skills/objectstack-i18n: the group table misses fields.{field}.inlineTitle and reports.{report}.blocks.{block}.{label,description} (follow-up owed by PR #26757)

Filing gate: ① a shipped-surface gap with a named location, owed by in-flight work (PR #26757, cards #26112 + #26142). It is filed now so the PR's Skills-docs: line can name it. Filed by domain:spec seat 3 · os-support-ai · session session_01WGufzgbJhFoXYsLSe5WSPP. ⛔ Not a claim.

Business reason. An AI author who writes a translation bundle from skills/objectstack-i18n/SKILL.md will not translate a joined report's block headings or a master-detail field's inline grid title. The skill's group table does not name either key, so zh-CN users keep seeing those headings in English, even though the platform now serves them translated. The same kind of gap was closed for the 18.2 groups by #26296.

Reader who acts. The skills lane (skills/**, Tier H, which needs a human merge). Do this after PR #26757 lands, so the keys exist on main.

What to change (measured on PR #26757 head `bcbc8374`)

skills/objectstack-i18n/SKILL.md, the fields.{field_name} row: add inlineTitle, the title of a master-detail field's inline child grid. When it is not translated it keeps its authored text; there is no fallback to the child's plural label.
The same file, the sentence saying reports holds label + description: add reports.{report}.blocks.{block_name}.{label,description}. Block keys go by the block's required name, as pages.{page}.components.{id} does.
One example row for each, in the skill's example bundle if it has one.
Same gap on the hand-written author page (added by the seat from contract review 6445499937, ③7): content/docs/ui/translations.mdx:71 (objects.NAME.fields.FIELD.label / help / placeholder) misses inlineTitle, and :85 (reports.NAME.label / description) misses reports.NAME.blocks.BLOCK.label / description. content/docs/ui/ is not governed, so that half can land in its own small PR ahead of the skill edit.
Stale header line: packages/lint/src/validate-translation-references.ts:163 ("Leaf keys below the name are the schema's (label / description, strict)") now misses blocks.

Dedupe

search_issues on this repo: "objectstack-i18n skill translation keys group table missing reports blocks inlineTitle". 20 hits. The closest are #26296 (closed, the 18.2 group sweep of the same skill), #21583 (closed, _actions row), and #26112 / #26142 (the open source cards). None covers these two keys. New.

Generated by Claude Code

Triage: first grade — `priority:p1` · `domain:skills` · `area:i18n` · `pm:queue` · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T16:55Z

⛔ Not a claim, ⛔ not a dispatch.

p1, it rides with protocol additions (i18n: a master-detail field's inlineTitle has no translation slot — the inline child grids on a zh-CN create form read "Versions / Deviations / Payment Plan" #26142 inlineTitle, spec/i18n: a joined report's block labels have no translation key, so under zh-CN the showcase's "Task Overview" shows a translated title above English block headings ("Open Tasks", "Completed Tasks") #26112 report block labels): this is their owed Skills-docs: follow-up from PR feat(spec): a joined report's block headings and a master-detail field's inline grid heading are translatable #26757 ([Decision] pm-dispatch: which PRs owe a Skills-docs: line (the #26184 audit Part B rule draft) — protocol-kind only, protocol-kind plus narrowing, or not adopted #26369 B). Add the fields.FIELD.inlineTitle and reports.REPORT.blocks.BLOCK rows to the group table, with an eval prompt. Tier H.

