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

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。

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.

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.
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.

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.

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

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

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.

"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.
"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.
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.
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.
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.

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

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.

Claim: PM loop round 1 (stage 2 of a staged card: the view container's body name)
Session: session_01YDt3PzwfrkuFzUBF89WPmM
Account: os-tesla (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-20301-view-container-name-retired
Worktree: objectstack-issue-20301
Domain: domain:spec
Seat: domain:spec#2
File surface:

packages/spec/src/ui/view.zod.ts: the container name (:4748 on 1d0600bf66) and its comment block, as a retiredKey tombstone with a prescription.
The retirement kit: an ADR-0087 D2 conversion or D3 entry, the generated region of migrations/registry.ts (generator only), liveness/view.json (the row stays dead, as for every tombstone), and the regenerated reference pages.
The in-repo writers that send a body name on a view container (the metadata-validation sweep named on the card, and any write path the measurement finds) are fixed in the same PR.
The carried notes from 5870707479: the stale i18n-resolver.ts:2086-2091 comment; the unreachable list-view tabs walks (lint checkTabs, metadata-protocol view?.tabs) and their fixtures.
One .changeset/20301-*.md.
⛔ Not :5522 (flattenedViewOverlayFields' name, stamped by the write path on a single inline view, a different key). ⛔ No objectui edit.
Stop on breach and explain in the report.

Container & model: M, mode:subagent, model: opus (dispatch-gates --tier at 1d0600bf66: no path-derived mandate; floor sonnet · default opus · ceiling CONTRACT_REVIEW_TIER). The at-tier contract review is owed and runs as a separate isolated subagent after the PR opens.
Clause-②: yes (narrowing)
Thread-read: 5925985857
Serial constraints cleared:

At 2026-10-02T12:37Z, 10 open PRs were read. PR docs: comments outside driver-memory name what replaced the retired reference matcher (#20822 group 4) #21395 (domain:engine) edits view.zod.ts at :602 only, which is disjoint.
PR feat(automation): a flow's credentials live in a write-only channel on the secret seam, not in its stored definition (#20790) #21377 (security(flows): move a flow's inbound-hook secret out of flow metadata into the write-only secret seam #7799 established — no read, the generic data door included, returns it #20790) edits metadata-protocol/src/protocol.ts and siblings. If the writer fix lands there, the second to land merges main.
PR fix(analytics)!: a query window outside the non-negative integers is refused at the door, and an offset with no limit runs on SQLite #21399 (analytics: a query window with no single answer across drivers (negative or fractional limit or offset, or an offset with no limit) answers 500 on the native face and a slice on the ObjectQL face; AnalyticsQuerySchema admits all of them #21365) and this seat's ai: guardrails, memory, structuredOutput, lifecycle and tool.outputSchema are enforced by the agent runtime (5 keys), starting with the guardrails the built-in agents already declare #20274 memory PR write migrations/registry.ts (and ai: guardrails, memory, structuredOutput, lifecycle and tool.outputSchema are enforced by the agent runtime (5 keys), starting with the guardrails the built-in agents already declare #20274 conversions/registry.ts). The second to land regenerates through scripts/pm/os-regen-merge.sh.
Verify-lock depth 1.

Premise, re-measured at claim time (as the landing record 5870707479 orders):

Cloud's writer does not send it: the repo:cloud#1 reading 5925985857 at cloud 72b8415a found no top-level name in service-ai's view artifacts.
⚠️ The declaration's own comment (view.zod.ts:4741–:4746) says the container name was declared because "undeclared, they were stripped, and the metadata door quietly discarded an item's own name on the way through" (未知键静默剥离仍是全仓默认:把 #3405 的 strict 收紧从一个 schema 推广到整个可授权面(ADR-0078 完整性闸门) #4001 batch 5). So the dispatch measures every in-repo and objectui writer and reader of a body name before any tombstone, and returns here, without building, if the platform's own door still writes it.

Seat terms: p3, under the maintainer's 「p3 的卡也可以派了」 (5952442849).

This PR is Fixes #20301, closing stage 2, the card's last.

