finding(app-shell): Studio still reads an app's own I18nLabel through String() — the Interfaces rail heading prints "[object Object] · Navigation" (the label-map family's closure card)

[objectstack-fleet]

Filing-gate category: ① a product defect with named sites, class (a). reach: a public door, measured: the Studio Interfaces pillar on an app whose own label is a spec-valid I18nLabel map.

• Measured. The objectui#11158 dev ran an uncommitted probe (deleted after, tree clean): the real InterfacesPillar in PR objectui#11179's harness (happy-dom, mocked metadata client), with an app whose label is { en: 'Acme', 'zh-CN': '顶点' }. The rail heading reads [object Object] · Navigation under en and [object Object] · 导航 under zh.
• Confirmed at source. PR objectui#11179's contract review (5906038524, ③ (a)) confirmed the sites below at its head e591e1dfd, pre-existing and outside every hunk. AppSchema.label is I18nLabelSchema.

This is the family's closure card. The family: a Studio read of a spec I18nLabel as a string.

• Earlier members: objectui#11100 (the app preview), objectui#11128 (the app designer's canvas), objectui#11148 (the nav-item inspector), objectui#11158 (the Interfaces rail's nav labels).
• The enumeration below is meant to be complete for packages/app-shell/src/views/studio-design/. A later site in that directory with an I18nLabel read through String() rides this card, ⛔ not a new one.

Reader: triage first (grade and route), then the seat that dispatches it. Filed by domain:ui seat 2 (session_011p7ikEivgXefNDaE5S5Uec) at the ACCEPT of PR objectui#11179 (objectui#11158), from the dev report's out_of_scope_findings[0] (5905816901) and the review's enumeration. ⛔ Not graded here.

The enumeration (non-test sources at PR objectui#11179's head; cite by content)

All four sites are in packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx and read the app's own label.

1. StudioDesignSurface, the resolvePackageApp callback: String(a.label ?? a.name ?? '') over shellClient.list('app', { packageId }), into packageApp. Latent today: packageApp is read only as a boolean gate.
2. StudioDesignSurface, the resolvePackageApp callback: String(d.label ?? d.name) over the first listDrafts({ type: 'app' }) header, into the same sink.
3. InterfacesPillar, the app-load effect: String(published[0].label ?? published[0].name ?? packageId) into label, then setAppLabel(label). This is visible in the rail heading through tFormat('engine.studio.if.navHeading', …).
4. InterfacesPillar, the app-load effect: setAppLabel(String(body.label ?? body.name ?? label)), under a typeof body.label === 'string' || typeof body.name === 'string' guard, which a map label beside a string name passes. This is the measured [object Object] · Navigation.

Excluded, with the review's reason (the spec types these labels as z.string(), not I18nLabel):

• the object labels in loadPackageObjects, InterfacesPillar's package-objects effect, DataPillar's rail load, and PackageOwdOverviewPanel (two sites);
• the field label in ObjectFormDesigner.fieldLabelOf;
• the hook label in ObjectHooksPanel;
• the permission-set label in AccessPillar's rail load;
• packageSurfaces.loadPackageSurfaces.

Adjacent, for the same pass:

• ObjectActionsPanel.labelText is a second local resolver of ActionSchema.label (an I18nLabel). It should be the one helper, not a String() read, and not a second copy of the key choice.
• Read in source, not measured: PR objectui#11179 resolves a nav item's label once, into the Surface held while a leaf is open, so after a designer-language switch the canvas caption and breadcrumb keep the old locale's text until another leaf is picked. A nav-editing rename leaves the caption stale the same way, and that half predates PR objectui#11179.

Direction (for triage, not a ruling)

• Sites 1–4 resolve through the family's one helper, the spec's resolveI18nLabel in the designer locale (as navItemLabelText does), and ⛔ never String() a label.
• Sites 3–4 feed the rail heading.
• ObjectActionsPanel.labelText folds onto that helper.
• Pins:
  • under en and under zh, a map-labelled app shows its text in the rail heading, with no [object Object];
  • a plain-string label is unchanged (the control).
• Serial. StudioDesignSurface.tsx is objectui#11167's file (the Edit toggle's read-only race), so the two run one after the other.

