finding(platform-objects): five system objects declare nameField: id beside a titleFormat — once objectui honours ADR-0079's order (objectui#9436), their record-page H1 becomes the raw record id

Filed by the objectui domain:ui execution seat (seat #1, session session_01BA3nKVUwKQJf8DBxrSVtNC). The source is the census in objectui#9436's dev stop report (objectstack-ai/objectui#9436, comment 5820114435), taken at objectstack origin/main 61609edf. ⛔ Not graded and not routed: domain:*, priority:* and type are the triage seat's.

Fact

objectui is moving its record page to the order ADR-0079 and @objectstack/spec's own titleFormat describe declare: 「an explicit nameField now takes precedence」. That move is objectui#9436, ruling C1, maintainer 「同意」, and it is in flight now. The census of objectstack objects that declare both keys:

71 non-test source files carry a titleFormat key. 42 of them are platform objects that also declare a pointer (nameField / displayNameField).
26 of the 42 have a single-field template equal to the pointer, so nothing visible changes.
16 render something else, so their H1 will move. Eleven move to another field's value: sys_import_job, sys_job_queue, sys_job_run, sys_session, sys_setting, sys_migration_journal, sys_activity, sys_audit_log, sys_comment, sys_sharing_rule, sys_webhook.
⚠️ Five declare nameField: id, so their H1 becomes the raw record id: sys_approval_action, sys_approval_approver, sys_approval_request, sys_automation_run, sys_http_delivery.

Why it is objectstack's

The metadata is objectstack's. The spec describe itself advises the remedy: express the title as a formula nameField instead of a render-only template. objectui cannot fix a declared nameField: id without breaking ADR-0079's order.

⛔ Not claimed

⛔ No per-object remedy is chosen. The five nameField: id objects are the ones that visibly regress. The other eleven may be intended.
⛔ The census was taken by the objectui dev from the objectstack tree at 61609edf. Re-run it before acting on it.
⛔ This card does not ask objectui#9436 to wait. The landing order is noted in that card's round report for the maintainer, who can veto.

Dedupe words

nameField id titleFormat · ADR-0079 record title raw id · sys_approval_request H1 · formula nameField migration · titleFormat nameField census

分诊首次定级：`priority:p2` · `bug` · `domain:services` · `pm:queue` —— 五个系统对象声明了 `nameField: id`，旁边另有 `titleFormat`；objectui 按 ADR-0079 改成「声明的名称字段优先」（objectui#9436）之后，这五种记录页的大标题会变成原始的记录 id

Path: packages/plugins/plugin-approvals/src/（sys-approval-action / sys-approval-approver / sys-approval-request）· packages/services/service-automation/src/sys-automation-run.object.ts · packages/services/service-messaging/src/objects/http-delivery.object.ts

Triage: the five objects that visibly regress all sit in plugin-approvals / packages/services/* ⇒ domain:services (lane table), bug, priority:p2, pm:queue (finding removed — graded); rationale: objectui#9436 (ruling C1, maintainer 「同意」, in flight) makes the record page honour the declared pointer over titleFormat, as the spec's own describe says ("an explicit nameField now takes precedence") — objects that declare nameField: 'id' beside a real titleFormat (e.g. sys_approval_request: nameField: 'id', titleFormat: '{process_name} · {record_id}') will then show the raw id as their H1; objectui cannot fix a declared pointer without breaking ADR-0079, so the metadata is objectstack's to change, ideally before or with that landing.

分诊席 #6015，2026-09-24T20:29Z。⛔ 不认领、不派发。本席读完了卡面（本卡尚无评论），并在 objectstack origin/main 44639665ee 上核对。

本席核对

sys-approval-request.object.ts：nameField: 'id', // [ADR-0079] canonical primary-title pointer …，下一行 titleFormat: '{process_name} · {record_id}'。
main 上声明 nameField: 'id' 的对象文件，本席实测有 8 个：
- sys-migration（platform-objects）；
- sys-approval-action / -approver / -request / -token（plugin-approvals）；
- sys-automation-run、sys-flow-dispatch（service-automation）；
- http-delivery（service-messaging）。
卡面点名的五个都在其中。另外三个（sys_migration、sys_approval_token、sys_flow_dispatch）要看它们有没有 titleFormat，接手时一并测量。
卡面的全量普查（71 个文件带 titleFormat，其中 16 个标题会变）是 objectui 开发在 61609edf 上做的，卡面要求动手前重跑。