Generated by Claude Code

os-dev-report

{
  "issue": 20301,
  "status": "needs_decision",
  "branch": "claude/issue-20301-view-container-name-retired",
  "pr": null,
  "session": "session_01YDt3PzwfrkuFzUBF89WPmM (mode:subagent, the PM's session; identity is the branch named by claim 5952493523, which is the newest Claim: on the card and names this branch)",
  "premise_still_valid": false,
  "summary": "STOPPED at the A2 stop condition, before any tombstone. Nothing was written to the tree. The metadata door itself writes a view container's body `name`. `saveMetaItem` runs `normalizeViewMetadata` (packages/metadata-protocol/src/protocol.ts:1220, stamp at :1227, call at :17256) before the schema gate, and it stamps the save name onto every view body that has none, containers included. A container saved with no name is stored as `{ list, name: 'crm_lead' }`, the by-name read serves that `name` back, and a GET then PUT round trip sends it and is accepted today. The door's own overlay identity paths also READ the body copy. `hydrateOverlayIntoRegistry` (:15990) does not register, or expand, a body that has no `name`. `mergePackageAwareOverlay` (:1835) skips an overlay row with no body `name` and keys its slot by `body.name`. The ObjectQL boot loop (engine.ts:7010) mints `name` onto every stack container it registers. With the tombstone in place, every container save through the door would 422 on the name the door itself stamped. The precondition that the landing record set (`the platform's own writers stop sending a body name`) does not hold, and the card's premise (a `dead` body copy that nobody reads) is false at 1d0600bf66. Two smaller corrections: the metadata-validation sweep does not author a body `name` (the door stamps it), and a 2026-09-03 maintainer ruling refused exactly this spec-level forbid (direction 3). The PM returns to the card; options and the four axes are in open_questions.",
  "tests": "All runs are at worktree HEAD 1d0600bf66 (== BASE, no commits). (1) Build closure: `OS_SKIP_DTS=1 bash scripts/pm/os-verify-lock.sh -c \"pnpm --workspace-concurrency=2 --filter '@objectstack/metadata-protocol^...' build\"` -> `VERDICT command-exit 0 · held the lock 63s`. (2) The pin on main that records the door's stamp: `pnpm --filter @objectstack/metadata-protocol exec vitest run --maxWorkers=2 src/view-container-runtime-expansion.test.ts` -> `Test Files 1 passed (1) / Tests 16 passed (16)`, `VERDICT command-exit 0`. Its case at :252 asserts that the stored row of a nameless container equals `{ ...leadContainer, name: 'crm_lead' }`, with the comment `plus the name the write door has always stamped`. (3) Probe A2(a)/(b): a temporary file, src/zz-probe-20301.test.ts, built on that file's stub engine, run, then deleted. Output: `PROBE stored body keys: list,name name= crm_lead` / `PROBE served body keys: _diagnostics,list,name name= crm_lead` / `PROBE registry container key present: true registered name= crm_lead` / `PROBE GET->PUT round-trip success: true sent body has name: true` / `PROBE hydrateOverlayIntoRegistry(nameless) = false ; with name = true`; `Tests 1 passed`, `VERDICT command-exit 0`. (4) Probe: a body name that differs from the row name, at the runtime door. Temporary file src/zz-probe2-20301.test.ts, deleted after the run. Output: `PROBE2 save error: none` / `PROBE2 rows: crm_lead -> body.name=lead_views` / `PROBE2 registry keys: lead_views,crm_lead.default` / `PROBE2 list names: crm_lead.default`; `VERDICT command-exit 0`. After each deletion `git status --porcelain` printed nothing. (5) Census A2(e) with the TypeScript AST. It finds object literals with a top-level list/form/listViews/formViews and no viewKind/type key, passed to defineView(...), in a `views: [...]` array, or in a View-typed declaration. examples/** + packages/platform-objects (260 files): `containers 10, withName 0`. packages/** production sources (2853 files): `containers 27, withName 0, withObject 24` (control: the key detector reads keys). packages/** including tests (7359 files): `containers 227, withName 90` (lit control: the name detector fires; all 90 are test fixtures). (6) `node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands` -> exit 2, `this branch changes nothing against 'origin/main' (merge base 1d0600bf6) — nothing to derive`. No ablation and no reverse verification: no implementation exists to verify.",
  "mcp_calls": "0",
  "api_writes": "1 — POST issues/20301/comments (this os-dev-report), through scripts/pm/post-stamped.mjs over the fleet-write relay. Reads were REST GET only (the card and its comments, PR #15319, comment 5520462984 -> 404).",
  "gates": [
    "dispatch-gates --commands: exit 2, `this branch changes nothing against 'origin/main' (merge base 1d0600bf6) — nothing to derive`",
    "Zone gates (spec build/check:generated/check:liveness/test/typecheck, consumer suites, check:adr-0087-registration, changeset gates, check:doc-authoring, check:nul-bytes): NOT RUN, reason: stopped at A2 before any edit, so there is no diff to gate",
    "verify-lock runs: build closure VERDICT command-exit 0; view-container-runtime-expansion.test.ts VERDICT command-exit 0 (16/16); probe 1 VERDICT command-exit 0; probe 2 VERDICT command-exit 0"
  ],
  "line_budget": "n/a — no diff; no skills/** or governed file touched",
  "files_changed": "none committed. Two temporary probe files were created and deleted in the worktree (packages/metadata-protocol/src/zz-probe-20301.test.ts, zz-probe2-20301.test.ts); the tree reads clean.",
  "semver": "none — no changeset written, because the stop came before the kit. The A5 level (minor + BREAKING + ADR-0087 marker) was not exercised, and no changeset gate judged anything.",
  "deviations": [
    "The empty branch was pushed per os-dev rule 1 (the write probe) before the A2 stop. origin/claude/issue-20301-view-container-name-retired exists at 1d0600bf66 with zero commits of its own, and no PR is open. The seat may delete it or reuse it.",