Generated by Claude Code

Claim: PM loop round 1
Session: session_01ADHkmSrRBdu9MdC6fc33Uu
Account: os-steve (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/project-thread-jdbl6z (the card thread's harness push branch, self-reported at 2026-10-27T17:40Z — the measured platform exception to claude/issue-N-slug)
Worktree: objectstack-issue-26759
Domain: domain:skills
Seat: domain:skills#1
File surface: skills/objectstack-i18n/SKILL.md (the fields.{field_name} row gains inlineTitle; the reports sentence gains reports.{report}.blocks.{block_name}.{label,description}; one example row each if the skill has an example bundle; the card's hand-written-twin item as the body states it), an eval only if its file has headroom, its row in scripts/check-skills-token-ratchet.mjs if it moves (trim first, then raise at the measurement under the standing ruling 「A:先削重复,余额按实测提高」), and a patch changeset for @objectstack/skills; nothing else (stop on breach; explain in the report)
Container & model: XS, mode:thread, model: CONTRACT_REVIEW_TIER (dispatch-gates --tier: MANDATORY — skills/**, clause ①, no one-line exit), effort: high. Tier H (skills/**): draft PR; ⛔ the dev never readies, arms or enqueues it.
Clause-②: no
Metadata-protocol: none
Responsibility: n/a — not a defect card (the owed Skills-docs: follow-up of PR #26757)
Thread-read: 6448132939
Premise: holds (seat re-measure on origin/main 0ea90c7a: PR #26757 merged; git grep inlineTitle -- skills/objectstack-i18n/ → none; SKILL.md:175 names only label + description for reports)
Serial constraints cleared: the Blocked-by PR #26757 merged; no open PR names skills/objectstack-i18n/** (read over every open PR's files); #26790 (same file) is held behind this card

os-dev-report

{
  "issue": 26759,
  "status": "done",
  "branch": "claude/project-thread-jdbl6z",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/26832",
  "session": "session_01ADHkmPqouEb6yELnSqBwRY",
  "premise_still_valid": true,
  "summary": "skills/objectstack-i18n/SKILL.md now names both keys where a bundle author picks a group: the fields.{field_name} row gains inlineTitle (the inline child-grid heading a master_detail field draws on its parent's form, keyed on the child object; untranslated keeps its authored text, no fallback to the child's plural label) and the reports sentence gains reports.{report}.blocks.{block_name}.{label,description} (keyed by the block's required name), with one example row each spelled as the shipped app-crm / app-showcase zh-CN bundles spell them. content/docs/ui/translations.mdx (the claim's hand-written-twin item) gains the same two rows. The eval file gains case 8 covering both keys; case 6's duplicate --no-objects-only pin was cut first (77 bytes) and the ratchet row is pinned at the measurement 1451 -> 1642 under the standing ruling. Patch changeset for @objectstack/skills. Draft PR (Tier H, skills/**), never readied.",
  "tests": "node scripts/check-skills-token-ratchet.mjs at head 4f490d8e: exit 0, '81 authored bundle file(s) within their ceilings; 10 generator-owned file(s) measured, not ratcheted'; objectstack-i18n/SKILL.md 5456/6333 (headroom 877), bundles-locales-coverage.json 1642/1642 (headroom 0). --self-test: '79 cases pass', exit 0. Eval JSON parses; no must_contain string sits in a must_not_contain. No package source touched, no package test/typecheck owed; no forced test edit; no ablation (docs/config diff).",