路由说明

卡面标题写的是 platform-objects。但真正会「退化成原始 id」的五个对象，都在 plugin-approvals 和 packages/services/*，按车道表属于 domain:services。另外 11 个「标题会换成别的字段」的对象多在 platform-objects（domain:engine），可能是有意为之，由接手者逐个判断。需要改的，另开给 domain:engine 的卡或在 PR 里说明。

定级说明

p2：objectui#9436 一落地，审批请求、自动化运行、HTTP 投递这些记录页的大标题就会显示成一串 id，用户看得见的退化。不给 p1：不影响数据。

执行要点

重跑普查：在当时的 main 上重跑卡面的普查，确认受影响对象的最新清单。
给 nameField: id 的对象换成有意义的名称字段：按 spec describe 的建议，把 titleFormat 表达的标题改成一个公式型 nameField，⛔ 不要继续依赖只用于渲染的模板。每个对象的具体写法由接手者决定。
时间上：最好在 objectui#9436 落地之前或同时完成。卡面说那张卡不会为此等待，落地顺序已写进它的回合报告，由维护者决定。
钉住：每个改过的对象，记录页标题取到的是有意义的名称，⛔ 不是 id。

Generated by Claude Code

Claim: PM loop round 1
Session: session_01Evb5jFDZGKQE9KG4jbMfMF
Branch: claude/issue-20015-name-field-id
Worktree: objectstack-issue-20015
Domain: domain:services
Seat: domain:services#1
File surface: the services-lane object definitions that declare nameField: 'id', namely packages/plugins/plugin-approvals/src/sys-approval-{action,approver,request,token}.object.ts, packages/services/service-automation/src/sys-automation-run.object.ts and sys-flow-dispatch.object.ts, and packages/services/service-messaging/src/objects/http-delivery.object.ts; the tests that pin their pointers (e.g. sys-automation-run-pointer-pair-verdict.test.ts); new test file(s); .changeset/20015-*.md. ⛔ Not packages/platform-objects (sys-migration.object.ts is domain:engine: report it, do not edit). ⛔ No packages/spec. Measurement first: re-run the census on current main. (Stop on breach; explain in the report.)
Container & model: M, mode:subagent, model: opus (dispatch-gates --tier: no path-derived mandate ⇒ default tier)
Clause-②: no
Thread-read: 5821737695
Serial constraints cleared at 2026-09-24T21:34Z: over all 19 open PRs, none touches any file above. objectui#9436 (objectui lane) is the consumer this card lands ahead of; ⛔ it is not held.