    "Zone 1 conflict, reported and not resolved: Zone 1 rules RETIRE for the container `name`. The merged PR #15319 records a maintainer ruling of 2026-09-03 (direction 2 adopted) that explicitly refused `direction 3 (forbid ViewSchema.name on object-scoped containers in spec)`. The PR body says so at its changeset paragraph (`forbidding ViewSchema.name in spec was direction 3, explicitly refused`), and the header of packages/objectql/src/view-container-divergent-name-registrars.test.ts says the same. A tombstone is a superset of direction 3. The ruling comment itself (5520462984, on #14666) answers 404 from this session, so its own text is NOT MEASURED; the PR body and the test header are the surviving records.",
    "Zone 2 A2(d) falsified: the metadata-validation sweep does not author a body `name`. Its view fixture (packages/objectql/src/metadata-validation-sweep.test.ts:164-178) is a nameless container, and `runOne` passes `name: fx.valid.name ?? 'sweep_view'` as the REQUEST name (:392). The door's `normalizeViewMetadata` then stamps it. There is no in-repo writer to `fix` short of the door.",
    "A4 carried notes NOT executed (under the stop, the budget is the report comment only). Locations re-measured at 1d0600bf66: the stale ListView-tabs comment is now packages/spec/src/system/i18n-resolver.ts:2355-2361 (it still says `ListViewSchema.tabs (multi-tab view interface)` is a live carrier); lint `checkTabs` is packages/lint/src/validate-list-view-field-refs.ts:786-799; metadata-protocol `view?.tabs` is packages/metadata-protocol/src/metadata-diagnostics.ts:231. These three do not depend on the container `name` and could ride a separate small PR if the seat wants them.",
    "History questions (when the sweep fixture lost its `name`, when the stamp was added) are NOT MEASURED. The checkout is shallow (`--is-shallow-repository` = true), and the stop does not need them."
  ],
  "a2_census": {
    "a_door_write": "WRITES. packages/metadata-protocol/src/protocol.ts `normalizeViewMetadata` (:1220; `return { ...it, ...(it.name ? undefined : { name: saveName }), ...patch }` at :1227) is called inside `ObjectStackProtocolImplementation.saveMetaItem` at :17256, ahead of the schema gate. For a view `projectStorableViewBody` keeps every key the body carried, so the stamped `name` is persisted. `SysMetadataRepository.put` (sys-metadata-repository.ts:638-640) writes the body verbatim and the row name to the `name` column. Measured: probe 1 stored keys `list,name`. Pinned on main: view-container-runtime-expansion.test.ts:252. The doors that call it do not inject a name: REST PUT /meta/:type/:name (rest-server.ts:7097-7125, request built field by field) and the dispatcher door (runtime/src/domains/meta.ts:1424).",
    "b_read_path": "SERVES AND ROUND-TRIPS IT. `getMetaItem` step 1 serves the stored overlay body, which carries the stamped `name`. Probe 1: served keys `_diagnostics,list,name`, and GET then PUT of that body returns success true while sending `name`. Step 3 (SchemaRegistry) serves the boot loop's copy, and packages/objectql/src/engine.ts `registerMetadataCollections` :7010 `toRegister = item.name === itemName ? item : { ...item, name: itemName }` mints `name` onto every stack container it registers. That half is a code read, and PR #15319's body states the same (`the boot loop's minting of a derived name onto a container that carries none`); it was not driven. The artifact/HMR loader (packages/metadata/src/plugin.ts:1167-1200 `_registerArtifactBodyCollections`) registers a container as authored, with no mint.",
    "door_readers_of_the_body_copy": "protocol.ts `hydrateOverlayIntoRegistry` :15990 (`!('name' in data)) return false` — no registry entry and no `hydrateExpandedViewItems`; probe 1 measured false without a name, true with one) · protocol.ts `mergePackageAwareOverlay` :1835 (overlay row skipped without a body `name`; slot keyed by `body.name`) · protocol.ts `expandRuntimeViewContainer` :16064 and packages/metadata/src/view-container.ts `deriveViewContainerObject` :81 (`name` is the last-resort object binding) · packages/objectql/src/view-container-name-refusal.ts `viewContainerNameRefusal` plus the artifact door's `assertMetadataRegisterContract` (both refuse a divergent container `name`). The ledger row packages/spec/liveness/view.json `name` (status dead, verifiedAt 2026-08-10, `a copy nobody reads`) is therefore falsified by measurement.",
    "c_objectui_at_pin_89cad75d55": "No objectui writer AUTHORS a container body `name`. Studio's view create path emits a ViewItem whose `name` is the qualified id (packages/app-shell/src/views/metadata-admin/anchors.ts:291 `createBuildBody`), a ViewItem key and not the container key. Studio's view list drops containers (packages/app-shell/src/services/builtinComponents.tsx:232 `listFilter: (item) => !isAggregatedViewContainer(item)`). The generic editor's edit mode PUTs the served body back (`{ ...current, ...partial }`, clientValidation.ts:635 comment), so a container opened by direct URL would round-trip the server-served `name`; that is the server's write echoed, not an objectui writer. The adapter `updateView` (packages/data-objectstack/src/index.ts:5864) unwraps `current.list` from a container and writes a flattened overlay, which carries no container `name`. Read from the local objectui clone at the pin commit (git show / git grep).",
    "d_metadata_validation_sweep": "Not a writer. packages/objectql/src/metadata-validation-sweep.test.ts:164-178: the view fixture has no body `name`; :392 passes `name: fx.valid.name ?? 'sweep_' + type` as the request name, and the door stamps it into the body (a_door_write).",
    "e_examples_and_platform_objects": "examples/** + packages/platform-objects: 10 containers, 0 author `name`. packages/** production sources: 27 containers, 0 `name` (24 carry `object`, the control). Authored container `name`s exist only in test fixtures (90 of 227 across packages/** tests: cli, lint and others). Cloud: reading 5925985857 (not re-measured) finds no `name` in service-ai's view artifacts."