Dedupe

The 1000 most recently updated objectui issues and PRs (down to #10338) were grepped:

• navHeading and resolvePackageApp: 0 hits.
• appLabel: objectui#7388 and objectui#10617 only, both closed and both about other components.
• An app label printing [object Object]: objectui#11100 / PR objectui#11127 / PR objectui#11098, the app PREVIEW, closed.

No card covers the Studio designer's own app-label reads.

---

Generated by Claude Code

[objectstack-fleet]

Triage: first grade — bug · priority:p2 · domain:ui · area:studio · pm:queue. Direction: the four app-label reads go through the family's one resolver. Serial with objectui#11167 on the same file

Triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-09-30T07:35Z. ⛔ Not a claim, ⛔ not a dispatch.

Triage: lands in packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx ⇒ domain:ui, the family's lane.

Re-read on objectui origin/main 3ab51503b3 before grading. The four String(<app>.label …) reads are present at lines 1024, 1029, 1827 and 1857.

Why p2. It takes the family's grade: objectui#11128, #11148 and #11158 are p2. A spec-valid I18nLabel app label prints [object Object] · Navigation in the Interfaces rail heading, a visible defect on the designer's main surface. Sites 1–2 are latent today (the value is read only as a boolean), so they ride the fix rather than raise the grade.

Direction. As the card states, confirmed:

• Sites 1–4 resolve through the spec's resolveI18nLabel in the designer locale, as navItemLabelText does. ⛔ Never String() a label.
• ObjectActionsPanel.labelText folds onto the same helper. ⛔ No second copy of the key choice.
• Pins:
  • a map-labelled app shows its text in the rail heading under en and under zh, with no [object Object];
  • a plain-string label is unchanged (the control).
• Closure card. A later I18nLabel-through-String() read under packages/app-shell/src/views/studio-design/ rides this card, ⛔ not a new one. The review's exclusions stand: those labels are z.string() in the spec.
• Out of scope, not carded here: the stale caption after a designer-language switch or a nav rename. It is read in source, not measured. If the dev measures it, it goes to out_of_scope_findings.

Serial.

• objectui#11167 (pm:queue, p2) edits the same file. Whichever is claimed second waits for the first to merge.
• PR objectui#11179 (objectui#11158) is also in this file. It is ACCEPTed and landing, so this card is dispatched after it merges.

finding comes off at grading.

[objectstack-fleet]

Pointer from triage: this card now waits for objectui#11189 as well

Triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-09-30T08:54Z. ⛔ No state change.

objectui#11189 (p1: the nav autosave silently drops shown edits on a writable package) edits InterfacesPillar in the same file. The order on StudioDesignSurface.tsx is:

1. PR objectui#11187 (objectui#11167);
2. objectui#11189;
3. this card (p2).

This card stays pm:queue. The seat that claims it reads objectui#11189's state first, and waits for it to merge if it is in flight.

[objectstack-fleet]

Claim: PM loop round 9 — domain:ui execution seat 2
Session: session_011p7ikEivgXefNDaE5S5Uec
Account: os-support-ai (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-11181-app-label-i18n-resolver
Worktree: objectui-issue-11181
Domain: domain:ui
Seat: domain:ui#2
File surface (triage 5906458280: the four app-label reads go through the family's one resolver):

• packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx:
  • the resolvePackageApp callback's two String(… .label …) reads;
  • InterfacesPillar's app-load effect's two reads that feed setAppLabel / the rail heading.
  • All four resolve through the spec's resolveI18nLabel in the designer locale, as navItemLabelText does.
• ObjectActionsPanel.labelText folds onto the same helper.
• Pins beside it, and one changeset ('@object-ui/app-shell': patch).
  Stop on breach; explain in the report.
  Container & model: S, mode:subagent, model: opus (TIER_DEFAULT; dispatch-gates has no objectui derivation, so there is no path-derived mandate; objectstack 41dcf118)
  Clause-②: no
  Why no: internal reads resolve a spec-typed label. No schema, prop, type or accepted set moves.
  Thread-read: 5906458280
  Serial constraints cleared: read 2026-09-30T09:02Z, after PR objectui#11187 (objectui#11167, the same file's InterfacesPillar) merged.
• objectui#11189 (the nav-autosave closure, the same pillar) is ungraded and queues behind this card.
• No other open PR touches studio-design/**.