os-dev-report
{
"issue": 20015,
"status": "done",
"branch": "claude/issue-20015-name-field-id",
"pr": "#20042",
"session": "session_01Evb5jFDZGKQE9KG4jbMfMF — the dispatching seat's session (this run is a subagent; parent's id per the role file)",
"premise_still_valid": true,
"summary": "Census re-run on main 9d81af7 reproduced the card: 8 object files declare nameField 'id' (each mirrored by displayNameField 'id'), and 5 of them carry a composite titleFormat, so under ADR-0079's order their record-page H1 would be the raw id. Each of the 5 (sys_approval_request / _action / _approver, sys_automation_run, sys_http_delivery) now declares display_title, a formula with returnType 'text' reproducing its titleFormat (null-guarded where a source column is nullable), and points nameField and the displayNameField mirror at it; titleFormat is kept unchanged, and the formula adds no stored column (measured on the synced SQLite table). sys_approval_token and sys_flow_dispatch (no titleFormat) and sys_migration (domain:engine, titleFormat '{id}') are left with one-line verdicts, and the one existing pointer pin (sys-automation-run-pointer-pair-verdict) is re-judged with a note, not dropped. Assignee read as os-sales and untouched; the newest Claim names this branch; no spec capability was missing, so no needs_decision.",
"census": "| Object | File | Lane | titleFormat | H1 under ADR-0079 order, before | Verdict |\n|:---|:---|:---|:---|:---|:---|\n| sys_approval_request | plugin-approvals/src/sys-approval-request.object.ts | services | {process_name} · {record_id} | raw id | changed → display_title |\n| sys_approval_action | plugin-approvals/src/sys-approval-action.object.ts | services | {action} · {step_name} | raw id | changed → display_title |\n| sys_approval_approver | plugin-approvals/src/sys-approval-approver.object.ts | services | {approver} · {request_id} | raw id | changed → display_title |\n| sys_automation_run | service-automation/src/sys-automation-run.object.ts | services | {flow_name} · {node_id} | raw run id | changed → display_title |\n| sys_http_delivery | service-messaging/src/objects/http-delivery.object.ts | services | {label} → {url} | raw delivery UUID | changed → display_title |\n| sys_approval_token | plugin-approvals/src/sys-approval-token.object.ts | services | none | raw id, same as today | left: no titleFormat, so the id is today's title too; no regression |\n| sys_flow_dispatch | service-automation/src/sys-flow-dispatch.object.ts | services | none | raw dispatch key, same as today | left: no titleFormat; the dispatch key is the row's identity by design |\n| sys_migration | platform-objects/src/system/sys-migration.object.ts | engine | {id} | raw id, same as titleFormat | reported, not edited (domain:engine): the template is the id itself, so no regression |",
"tests": "All at HEAD a02c867 (the final commit). New pins: plugin-approvals src/sys-approval-display-title.test.ts (5 tests), service-automation src/sys-automation-run-display-title.test.ts (4) plus the re-judged src/sys-automation-run-pointer-pair-verdict.test.ts (4), service-messaging src/objects/http-delivery-display-title.test.ts (4): real ObjectQL + SqlDriver(better-sqlite3 :memory:) with the real object declarations; each asserts resolveDisplayField -> 'display_title', findOne H1 equals the intended text (e.g. 'flow:deal_desk · opp_42') and is not the row id, equals the titleFormat rendering, resolveRecordTitle agrees, the nullable leg, and display_title absent from columnInfo(). Full consumer suites (vitest run --maxWorkers=2, under os-verify-lock, VERDICT command-exit 0 each): plugin-approvals 50 files / 785 passed; service-automation 145 files / 1729 passed; service-messaging 45 files / 483 passed. typecheck of all three (incl. check:test-typecheck ledgers): VERDICT command-exit 0; tsc --listFiles confirms every new test file is inside a typecheck program. Ablations (committed state; scripts/ablation-replace.mjs wrap mode inside an outer trap-restore script; relative source import, so no dist on the resolution path and no rebuild): 5 pointer reverts nameField 'display_title' -> 'id' all red (request 2/5 'expected 1noNVpCJW8_iN7Aq to be flow:deal_desk · opp_42'; action 2/5; approver 2/5; run 4/8 incl. the re-judged pointer-pair pin 'expected id to be display_title'; http 3/4); 3 stored-column mutations Field.formula -> Field.text all red incl. 'expected [ id, created_at, …(20) ] to not include display_title' (request 3/5, run 5/8, http 4/4); every leg printed 'ok mutation landed' (anchor 1 -> 0, blob changed) and 'ok restored: blob == HEAD'; post-batch git status clean. Gates: dispatch-gates --commands derived 62 families; all 62 run; --ran reconcile: '62 derived, 62 run, 0 NOT-MEASURED, 0 UNRUN' (check:dual-build-cjs-loads first PREREQUISITE NOT MET exit 3 on six unrelated missing dists, re-run green after building them: '104 published require entry point(s) across 67 package(s) load'). Also: check:i18n OK (9 packages in sync); check:i18n-coverage OK (621 baselined, none new); check:i18n-stale-fill OK; GITHUB_TOKEN=$GH_TOKEN node scripts/check-issue-citations.mjs exit 0; check:nul-bytes OK; check:test-source-alias OK; check:cross-package-test-inputs OK; check-empty-changeset: 1 declaring changeset added, none modified. Narrowed eslint (repo-wide pnpm lint is CI's): population = the 17 changed .ts files, none ignored by eslint.config.mjs (json messages carry no 'ignored' notice); --format json count 17 files / 0 errors / 0 warnings; invariance: eslint.config.mjs never enables type-aware linting (its own statement near line 328), so this diff cannot move a verdict on an untouched file. Derivation note: tree is 4 commits behind origin/main; the one stale input is scripts/engine-double-contract.pinned.json, and this diff adds no fake engine. CI: not awaited (in_progress at report time), per the role file.",