  },
  "open_questions": [
    {
      "question": "The view container's body `name` is written by the platform's own metadata door (`normalizeViewMetadata` stamps it on every view save), served back by the by-name read, and read by the door's overlay identity paths (`hydrateOverlayIntoRegistry`, `mergePackageAwareOverlay`). The A2 stop fired, and a 2026-09-03 maintainer ruling refused forbidding `ViewSchema.name` in spec (direction 3). How should stage 2 of #20301 close?",
      "options": [
        "A — Keep the key and correct the record. Flip packages/spec/liveness/view.json `name` from dead to live, with the measured evidence: the writer `normalizeViewMetadata` protocol.ts:1227; the readers `hydrateOverlayIntoRegistry` :15990, `mergePackageAwareOverlay` :1835 and `deriveViewContainerObject`; the boot-loop mint engine.ts:7010. Rewrite its note, regenerate the README counts, and close stage 2 (and the card) as premise falsified, with stage 1 as the card's delivered half. Cost: one ledger row plus the README count, in a small spec PR. liveness/ is in @objectstack/spec `files[]`, so the changeset rule's published-surface test applies to that PR. No tombstone, no conversion, no door change. This is consistent with the 2026-09-03 ruling.",
        "B — Retire anyway by moving the door's view-container identity off the body. Stop the stamp for containers (normalizeViewMetadata). Key overlay rows by the sys_metadata row column instead of `body.name` in hydrateOverlayIntoRegistry, mergePackageAwareOverlay and the other `data.name` callers. This seam is shared by EVERY overlay type, so either it becomes generic, or it becomes a view-only identity path, which is a dialect. Stop the boot loop's mint (engine.ts:7010), or strip it at serve time. Add an ADR-0087 D2 conversion stripping `name` from every stored container row: every door-saved container carries one. That is lossless when it equals the row name; the divergent case has no lossless answer, so it needs a D3 semantic entry. Then the full tombstone kit. Cost: multi-package (metadata-protocol x3 seams on the generic overlay identity path, objectql boot loop, spec kit, conversion), well beyond what the card estimated, plus a maintainer word that supersedes the 2026-09-03 refusal of direction 3.",
        "C — A plus enforcement: keep the key, mark it live, and make the runtime door refuse a container body `name` that differs from the save (row) name. That is the same divergence the 2026-09-03 ruling already refuses at the two SOURCE registrars. Measured open at the runtime door by probe 2: the save is accepted, the row is `crm_lead` with body name `lead_views`, and the container is registered under `lead_views`. Cost: a small metadata-protocol change plus a pin, reusing `viewContainerNameRefusal`'s judgment and words (one judge) instead of a second rule."
      ],
      "recommendation": "A, with C's divergence routed separately as the class-b finding below rather than folded in. By the four axes: (实际业务需求) measured — no author writes it (0 of 10 example containers, 0 of 27 production containers), but the platform's own door writes it on every view save and two door readers key on it, so it is the stored identity copy, not an unused authoring capability, and the RETIRE premise `a copy nobody reads` is false; (项目长远合理性) every overlay type stores `name` in its body, equal to the row (objects, dashboards, pages), so taking it off views alone makes a view-only identity dialect, and B is a re-architecture of the generic overlay identity that deserves its own card if ever; (防 AI 写错) the trap the card names (an AI writes a container `name` that does nothing) does not exist as described, because the key is not inert. The real trap is divergence (body name differs from row name), measured accepted at the runtime door, and C or the finding closes it loudly. A tombstone would make the platform 422 its own stamped writes, the exact `rejected shapes the platform itself writes` failure the schema comment at view.zod.ts:4720-4728 already records; (创业阶段不扩散) A costs one ledger row, B expands far past the card's estimate, and the maintainer's 2026-09-03 ruling already refused the spec-forbid direction. Four-axis conflict to surface: Zone 1's RETIRE verdict and the 2026-09-03 refusal of direction 3 point opposite ways on this key; this report does not pick between them, and that is for the maintainer."
    }
  ],
  "out_of_scope_findings": [
    "class: b · Seam: spec:ViewSchema.name (packages/spec/src/ui/view.zod.ts:4748, `.describe('Item name — supplied by the metadata door; for an object-scoped container it is the object name.')`) → runtime:metadata-protocol saveMetaItem → hydrateOverlayIntoRegistry (protocol.ts:15990) / mergePackageAwareOverlay (protocol.ts:1835) · reach: the save door, measured at the protocol `saveMetaItem` seam that REST PUT /meta/view/:name and the dispatcher door call, with the stub engine of view-container-runtime-expansion.test.ts (not over HTTP): `saveMetaItem({ type: 'view', name: 'crm_lead', item: { name: 'lead_views', list: {...} } })` is accepted with no error, the row is `crm_lead`, the body name is `lead_views`, and the registry keys are `lead_views,crm_lead.default`. One document is filed under two keys. The two SOURCE registrars refuse this same document (objectql `viewContainerNameRefusal`, the artifact door `assertMetadataRegisterContract`, VALIDATION_ERROR/400) under the 2026-09-03 ruling; the third, runtime door does not. · dedupe words: `view container name divergent runtime door`, `saveMetaItem container body name row name mismatch`, `hydrateOverlayIntoRegistry body name key`, `viewContainerNameRefusal metadata door`",
    "carrier: the seat's stage-2 decision on this card · noted, not filed — the liveness ledger row packages/spec/liveness/view.json `name` (dead, verifiedAt 2026-08-10) is falsified by the measurements above (one door writer, two door readers). The correction rides option A or B, whichever the maintainer picks."
  ]
}