  "gates": "local subset at head 4f490d8e: dispatch-gates --commands --repo objectstack-ai/objectstack derived 87 commands; all 87 ran serially (70 source-reading, then the 17 dist-reading ones after the build), each exit captured before any pipe, 87 x exit 0. --ran: '87 derived famil(ies) accounted for — 87 run, 0 NOT-MEASURED (a DERIVED zero — all 87 recorded an exit code and none of them is 3)', exit 0, no UNRUN on stderr. --tier: MANDATORY (skills/**, clause (1), no one-line exit); changed lines 51 vs 3000. Build closure: turbo run build --filter='./packages/**' --concurrency=2 under os-verify-lock, VERDICT command-exit 0, held 8m20s. CI families (wide-population, artifact roster, pnpm lint, the required nine) are the PR's check-runs: in_progress at report time, not re-run locally.",
  "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm",
  "mcp_calls": "6 — all reads: mcp__github__issue_read get x1 (#26759), get_comments x3 (claim wait), mcp__github__pull_request_read get x2 (#26774 body shape, #26832 read-back); zero MCP writes",
  "api_writes": "2 relay strokes + git push — (1) fleet-write pr_create fw-20261027T180431Z-6bffd2 run 40556538796: POST /repos/objectstack-ai/objectstack/pulls (draft) + the assignee leg POST /issues/26832/assignees (os-steve) + the relay's own Tier H governed-review request leg; (2) this os-dev-report comment via post-stamped.mjs --comment=26759 (POST /issues/26759/comments). No labels sent (none named by the dispatch; the diff publishes @objectstack/skills so no skip-changeset).",
  "open_questions": [
    { "question": "Where should the two example rows live — the card says 'in the skill's example bundle if it has one'?", "options": ["A: a two-line fragment fence after the group table, copied from the shipped app-crm / app-showcase zh-CN bundles (what the PR does)", "B: invent a master-detail field and a joined report into the todo-mirroring os:check bundle"], "recommendation": "A, because the skill's bundle mirrors examples/app-todo, which has neither key's source; B would teach names no shipped example carries." },
    { "question": "content/docs/ui/translations.mdx rides this Tier H PR rather than its own small docs PR (the card body allowed either).", "options": ["A: keep it here (the claim's file surface names it; one PR per card thread)", "B: the seat splits it into a docs-only PR to land ahead"], "recommendation": "A unless the seat wants the docs row on main before the governed approval; the split is a revert-and-cherry-pick for the seat, not the dev." }
  ],
  "out_of_scope_findings": [
    "class: none (stale code comment, not a defect) · packages/lint/src/validate-translation-references.ts L163 still says a report's leaf keys are label / description only; the validator reads the Zod shape, so behaviour is right and only the header is stale · carrier: seat (one-line comment edit, outside the claim's surface) · dedupe words: validate-translation-references blocks header stale report leaf keys"
  ]
}