"mcp_calls": "0 — no MCP GitHub calls",
"api_writes": "2 — both through the fleet-write relay as objectstack-fleet[bot]: (1) POST /repos/objectstack-ai/objectstack/dispatches (204) executing pr_create -> POST /repos/objectstack-ai/objectstack/pulls draft:true = PR #20042 (run 36070248629, success); (2) this os-dev-report comment, post-stamped -> POST /repos//issues/20015/comments. label-write: 0 (the dispatch names no label beyond the role file's, and skip-changeset does not apply because the diff publishes). git push x4 to the branch (not REST). Reads: plain REST GETs (issue, comments, pulls?head=).",
"open_questions": [
{
"question": "Clause-② is carried as 'no' from the claim. The change adds display_title, a read-only virtual field, to five objects' read and write-response shapes. Does that count as public-surface growth?",
"options": [
"A keep 'no' (patch): no accepted input widens, since a formula field is not writable; the field only makes the declared title readable",
"B 'yes' (minor): treat any new field in a served object shape as surface growth"
],
"recommendation": "A, because the declaration answers whether the accepted set or public surface widens, and this fix only realises a title the objects already declared in titleFormat. The PR and changeset carry 'no' as the claim dictates; flip both lines if the seat rules B."
}
],
"out_of_scope_findings": [
"class: b · Seam: spec:ObjectSchema.titleFormat (describe: 'an explicit nameField now takes precedence') → runtime:SchemaRegistry.registerObject provisionPrimary(synthesize:false), served via materializeServedObjectOnto (packages/spec/src/data/display-name.ts:331: 'A body served by a /meta READ EXIT is already designated') | renderer: objectui record page (objectstack-ai/objectui#9436) · evidence: registering each declaration in a real ObjectQL registry at a02c867 stamps nameField 'id' on 19 objects that declare titleFormat and NO pointer, so the served body presents a pointer the author never wrote as if explicit, and the renderer draws the raw id as H1, the same regression as #20015, reachable on the same premise (objectui#9436 landing). By lane: platform-objects (domain:engine) sys_account, sys_business_unit_member, sys_invitation, sys_member, sys_scim_group_member, sys_scim_projection_grant, sys_scim_subject, sys_team_member, sys_two_factor, sys_verification; plugin-approvals sys_approval_delegation; plugin-security sys_position_permission_set, sys_user_permission_set, sys_user_position; service-messaging sys_notification_delivery, sys_notification_preference, sys_notification_receipt, sys_notification_subscription; service-realtime sys_presence. Remedy has two readings (per-object formula titles, as this PR does, vs a designation pass that never stamps id), so it is not fixed here and was outside the claim's file surface. Dedupe words: titleFormat no nameField designated id · provisionPrimary designate id H1 · registry stamps nameField id titleFormat · ADR-0079 derived pointer served as explicit · sys_notification_delivery record title raw id",
"carrier: 承接者:无 · noted in PR Acceptance notes, not filed — six more no-pointer titleFormat objects get a non-id field designated, so their H1 moves to that field: sys_attachment→file_name, sys_report_schedule→name, sys_audience_binding_suggestion→permission_set_name, sys_record_share→object_name, sys_share_link→label, sys_notification_template→subject (same shape as the card's eleven 'moves to another field' objects; may be intended)",
"carrier: 承接者:无 · noted in the PR i18n section, not filed — sys_approval_approver and sys_approval_token are outside plugin-approvals' i18n extract config (and sys_automation_run has no bundle at all), so their labels, the new display_title label included, ship untranslated; pre-existing scope, unchanged by this PR"
],
"deviations": "Commit trailers use AGENTS.md's model-free pair (Claude-Session + 'Co-authored-by: Claude'); the harness reminder's model-named Co-Authored-By was not used because the pre-push hook refuses a model identifier. The PR body ends with the role file's session-URL footer, not the harness's footer form. Worktree ../objectstack-issue-20015 is removed after this comment (node_modules first, no --force).",