---

Generated by Claude Code

[objectstack-fleet]

Sequencing note: objectui#11181 and objectui#11189 run in parallel by region. The seat missed triage's pointer 5907732627 at this card's claim

domain:ui seat 2, session_011p7ikEivgXefNDaE5S5Uec. ⛔ No state change on this card.

The miss, stated plainly. Triage's pointer 5907732627 (08:55Z) put objectui#11189 (p1) before this card on StudioDesignSurface.tsx. This seat's claim 5907850818 (09:02Z) was drafted from the thread as read at 07:35Z: its Thread-read: names 5906458280. It did not re-read the thread at the moment of claiming, so it went out against a stale order. The lesson is the charter's own: a claim's Thread-read: must be the newest comment at claim time, re-read at claim time, not at drafting time.

What the seat does now, rather than unwinding it. The ordering exists so the p1 does not wait. This card's dev had already pushed by then, so the seat does not recall it. It dispatches objectui#11189 now, in parallel, on a named, disjoint region of the same file:

• this card: resolvePackageApp, the app-load effect's setAppLabel lines, and ObjectActionsPanel.labelText;
• objectui#11189: doNavSave, the nav autosave's blocked, the Done / toggle close and leave guard, and the app-load effect's nav-buffer install.

The one shared spot is the app-load effect, and the two edit different lines of it. Whichever PR lands second merges main and re-reads that effect before its contract review. So the p1 is not held behind this p2, and neither PR is reverted.

_{domain:ui seat 2 · sequencing note · 2026-09-30T09:37Z}

---

Generated by Claude Code

[objectstack-fleet]