Generated by Claude Code

Stage 2: the stop condition fired. The premise is falsified, the retirement is withdrawn, and the stage is re-scoped to a ledger correction (claim amendment)

domain:spec seat 2 (session_01YDt3PzwfrkuFzUBF89WPmM), holder of claim 5952493523 · 2026-10-02T13:06Z. From the os-dev report 5953000865, which stopped at A2 before any edit. Nothing was committed, and the branch is empty at 1d0600bf66.

What was measured (file, symbol and probe lines are in 5953000865):

The door writes it. metadata-protocol normalizeViewMetadata (protocol.ts:1227, called by saveMetaItem at :17256, ahead of the schema gate) stamps the save name onto every view body that has none, containers included. This is pinned on main: view-container-runtime-expansion.test.ts:252, "plus the name the write door has always stamped".
The read path serves it, and a GET then PUT round-trips it. The objectql boot loop mints it onto every stack container (engine.ts:7010).
The door reads it. hydrateOverlayIntoRegistry (:15990) does not register a body without name. mergePackageAwareOverlay (:1835) keys its slot by body.name.
No author writes it: 0 of 10 example containers, 0 of 27 production containers. Cloud does not send it (5925985857). objectui writes none of its own.

Why the tombstone is off, by rulings already on record:

Triage's own guard on this key (5859554364, note 2): "⛔ Never retire a key the platform's own writer still sends." The platform's own writer is the door itself, and the measurement found no in-repo or cloud writer the door does not cover.
The maintainer's ruling of 2026-09-03 (PR fix(objectql): the boot loop refuses a view container whose name disagrees with its derived object key (#14666) #15319 adopted direction 2) explicitly refused "direction 3 (forbid ViewSchema.name on object-scoped containers in spec)". This is quoted from PR fix(objectql): the boot loop refuses a view container whose name disagrees with its derived object key (#14666) #15319's body and the header of objectql/src/view-container-divergent-name-registrars.test.ts. The ruling comment itself answers 404 from here. A tombstone is a superset of direction 3.
The verdict's reason, "a body copy of the container's identity … duplicates the row name", read the copy as unused. Measured, it is the stored identity copy that two door readers key on. Per this repo's rule, the inherited verdict falls with its reason.

The stage-2 scope from this act (claim amendment; same branch claude/issue-20301-view-container-name-retired):

Correct the ledger. packages/spec/liveness/view.json name: dead → live. Cite the writer and readers above (verifiedAt, measured in-repo). The note says the key is the door-stamped identity copy, and that a divergent body name is refused at the two source registrars under the 2026-09-03 ruling. Regenerate the counts through gen:liveness-counts.
The carried notes from 5870707479, re-located by the dev:
- the stale ListViewSchema.tabs comment, now packages/spec/src/system/i18n-resolver.ts:2355-2361;
- lint checkTabs (packages/lint/src/validate-list-view-field-refs.ts:786-799);
- metadata-protocol view?.tabs (packages/metadata-protocol/src/metadata-diagnostics.ts:231), each deleted if unreachable since stage 1, with their fixtures.
The release: a patch changeset (the ledger ships in @objectstack/spec's files[]) plus each touched package. Clause-②: no. Fixes #20301: stage 1 delivered list.tabs, and this delivers the corrected record.

⛔ No tombstone, no conversion, no door change, no view.zod.ts key change.

Not taken:

Taking the identity off the body (the report's option B): stop the stamp, re-key the generic overlay identity path by the row column, add a conversion for every stored container, then the kit. It supersedes a maintainer ruling and is a multi-package re-architecture nobody is pulling for. If ever wanted, it is a new card in the decision box, not this one.
The divergence the runtime door accepts (the report's option C) is a defect of its own, filed as a card in this act and named in the next comment.

The at-tier contract review still follows the PR (packages/spec/src/** is touched by the comment edit).

Generated by Claude Code

Correction to `5953063959` item 1: the `name` row stays `dead`; only its note is corrected

domain:spec seat 2 (session_01YDt3PzwfrkuFzUBF89WPmM), holder of claim 5952493523 · 2026-10-02T13:08Z. A seat assumption, falsified within the hour, is corrected here on the same card.

What I wrote: flip liveness/view.json name from dead to live.
Why that is wrong:
- The ledger's definition (packages/spec/liveness/README.md:180) reads: "live means authoring the property changes runtime behaviour". An AUTHORED container name restates the registry key or contradicts it.
- The contradiction is refused at the two source registrars under the 2026-09-03 ruling, whose remedy is "drop name".
- The runtime door's acceptance of a contradicting one is a defect. It is filed separately, as the next comment names.
- What the door stamps and keys on is its own identity copy, not an authored value. PR fix(cli)!: os generate binds declared metadata (--object, --flow), never the item name, and every scaffold passes validate, build and lint #21369 (merged) read the same key from its readers and reached dead for the same reason. os validate's liveness-dead-property warning, "drop it", is correct advice for an author and for an AI.
Item 1, corrected:
- name stays dead. Its note is rewritten to the measured facts: the metadata door stamps the save name into the body (normalizeViewMetadata) and its overlay paths key on that stamped copy (hydrateOverlayIntoRegistry, mergePackageAwareOverlay), which replaces the false "a copy nobody reads".
- The note says why the key is not retired: the door's own writes, the 2026-09-03 refusal of direction 3, and triage's guard. It is kept on the "KEPT deliberately, exempt from enforce-or-remove" precedent the ledger already uses.
- verifiedAt is updated. No status moves, so state-counts does not move.
Unchanged from 5953063959: items 2–3, the "not taken" list, Clause-②: no, a patch changeset, and Fixes #20301.

Generated by Claude Code

Pointer from domain:spec seat 2 (session_01YDt3PzwfrkuFzUBF89WPmM) · 2026-10-02T13:09Z, named by 5953063959 and 5953104466. The runtime save door's acceptance of a container body name that contradicts the row name, which the two source registrars refuse under the 2026-09-03 ruling, is filed as #21412 (bare; routing and grading are triage's).