Generated by Claude Code

Seat note · domain:skills#1 · session_01ADHkmSrRBdu9MdC6fc33Uu · 2026-10-27T18:07Z · claim amendment to 6449271046 (same branch claude/project-thread-jdbl6z)

File surface adds: packages/lint/src/validate-translation-references.ts:163, the header comment only (the card's "Stale header line" item: leaf keys under a report name now include blocks). Reason: PR #26832 says Fixes #26759, so the card's last item rides the same PR rather than being left behind a closed card. content/docs/ui/translations.mdx (the card's hand-written-twin item) was already in the claim and stays in this PR. No behaviour change; nothing else.

os-dev-report

Round 2 (seat review of PR #26832 + claim amendment 6449881440). Supersedes the round-1 report's gates, tests and out_of_scope_findings; everything else stands.

{
  "issue": 26759,
  "status": "done",
  "branch": "claude/project-thread-jdbl6z",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/26832",
  "session": "session_01ADHkmPqouEb6yELnSqBwRY",
  "premise_still_valid": true,
  "summary": "Head 73e5a8f1 (on top of 4f490d8e). Two seat items: (1) the claim amendment's file surface — packages/lint/src/validate-translation-references.ts Reports header (comment only) now lists `blocks.<block_name>.label` / `.description` by the block's `name` among a report's leaf keys, which the validator already accepts from the Zod shape; (2) SKILL.md's fence lead-in names the `inlineTitle` and report-block keys instead of 'Both deeper keys' (which read as the page:tabs / approval keys of the previous sentence), and a blank line now separates the closing fence from the `settings` paragraph. No behaviour change. The PR body's 'Not in this PR … L163' bullet and its Acceptance-notes 'stale comment, not filed' bullet are now stale: the dev writes the body once, so the seat is asked to drop both bullets (or note that the header rode this PR) — see open_questions.",
  "tests": "At head 73e5a8f1: node scripts/check-skills-token-ratchet.mjs exit 0 — objectstack-i18n/SKILL.md 5462/6333 (headroom 871; +6 tokens from the lead-in and the blank line), bundles-locales-coverage.json 1642/1642 unchanged, '81 authored bundle file(s) within their ceilings'. @objectstack/lint (the package whose source file carries the comment edit) under os-verify-lock: `pnpm --filter @objectstack/lint typecheck && pnpm --filter @objectstack/lint test` → typecheck OK (check:test-typecheck OK, debt ledger unchanged 2 files / 6 errors), `Test Files 168 passed (168) · Tests 7430 passed (7430)`, VERDICT command-exit 0 (held 1m59s). No forced test edit; no ablation (comment/docs diff).",
  "gates": "local subset at head 73e5a8f1: dispatch-gates --commands --repo objectstack-ai/objectstack derived 108 commands (the 87 of round 1 plus 21 the lint source path adds: check-comment-mask-adoption, check-issue-citations, check-plugin-teardown-shape, docs-audit check-affected-docs / check-drift-comment, check:duration-unit-keys, check:keyed-text-bounds, check:logger-receiver-detach, check:org-identifier, check:page-declaration-shape, check:platform-object-tenancy-census, check:registry-log-declared, check:rest-log-spy-declared, check:slot-lookup, check:system-context-census, check:test-source-alias, check:tier-file-adoption, check:undeclared-dep-imports …). All 108 ran serially after the commit (91 source-reading, 17 dist-reading after the build closure), each exit captured before any pipe: 108 x exit 0. One prerequisite round: `node scripts/check-plugin-teardown-shape.mjs --self-test` first exited 3 (PREREQUISITE NOT MET — its positive-control fixture is pinned to commit 621a4876, unreachable in the shallow clone); after `git fetch origin 621a4876…` it was re-run and passed ('48 cases pass', exit 0), and the ledger carries that exit. --ran: '108 derived famil(ies) accounted for — 108 run, 0 NOT-MEASURED (a DERIVED zero — all 108 recorded an exit code and none of them is 3)', exit 0, no UNRUN on stderr. Build closure: turbo run build --filter='./packages/**' --concurrency=2 under os-verify-lock, VERDICT command-exit 0 (held 4m55s, turbo cache hits). CI families are the PR's check-runs on 73e5a8f1, not re-run locally.",
  "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm",
  "mcp_calls": "cumulative 8 — all reads: mcp__github__issue_read get x1, get_comments x4 (claim wait x3, amendment read-back x1), mcp__github__pull_request_read get x2 (#26774, #26832), get_check_runs x1 (#26832, subscription check-in); zero MCP writes",
  "api_writes": "cumulative 3 relay strokes + 2 git pushes — round 1: pr_create (fw-20261027T180431Z-6bffd2) and the os-dev-report comment 6449843531; round 2: this comment via post-stamped.mjs --comment=26759 (POST /issues/26759/comments). No PR-body PATCH (the dev writes the body once; the seat edits it).",
  "open_questions": [
    { "question": "PR body text now stale after the amendment: the 'Not in this PR … validate-translation-references.ts L163' bullet under What changes and the 'Stale comment, not filed' bullet under Acceptance notes.", "options": ["A: the seat replaces the first with 'packages/lint/src/validate-translation-references.ts: the Reports header lists blocks.BLOCK.label / .description among a report's leaf keys (comment only; claim amendment 6449881440)' and deletes the second", "B: leave the body; the round-2 commit message carries the change"], "recommendation": "A, because the reviewer reads the body and it currently says the opposite of the diff." }
  ],
  "out_of_scope_findings": []
}

Generated by Claude Code

ACCEPT → PR #26832 comment 6450504495 · domain:skills#1 · 2026-10-27T18:37Z · Tier H; draft until a GOVERNED_APPROVERS approval; contract review requested from the director seat.

os-dev-report

Round 3 (director contract review 6450590107, VERDICT PASS, should-fix in this PR's own hunk). Supersedes the round-2 report's tests, gates, mcp_calls and api_writes; everything else stands.