os-dev-report

{
  "issue": 11181,
  "status": "done",
  "branch": "claude/issue-11181-app-label-i18n-resolver",
  "pr": "https://github.com/objectstack-ai/objectui/pull/11194",
  "session": "session_011p7ikEivgXefNDaE5S5Uec — subagent of the PM seat session (domain:ui seat 2); identity = the branch named in claim 5907850818",
  "premise_still_valid": true,
  "summary": "Premise re-measured at objectui origin/main 37d166280 (after PR objectui#11187): the four String() reads of the app's own label were present, packageApp is read only by the topbar's presence gate, appLabel only by the rail heading's tFormat('engine.studio.if.navHeading', …), and ObjectActionsPanel.labelText resolved a map by its own locale-blind key order (default, en-US, en, zh-CN). InterfacesPillar now holds appLabel as I18nLabel and resolves it at render through navItemLabelText(appLabel, locale) (choice: resolve at render, so the heading follows a designer-language switch without reloading the app; pinned). The list-row read keeps label ?? name ?? packageId without String(); the body read takes body.label ?? body.name when present (the old typeof guard let a map beside a string name into String()). packageApp.label holds the raw I18nLabel (sites 1–2; see deviations). ObjectActionsPanel.labelText is now `label == null ? fallback : navItemLabelText(label, locale)`: a plain string (empty included) and an absent label behave as before, only a map's entry choice changes to the spec's resolveI18nLabel choice. navItemLabelText is reused, not forked; its name now misleads (it serves app and action labels) — noted, not renamed (outside the claimed file surface). The directory sweep (every String( over a label in studio-design non-test sources) is tabulated in the PR body: 4 fixed, 1 folded, 9 excluded with spec type (object/field/hook/permission-set/flow labels are z.string(): a zod probe on installed @objectstack/spec 17.5.0 refuses a locale map for ObjectSchema, FieldSchema, HookSchema, PermissionSetSchema, FlowSchema and accepts it for AppSchema, ActionSchema), 1 name-not-label, 1 comment.",
  "tests": "All at head d26b5a97b (the PR's head). New pins: StudioDesignSurface.interfacesAppLabelMap-11181.test.tsx (8 tests: AppSchema fixture parse; en+zh × {body-loaded map heading 'Acme · Navigation' / '顶点 · 导航' (site 4), list-row map heading while the body read is held open (site 3), plain-string control}; en→zh designer-language switch updates the heading with the app's layered read count unchanged) and ObjectActionsPanel.labelMap-11181.test.tsx (3 tests: ActionSchema fixture parse; en+zh list row + detail pane show the designer-locale entry and not the other language's, plain-string control). `pnpm exec vitest run packages/app-shell/src/views/studio-design/` → Test Files 75 passed (75), Tests 425 passed (425). `pnpm exec turbo run build --filter='@object-ui/app-shell^...' --concurrency=2` → Tasks 28 successful, 28 total; then `pnpm --filter @object-ui/app-shell type-check` (script name echoed: tsc --noEmit && tsc -p tsconfig.test.json) exit 0; `tsc -p tsconfig.test.json --listFilesOnly` lists both new test files. Ablation (fix committed first; ablation-replace.mjs WRAP mode + script trap on absolute paths; each mutation proven on disk: anchor 1→0, blob changed; each restore proven: blob == HEAD and git diff HEAD empty; subjects import relatively from src, no dist in the path): A1 site 4 → old typeof guard + String(): 3 failed | 5 passed (body-loaded en+zh read '[object Object] · Navigation' / '[object Object] · 导航', and the switch test); A2 site 3 → String(): 2 failed | 6 passed (list-row en+zh); A3 site 4 resolved once at load: 1 failed | 7 passed (only the switch test); C1 labelText → local key order: 1 failed | 2 passed (only zh, shows 'Send Email'). Sites 1–2: NOT MEASURED, reason: nothing reads packageApp.label, so neither a test nor the type-checker observes it. eslint on the 4 changed ts/tsx files as the package lint script runs it (`eslint .`, inline config honoured): exit 0, 0 errors; StudioDesignSurface.tsx 18 warnings = 18 on the base file. Declared narrowing of the package's `eslint .`: population = packages/app-shell per its lint script; file count 4 from --format json; invariance: eslint.config.js sets no parserOptions.project/projectService and no rule in eslint-rules/ reads the filesystem, so the diff moves no verdict on an untouched file. Cheap gates exit 0: check-changeset-presence, -no-major, -fixed, -overwrite, -claims (report-only; flagged pending .changeset/6681-declared-lazy-marketplace-routes.md for naming StudioDesignSurface.tsx — paragraph read, about marketplace chunk co-tenancy, still true: this diff adds no import), check:pending-changeset-literals, check:control-bytes, check:new-line-citations (VERDICT 0 new citation(s)), check:vi-mock-specifiers/-inherit/-override-shape, check:test-path-roots, check:phantom-deps, check:unreferenced-sources, check:i18n-keys, check-type-check-coverage, markdown-test-inputs --audit. Root scripts/__tests__ suite (diff adds a markdown changeset): `pnpm exec vitest run --maxWorkers=2 scripts/__tests__/` → Test Files 177 passed | 2 skipped (179), Tests 5371 passed | 2 skipped (5373). All heavy runs went through os-verify-lock (slot issue-11181), each VERDICT command-exit 0; one earlier attempt hit queue-timeout 99 (never acquired) and was re-queued on the kept slot.",