Generated by Claude Code

os-dev-report

{
  "issue": 20301,
  "status": "done",
  "branch": "claude/issue-20301-view-container-name-retired",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/21423",
  "session": "session_01YDt3PzwfrkuFzUBF89WPmM (mode:subagent, the PM's session; identity is the branch named by claim 5952493523, re-scoped by 5953063959 as corrected by 5953104466)",
  "premise_still_valid": true,
  "summary": "Built the reduced stage-2 scope on the same branch, fast-forwarded to origin/main 6d67ad5eca, and opened draft PR #21423 (`Fixes #20301`, `Clause-②: no`). (1) liveness/view.json `name` STAYS `dead`. Only its note is rewritten to the measured facts: the door stamps the save name into the body (normalizeViewMetadata); the overlay paths key on it (hydrateOverlayIntoRegistry, mergePackageAwareOverlay); a contradicting authored value is refused at the two source registrars under the 2026-09-03 ruling and accepted at the runtime door (#21412); and the three reasons it is kept. verifiedAt is now 2026-10-02 and evidenceScope stays cross-repo, with ViewPreview re-read at the pin 89cad75d55. gen:liveness-counts rewrote 0 shards. (2) The carried notes: the stale i18n-resolver comment is corrected; the unreachable list-view `tabs` reads in lint (`checkTabs(listView.tabs, ...)`) and in metadata-protocol `computeViewReferenceDiagnostics` (`view?.tabs`) are deleted with their fixtures, and each keeps its userFilters.tabs walk; the stage-1 absence pin's self-expiring RESIDUE drops the two fixture files. (3) Three patch changesets: spec, lint, metadata-protocol. No tombstone, conversion, door change or view.zod.ts edit.",
  "tests": "All at HEAD 7ee481ef2d (base 6d67ad5eca); every os-verify-lock line reads `VERDICT command-exit 0` unless stated. Build: `pnpm --workspace-concurrency=2 --filter '@objectstack/objectql^...' build` (with declarations, held 267s). Spec: `check:generated` -> `✓ All 15 generated artifacts are up to date`; `typecheck` exit 0 (`SPEC_GENERATED_EXIT=0 SPEC_TYPECHECK_EXIT=0`); `check:liveness` exit 0, `✓ packages/spec/liveness/state-counts/ is current`; `gen:liveness-counts` -> `0 shard(s) rewritten, 0 pruned`, totals `988 live · 3 experimental · 1 live-elsewhere · 109 dead · 10 planned = 1111` (no status moved); `test` -> `Test Files 600 passed (600) / Tests 17601 passed | 1 todo`, including view-list-tabs-retirement.test.ts with RESIDUE reduced to the CLI file. Lint: `test && typecheck` -> `Test Files 119 passed (119) / Tests 5574 passed (5574)`, `check:test-typecheck: OK`. metadata-protocol: `exec vitest run --maxWorkers=2 && typecheck` -> `Test Files 201 passed | 3 skipped (204) / Tests 2983 passed | 19 skipped (3002)`. objectql (test file only): `exec vitest run --maxWorkers=2 src/metadata-diagnostics.test.ts src/view-container-divergent-name-registrars.test.ts src/metadata-validation-sweep.test.ts && typecheck && build` -> `Test Files 3 passed / Tests 23 passed`, `check:test-typecheck: OK`. Derived gates: `dispatch-gates --commands` listed 91; each was run with its exit recorded before any pipe; `--ran` -> `✓ dispatch-gates --ran: 91 derived famil(ies) accounted for — 90 run, 1 NOT-MEASURED`. No ablation: this is a deletion of reads that are unreachable on parsed input, not a new guard.",
  "mcp_calls": "0",
  "api_writes": "4 across both rounds — POST issues/20301/comments (the A2-stop report 5953000865, via post-stamped over the relay); POST repos/objectstack-ai/objectstack/pulls (pr_create, draft forced, via with-fleet.sh --via dispatch, relay run 37019721873; read-back 7882 bytes sent, 7882 stored, identical); POST issues/21423/assignees os-tesla (label-write.mjs over the relay, run 37019851709; read-back MATCHES); POST issues/20301/comments (this report). git push is not counted. No label was added by me.",
  "gates": [
    "node scripts/check-adr-0087-registration.mjs --base origin/main :: exit 0 — `✓ check-adr-0087-registration: this PR adds no declared-breaking changeset (3 non-breaking changeset(s) seen).`",
    "node scripts/check-changeset-no-major.mjs --base origin/main :: exit 0 — `✓ This diff introduces no major bump.`",
    "node scripts/check-empty-changeset.mjs --base origin/main :: exit 0; pnpm check:changeset-gate-self-tests :: exit 0",
    "pnpm check:doc-authoring :: exit 0; pnpm check:nul-bytes :: exit 0; pnpm check:cross-package-test-inputs :: exit 0",
    "pnpm --filter @objectstack/spec run check:liveness / check:api-surface / check:authorable-surface / check:docs / check:variant-docs / check:empty-state / check:exported-any / check:dual-source-exports / check:skill-refs / check:strictness-ledger / check:objectui-pin-citations (+8 more spec check:*) :: all exit 0",
    "pnpm check:lean-entry-closure :: first exit 3 (PREREQUISITE NOT MET: objectql dist absent), re-run after building objectql :: exit 0 — `✓ check-lean-entry-closure: 2 published condition(s) measured from a real load ... Admitted set held exactly (15 packages)`",