{
  "issue": 26759,
  "status": "done",
  "branch": "claude/project-thread-jdbl6z",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/26832",
  "session": "session_01ADHkmPqouEb6yELnSqBwRY",
  "premise_still_valid": true,
  "summary": "Head 57406c6f (on 73e5a8f1). The director's should-fix: eval case 6's ideal answer said 'never `permissionSets`' while its must_not_contain lists `permissionSets`, so the ideal answer graded wrong. Reworded at the same byte length to 'not the camelCase name' (22 bytes for 22), so the file stays at its 1642-token pin and the ratchet row does not move. The dev's acceptance check now also covers every case's prompt and expected_output against its must_not_contain, and every must_contain string against its expected_output: 8 cases, 0 findings at this head (the round-1 check only compared the two assertion lists). No other file changed.",
  "tests": "At head 57406c6f: node scripts/check-skills-token-ratchet.mjs exit 0 — bundles-locales-coverage.json 1642/1642 (headroom 0, bytes 6567 before and after), '81 authored bundle file(s) within their ceilings'; --self-test '79 cases pass', exit 0. The eval file parses; per-case assertion cross-check 0 findings. No test file edited; no forced test edit.",
  "gates": "The changed set is the same five files as round 2, so the derivation is the same 108 families; at this head the dev re-ran the ratchet family (the only family whose reading a 22-byte same-length reword inside a JSON string can move): check-skills-token-ratchet and its --self-test, both exit 0. The remaining 106 were run at 73e5a8f1 (--ran 108/108, 0 NOT-MEASURED) and are CI's check-runs on 57406c6f, not re-run locally.",
  "dogfood_battery": "n/a — no `!` in the PR title and no (narrowing) arm",
  "mcp_calls": "cumulative 11 — all reads: issue_read get x1, get_comments x4, pull_request_read get x2, get_check_runs x2, get_comments x1 (#26832, the director record), get_job_logs x1 (run 40556644688: the two round-1 aggregator reds were cancelled lanes, superseded by the round-2 push); zero MCP writes",
  "api_writes": "cumulative 4 relay strokes + 3 git pushes — pr_create; os-dev-report comments 6449843531, 6450462144 and this one (post-stamped.mjs --comment=26759). No PR-body PATCH.",
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

Landed · domain:skills#1 · session_01ADHkmSrRBdu9MdC6fc33Uu · 2026-10-27T23:39Z — PR #26832 merged through the queue (merge commit 4b8f8cdc07) on the os-zhuang approval, after the director's PASS re-stamped on round 3 (6450750169); the i18n skill names fields.{field}.inlineTitle and a joined report's blocks.{block_name} keys, the author page and the lint header comment agree, eval case 8 pins both. The card closed by its Fixes line; pm:dispatched stripped in this act. #26790 (same SKILL.md) is dispatched next.

Navigation Menu

Repository navigation

skills/objectstack-i18n: the group table misses `fields.{field}.inlineTitle` and `reports.{report}.blocks.{block}.{label,description}` (follow-up owed by PR #26757) #26759

What to change (measured on PR #26757 head `bcbc8374`)

Dedupe

Triage: first grade — `priority:p1` · `domain:skills` · `area:i18n` · `pm:queue` · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T16:55Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

skills/objectstack-i18n: the group table misses fields.{field}.inlineTitle and reports.{report}.blocks.{block}.{label,description} (follow-up owed by PR #26757) #26759

Description

What to change (measured on PR #26757 head bcbc8374)

Dedupe

Activity

objectstack-fleet commented on Oct 27, 2026

Triage: first grade — priority:p1 · domain:skills · area:i18n · pm:queue · triage seat · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-27T16:55Z

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

objectstack-fleet commented on Oct 27, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

skills/objectstack-i18n: the group table misses `fields.{field}.inlineTitle` and `reports.{report}.blocks.{block}.{label,description}` (follow-up owed by PR #26757) #26759

What to change (measured on PR #26757 head `bcbc8374`)

Triage: first grade — `priority:p1` · `domain:skills` · `area:i18n` · `pm:queue` · triage seat · `session_01AavokzJ5DndAwitDXvKy4U` · 2026-10-27T16:55Z