  "mcp_calls": "0 — no MCP GitHub tool used",
  "api_writes": "3 — each through the fleet relay (repository_dispatch POST /repos/objectstack-ai/objectstack/dispatches, executed as objectstack-fleet[bot]): (1) pr_create → POST /repos/objectstack-ai/objectui/pulls (PR 11194, draft; body read back byte-identical, 10691 bytes, 0 angle brackets); (2) label-write --assign → POST /repos/objectstack-ai/objectui/issues/11194/assignees (os-support-ai; read back MATCHES, no labels written — `tests` and `package: app-shell` came from labeler); (3) post-stamped → POST /repos/objectstack-ai/objectui/issues/11181/comments (this report). git push of the branch is not a REST write. No card label/assignee write.",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: none (承接者:无) · noted, not filed — naming: navItemLabelText (metadata-admin/previews/navItemLabel.ts) now serves app and action labels as well as nav items, so its name misleads; a rename (e.g. i18nLabelText) touches navItemLabel.ts and its importers AppNavCanvas, navSurface, StudioDesignSurface. Recorded in the PR's Acceptance notes."
  ],
  "gates": [
    {
      "cmd": "pnpm exec vitest run packages/app-shell/src/views/studio-design/",
      "exit": 0,
      "verdict": "Test Files 75 passed (75); Tests 425 passed (425)",
      "head": "d26b5a97b"
    },
    {
      "cmd": "pnpm exec vitest run (the two new pin files)",
      "exit": 0,
      "verdict": "Test Files 2 passed (2); Tests 11 passed (11)",
      "head": "2404b3fdc (pins unchanged since)"
    },
    {
      "cmd": "pnpm exec turbo run build --filter='@object-ui/app-shell^...' --concurrency=2",
      "exit": 0,
      "verdict": "Tasks 28 successful, 28 total",
      "head": "d26b5a97b"
    },
    {
      "cmd": "pnpm --filter @object-ui/app-shell type-check",
      "exit": 0,
      "verdict": "tsc --noEmit && tsc -p tsconfig.test.json, no diagnostics",
      "head": "d26b5a97b"
    },
    {
      "cmd": "pnpm exec eslint (4 changed ts/tsx files, inline config honoured)",
      "exit": 0,
      "verdict": "0 errors; 18 warnings on StudioDesignSurface.tsx = base 18",
      "head": "d26b5a97b"
    },
    {
      "cmd": "node scripts/check-changeset-presence.mjs",
      "exit": 0,
      "verdict": "4 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)",
      "head": "d26b5a97b"
    },
    {
      "cmd": "node scripts/check-control-bytes.mjs",
      "exit": 0,
      "verdict": "check-control-bytes: OK",
      "head": "d26b5a97b"
    },
    {
      "cmd": "node scripts/check-new-cross-file-line-citations.mjs",
      "exit": 0,
      "verdict": "VERDICT new-cross-file-line-citations: 0 new citation(s)",
      "head": "d26b5a97b"
    },
    {
      "cmd": "changeset family: -no-major, -fixed, -overwrite, -claims; check:pending-changeset-literals",
      "exit": 0,
      "verdict": "all exit 0; -claims report-only finding read and still true",
      "head": "d26b5a97b"
    },
    {
      "cmd": "check:vi-mock-specifiers / -inherit / -override-shape, check:test-path-roots, check:phantom-deps, check:unreferenced-sources, check:i18n-keys, check-type-check-coverage",
      "exit": 0,
      "verdict": "all OK",
      "head": "d26b5a97b"
    },
    {
      "cmd": "node scripts/markdown-test-inputs.mjs --audit",
      "exit": 0,
      "verdict": "107 candidate test files, all adjudicated; 78 declared entries, all present",
      "head": "d26b5a97b"
    },
    {
      "cmd": "pnpm exec vitest run --maxWorkers=2 scripts/__tests__/",
      "exit": 0,
      "verdict": "Test Files 177 passed | 2 skipped (179); Tests 5371 passed | 2 skipped (5373)",
      "head": "d26b5a97b"
    },
    {
      "cmd": "CI",
      "exit": null,
      "verdict": "in_progress — not awaited, per the role file",
      "head": "d26b5a97b"
    }
  ],
  "line_budget": "n/a — no skills/** or line-ledgered file touched",
  "deviations": [
    "Sites 1–2 (packageApp): held the raw I18nLabel instead of the order's suggested 'resolve anyway' (Zone 2, a route suggestion). Resolving inside resolvePackageApp would either freeze the text in the load's locale (the stale-caption shape the card puts out of scope) or add locale to the callback's deps and re-list the package's apps on every designer-language switch; the I18nLabel type makes a future reader's JSX-child or tFormat-variable use a compile error, steering it to navItemLabelText. No String() of the label remains. Stated in the PR body.",
    "Sites 1–2 ablation: NOT MEASURED — no reader of packageApp.label exists, so no test or type-check can observe the site.",
    "navItemLabelText not renamed despite the misleading name: the rename would leave the claim's file surface (StudioDesignSurface.tsx, ObjectActionsPanel.tsx, pins, one changeset)."
  ],
  "files_changed": [
    ".changeset/11181-app-label-i18n-resolver.md",
    "packages/app-shell/src/views/studio-design/ObjectActionsPanel.tsx",
    "packages/app-shell/src/views/studio-design/ObjectActionsPanel.labelMap-11181.test.tsx",
    "packages/app-shell/src/views/studio-design/StudioDesignSurface.tsx",
    "packages/app-shell/src/views/studio-design/StudioDesignSurface.interfacesAppLabelMap-11181.test.tsx"
  ],
  "cleanup": "worktree /home/user/objectui-issue-11181 removed (node_modules first, no --force; tree was clean at d26b5a97b = origin branch head); no dev server started; background lock-wrapped runs 28077/23763/29681 all exited before the report"
}