    "NOT MEASURED: pnpm check:dual-build-cjs-loads :: exit 3, reason: PREREQUISITE NOT MET — it loads every package's built dist and this run built only the objectql closure (`Run pnpm build first. This is NOT a pass: nothing was measured.`)",
    "NOT MEASURED: node scripts/check-engine-split-ratio.mjs --days 90 :: exit 2, reason: the shallow clone's oldest visible commit (2026-09-18) sits inside the 90-day window (`cannot compute the ADR-0076 D7 trigger metric`). The --ran tally counts this line as run; it measured nothing.",
    "All other 82 derived commands :: exit 0 (record: ran.list reconciled by --ran)",
    "CI at the PR head: in_progress when this report was written (Test Core 1-6, Dogfood 1-3, Temporal Conformance running; Build Docs and Console Pin Gate skipped). Not waited on."
  ],
  "line_budget": "n/a — no skills/** or governed surface touched; diff is 11 files, +88 / -59.",
  "files_changed": [
    ".changeset/20301-spec-view-container-name-ledger-note.md (new, @objectstack/spec patch)",
    ".changeset/20301-lint-list-view-tabs-walk-deleted.md (new, @objectstack/lint patch)",
    ".changeset/20301-metadata-protocol-list-view-tabs-walk-deleted.md (new, @objectstack/metadata-protocol patch)",
    "packages/spec/liveness/view.json (name note + verifiedAt; list.tabs note's follow-up sentence)",
    "packages/spec/liveness/README.md (view cell's name sentence)",
    "packages/spec/src/system/i18n-resolver.ts (comment only)",
    "packages/spec/src/ui/view-list-tabs-retirement.test.ts (RESIDUE drops the two deleted fixtures)",
    "packages/lint/src/validate-list-view-field-refs.ts (list.tabs call deleted; header prose)",
    "packages/lint/src/validate-list-view-field-refs.test.ts (list.tabs fixtures and HARD_CODED_FILTER_WALKS row deleted)",
    "packages/metadata-protocol/src/metadata-diagnostics.ts (view?.tabs read deleted)",
    "packages/objectql/src/metadata-diagnostics.test.ts (list.tabs fixtures deleted; the deleted-read case now asserts the surviving userFilters.tabs read)"
  ],
  "semver": "patch x3 (@objectstack/spec, @objectstack/lint, @objectstack/metadata-protocol). Each changeset carries `Clause-②: no`, matching the PR body's second line. No ADR-0087 marker is owed, because there is no breaking changeset. Gate lines: `✓ check-adr-0087-registration: this PR adds no declared-breaking changeset (3 non-breaking changeset(s) seen).` and `✓ This diff introduces no major bump.` objectql gets no changeset: test-only change, and its files[] is dist/README/CHANGELOG.",
  "deviations": [
    "Scope reading: besides the name note, I corrected two neighbouring ledger sentences that carry the same false claims. The README `view` cell said `the platform's own writers send name (artifact-shipped containers, the metadata-validation sweep)`. The list.tabs row's note said the two walks `are left for their own follow-up`. The spec-property-retirement skill's correction rule requires the README row to follow a ledger correction. Both are prose; no status changed.",
    "lint header prose that listed `tabs[].filter` and `tabs[].view` as walked positions, and the lint test's comments that counted `three` hard-coded walks, were updated to match the deletion.",
    "The objectql case `flags tab filter rules pointing at unknown fields` was converted, not just deleted. It now asserts the same unknown field on the surviving `userFilters.tabs` read, which had no negative case in that file.",
    "Not merged with origin/main (declared narrowing). ceb4a939b4 is 7 commits ahead of base 6d67ad5eca, and `git diff --stat 6d67ad5eca ceb4a939b4` over the 11 paths is empty; CI's merge ref judges the combination.",
    "`default PR labels` was read through os-dev's label rule (named labels + skip-changeset by its criterion): no label was named and skip-changeset does not apply (three changesets), so my label-write carried only `--assign os-tesla`. The automation's path/size labels (documentation, size/m, tests, tooling, protocol:ui, protocol:system) were already on the PR at read-back; I wrote none of them.",
    "Commit trailers: the model-free pair (`Claude-Session` + `Co-authored-by: Claude`) per AGENTS.md and the pre-push hook, not the harness reminder's model-named `Co-Authored-By` line; neither commit names a card.",
    "The worktree was removed after the PR opened (node_modules deleted, `git worktree remove` without --force, exit 0)."
  ],
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: the next PR that touches the ViewSchema guidance block, else none · noted, not filed — packages/spec/src/ui/view.zod.ts, about lines 4720-4728 (the comment above `guidance:` in ViewSchema), still says `saveMetaItem sends the name, artifact-shipped containers do (service-ai/ai_traces), the validation sweep injects it`. Measured: only the door's stamp is real. It is the same misattribution this PR corrects in the ledger, left in place by this stage's no-view.zod.ts scope. Also in the PR's Acceptance notes.",
    "carrier: objectui's next pin bump · noted, not filed — objectui's `tabs?: ListViewSchema['tabs']` mirror (types objectql.ts) is objectui's, as carried from 5870707479; named in the PR's Acceptance notes."
  ]
}

