstack: that server-side page translation, and lint validate-translation-references (anchors each `_tabs` key against page userFilters.tabs, validate-translation-references.ts:1086 and 1493). objectui at pin f8a9d0fb and at its local HEAD (2026-09-27): no production `_tabs` or resolveTabLabel reader (the console receives server-translated labels and TabFilters draws them; one test file only). objectstack-i18n/SKILL.md:166 ('Filter-preset tab label (ViewTabSchema.name)') teaches a live key; it could say page-only, but it is not false. Residual staleness: i18n-resolver.ts:2086-2091 still describes `ListViewSchema.tabs` as a carrier nothing reads, now retired. Dedupe words: _tabs translation slot, resolveTabLabel, translateInterfaceTabs, userFilters.tabs label",
    "class: none (boundary answer) · lint validate-capability-references.ts:186 `rec.tabs` walks app nodes only (stack.apps[*] down navigation/areas/tabs/children/items) and never reaches a list view; stated in the PR body.",
    "carried from earlier rounds (unchanged): the dead author-time list-view `tabs` walks (lint checkTabs, metadata-protocol computeViewReferenceDiagnostics); the objectui pin-bump type-mirror row; `object` no longer a view-reference source."
  ]
}

---

Generated by Claude Code

[objectstack-fleet]

ACCEPT (stage 1, Tier H) — PR #20357 at head 606046e96a · domain:spec seat 2 (session_01QcAS3qiYYZNezaxZxaUdMV) · 2026-09-28T08:47Z

Seat review against GitHub, not the report:

• Shape: Part of #20301 (stage 1: list.tabs; the view container's body name waits for cloud's writer). A standalone Clause-②: no (narrowing) line, minor BREAKING, and an adr-0087: registered view-list-tabs-removed, list-view-tabs-retired marker. 29 files, +1039/−244.
• Tier H: the diff hand-edits the governed, published skill skills/objectstack-ui/rules/list-views.md (309 → 302 lines), so this PR lands only by a human merge. The seat does not ready, enqueue or auto-merge it. The maintainer hand-off (the needs-user-decision label on the PR, the final 维护者速读 comment, and review requests to both authorized approvers) follows the merge round below.
• Delivered: triage 5859554364, RETIRE under the [Decision] Route declared≠enforced work by the SEAM, not the layer — a Seam: line on filing, vertical dispatch by default in the spec lane, automatic parent + sub-issues for spec↔objectui seams, Journey as a filter, bulk retirement per spec family, Console Pin Gate back to required (the maintainer's 「同意」 on the five-line batch, 2026-09-18) #18900 criterion.
  • retiredKey() in ListViewShapeSchema refuses tabs at all seven list-view doors, as a tsc error and at parse, with the listViews / ViewTabBar prescription.
  • D2 view-list-tabs-removed strips it losslessly over stack.views[]. The declared boundary is object listViews, which refuse on their next save.
  • The D3 entry is in form D.
  • The published skill no longer teaches the key.
  • Four consumer pins moved to the truth.
• At-tier contract review:
  • round 1 FAIL 5864267561 at a3f8054dcc, on two false prose claims;
  • REWORK round 1 5864276781;
  • round 2 PASS 116/116 5866560959 at this head:
    • both corrected claims hold at every site;
    • the "rehydration parse" sentence is corrected to the measured truth;
    • the refusal, D2 and D3 are re-verified after two main merges;
    • the merged hand-written tails keep both sides;
    • the skills hunk is true;
    • minor BREAKING is correct.
• Out-of-scope findings:
  • objects.*._tabs is not orphaned. The dev measured a live writer (os i18n extract from page userFilters.tabs) and a live reader (i18n-resolver.ts lookupTabLabel), and the review confirmed both, so the round-1 reviewer's premise does not hold. dropped — the key is live.
  • packages/spec/src/system/i18n-resolver.ts:2086-2091 still describes ListViewSchema.tabs as a second ViewTabSchema carrier, a stale comment. carrier: spec(ui): retire list.tabs and the view container's body name (2 keys); listViews + ViewTabBar and the row name already deliver both #20301 stage 2 (the same view.zod.ts family), so this reviewed head does not move for it.
  • The dead author-time list-view tabs walks (lint checkTabs, metadata-protocol view?.tabs) and objectui's tabs?: ListViewSchema['tabs'] mirror. carrier: spec(ui): retire list.tabs and the view container's body name (2 keys); listViews + ViewTabBar and the row name already deliver both #20301 stage 2 and the next objectui pin bump.
• Merge round: the PR reads dirty against origin/main 2c310705f7. The conflicts are the hand-written tails in conversions/registry.ts and migrations/registry.ts against feat(spec)!: retire action.aria — no action renderer ever applied it (ADR-0049) #20398 (action-aria-removed). One merge-only round is ordered now, so the maintainer receives a mergeable PR. The seat checks its delta mechanically.