Generated by Claude Code

ACCEPT — PR #21423 @ `7ee481ef` (stage 2, as re-scoped)

domain:spec seat 2 (session_01YDt3PzwfrkuFzUBF89WPmM), holder of claim 5952493523 · 2026-10-02T14:51Z

Shape (read on GitHub): a draft against main. The first line is Fixes #20301, and Clause-②: no sits at a line start. 11 files, +88 / −59, all inside the re-scope 5953063959 as corrected by 5953104466. The two neighbouring ledger sentences that carried the same false claim (the README.md view cell and the list.tabs note's follow-up sentence) are corrected with them, by the playbook's correction rule.
Contract review: at-tier PASS 5954879465 on this exact head.
CI on 7ee481ef: 32 success and 3 skipped by design (Build Docs path-filtered, Console Pin Gate with no .objectui-sha change, and the opt-in packed-tarball smoke).
Governed surface: check-governed-merges --pr 21423 reads NOT governed, 147 changed lines.
The re-scope, item by item:
- liveness/view.json name stays dead. Only its note is rewritten to the measured facts: the door stamps the save name into the body, the overlay paths key on that stamped copy, and an authored value restates the key or contradicts it. The note gives why the key is kept. No status moved (gen:liveness-counts: 0 shards rewritten).
- Stage 1's carried residue is gone:
  - the stale i18n-resolver.ts comment is corrected;
  - the unreachable list.tabs reads in lint and metadata-protocol are deleted with their fixtures, each keeping its userFilters.tabs walk;
  - the stage-1 absence pin's RESIDUE drops the two files.
- Releases: three patch changesets (spec, lint, metadata-protocol).
- ⛔ Not changed: no tombstone, conversion, door change or view.zod.ts key change.
Prose read sentence by sentence by this seat: the name note and the three changesets.

Out-of-scope findings:

The comment above ViewSchema's guidance: (view.zod.ts, about :4720–:4728) still attributes the container name to artifact-shipped containers and the validation sweep. → carrier: finding(metadata-protocol): the runtime save door accepts a view container whose body name contradicts its row name and registers it under both keys; the two source registrars refuse the same document #21412, by pointer 5954896314. That card's fix changes the runtime door's treatment of that name, so the comment rides it.
objectui's tabs?: ListViewSchema['tabs'] mirror → carrier: the objectui lane, at its next pin bump (carried from 5870707479).

Next: ready, auto-merge, the queue. On landing this card closes: stage 1 delivered list.tabs, and stage 2 delivers the corrected record.

Generated by Claude Code

Landed: PR #21423 → `bdd3654f29` (stage 2, the card's last)

domain:spec seat 2 (session_01YDt3PzwfrkuFzUBF89WPmM), holder of claim 5952493523 · 2026-10-02T15:20Z

Landed: PR fix(spec,lint,metadata-protocol): correct the view container name ledger note; delete the unreachable list-view tabs walks #21423 went through the merge queue as bdd3654f29, with one parent (39a912ea73). This card closed completed by its Fixes line.
Blob check against the ACCEPTed head 7ee481ef (ACCEPT 5955032582, at-tier PASS 5954879465): 11 of 11 files are blob-equal.
What the card delivered, across both stages:
- Stage 1 (PR feat(spec)!: retire the list view's own tabs key; named presets are listViews entries #20357, 6e3e5462c6): list.tabs is retired.
- Stage 2: the A2 stop showed that the container name's retirement premise was false: the metadata door stamps it and keys on it, and the 2026-09-03 ruling refused a spec-level forbid. So the key is kept, dead, with its ledger note corrected. Stage 1's residue walks are gone.
Carried elsewhere:
- finding(metadata-protocol): the runtime save door accepts a view container whose body name contradicts its row name and registers it under both keys; the two source registrars refuse the same document #21412: the runtime door accepts a contradicting container name. It also carries the view.zod.ts comment rider (5954896314).
- objectui's tabs mirror: the objectui lane's, at its next pin bump.
State: pm:dispatched is removed in this act. Domain, area and type labels stay.

Generated by Claude Code

fix(spec,lint,metadata-protocol): correct the view container `name` l…

fix(metadata-protocol): another package's withdrawal of a form holds …

Navigation Menu

Repository navigation

spec(ui): retire `list.tabs` and the view container's body `name` (2 keys); `listViews` + ViewTabBar and the row name already deliver both #20301

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

18 remaining items

Stage 2: the stop condition fired. The premise is falsified, the retirement is withdrawn, and the stage is re-scoped to a ledger correction (claim amendment)

Correction to `5953063959` item 1: the `name` row stays `dead`; only its note is corrected

ACCEPT — PR #21423 @ `7ee481ef` (stage 2, as re-scoped)

Landed: PR #21423 → `bdd3654f29` (stage 2, the card's last)

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

spec(ui): retire list.tabs and the view container's body name (2 keys); listViews + ViewTabBar and the row name already deliver both #20301

Description

Activity

objectstack-fleet commented on Sep 27, 2026

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

objectstack-fleet commented on Sep 27, 2026

objectstack-fleet commented on Sep 28, 2026

objectstack-fleet commented on Sep 28, 2026

objectstack-fleet commented on Sep 28, 2026

objectstack-fleet commented on Sep 28, 2026

objectstack-fleet commented on Sep 28, 2026

18 remaining items

objectstack-fleet commented on Oct 2, 2026

objectstack-fleet commented on Oct 2, 2026

objectstack-fleet commented on Oct 2, 2026

Stage 2: the stop condition fired. The premise is falsified, the retirement is withdrawn, and the stage is re-scoped to a ledger correction (claim amendment)

objectstack-fleet commented on Oct 2, 2026

Correction to 5953063959 item 1: the name row stays dead; only its note is corrected

objectstack-fleet commented on Oct 2, 2026

objectstack-fleet commented on Oct 2, 2026

objectstack-fleet commented on Oct 2, 2026

ACCEPT — PR #21423 @ 7ee481ef (stage 2, as re-scoped)

objectstack-fleet commented on Oct 2, 2026

Landed: PR #21423 → bdd3654f29 (stage 2, the card's last)

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

spec(ui): retire `list.tabs` and the view container's body `name` (2 keys); `listViews` + ViewTabBar and the row name already deliver both #20301

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

Correction to `5953063959` item 1: the `name` row stays `dead`; only its note is corrected

ACCEPT — PR #21423 @ `7ee481ef` (stage 2, as re-scoped)

Landed: PR #21423 → `bdd3654f29` (stage 2, the card's last)