"pr_body_new": "Fixes #20015\n\nClause-②: no\n\n## What this changes\n\nADR-0079 resolves a record's title as nameField, then displayNameField, then a derivation. An explicit nameField takes precedence over the render-only titleFormat. The spec says so in the titleFormat describe (packages/spec/src/data/object.zod.ts:2149 on main):\n\n> [DEPRECATED → nameField (ADR-0079)] Render-only title template; the server cannot return or query it, and an explicit nameField now takes precedence. Migrate a single-field title to nameField, a composite to a formula field designated as nameField.\n\nobjectstack-ai/objectui#9436 makes the record page follow that order. Five services-lane objects declared nameField: 'id' beside a composite titleFormat, so their record-page H1 would become the raw id.\n\nEach of the five now declares display_title, a formula field with returnType: 'text' that reproduces its titleFormat. nameField and the displayNameField mirror point at it. This is the migration the describe prescribes.\n\n| Object | Formula (display_title) |\n|:---|:---|\n| sys_approval_request | record.process_name + ' · ' + record.record_id |\n| sys_approval_action | record.step_name != null ? record.action + ' · ' + record.step_name : record.action |\n| sys_approval_approver | record.approver + ' · ' + record.request_id |\n| sys_automation_run | record.node_id != null ? record.flow_name + ' · ' + record.node_id : record.flow_name |\n| sys_http_delivery | record.label != null ? record.label + ' → ' + record.url : record.url |\n\nWhere a source column is nullable, a row without it is titled by the other column alone, so the formula does not fail to evaluate.\n\n- No stored column. A formula is virtual. SqlDriver.createColumn emits nothing for type: 'formula', and the new tests read the synced table's columns to confirm display_title is absent. No migration runs.\n- titleFormat is unchanged on all five objects, for renderers that still read it first. The new tests hold the formula and the template to the same text for a fully populated row.\n- $search is unchanged. Neither id nor a formula is ever a search target (SEARCH_AUTO_EXCLUDED_FIELDS, SEARCH_VIRTUAL_TYPES), so the lead and the membership of the auto-default set are identical.\n- No server-side sort reads the title pointer. A formula cannot be an ORDER BY target, and a grep for sorts over nameField / displayField / titleField found none.\n- The formula form follows the ADR-0079 worked example (content/docs/data-modeling/formulas.mdx, display_title with returnType: 'text') and the system-object import precedent for expression tags (import { P } from '@objectstack/spec' in sys-permission-set.object.ts). No system object declared a formula field before this PR, and no object in this repo designated a formula nameField.\n\n## Census (re-run on main 9d81af714f)\n\nEvery object definition declaring nameField: 'id' (all eight also declare displayNameField: 'id'):\n\n| Object | File | Lane | titleFormat | H1 under ADR-0079 order, before | Verdict |\n|:---|:---|:---|:---|:---|:---|\n| sys_approval_request | plugin-approvals/src/sys-approval-request.object.ts | services | {process_name} · {record_id} | raw id | changed → display_title |\n| sys_approval_action | plugin-approvals/src/sys-approval-action.object.ts | services | {action} · {step_name} | raw id | changed → display_title |\n| sys_approval_approver | plugin-approvals/src/sys-approval-approver.object.ts | services | {approver} · {request_id} | raw id | changed → display_title |\n| sys_automation_run | service-automation/src/sys-automation-run.object.ts | services | {flow_name} · {node_id} | raw run id | changed → display_title |\n| sys_http_delivery | service-messaging/src/objects/http-delivery.object.ts | services | {label} → {url} | raw delivery UUID | changed → display_title |\n| sys_approval_token | plugin-approvals/src/sys-approval-token.object.ts | services | none | raw id, same as today | left: no titleFormat, so the id is today's title too; no regression |\n| sys_flow_dispatch | service-automation/src/sys-flow-dispatch.object.ts | services | none | raw dispatch key, same as today | left: no titleFormat; the dispatch key is the row's identity by design |\n| sys_migration | platform-objects/src/system/sys-migration.object.ts | engine | {id} | raw id, same as titleFormat | reported, not edited (domain:engine): the template is the id itself, so no regression |\n\n## Pointer pins, re-judged\n\n- service-automation/src/sys-automation-run-pointer-pair-verdict.test.ts pinned SysAutomationRun.nameField to 'id' as one leg of the #11386 "pointer pair stays undeclared" verdict ("no natural key a seed dataset could address its rows by"). Re-judged, not dropped. The leg still holds: a formula has no stored column a seed row could match, and the seed loader's default externalId is name, which this object does not have. The pin now asserts that the pointer names display_title and that its type is formula, so it goes red if the pointer ever names a stored column. The object's own comment carrying the same leg is updated to match.\n- No other test pins these objects' pointers. grep -rln \"sys_approval_request\\|sys_automation_run\\|sys_http_delivery\" packages --include=*.test.ts finds 37 files in 15 other packages. None reads these objects' nameField / displayNameField / titleFormat or enumerates their real field sets. They use hand-built fixtures or mention the names in prose.\n\n## Tests\n\nNew, one per package, each through the real ObjectQL engine on in-memory SQLite and the real object declarations:\n\n- plugin-approvals/src/sys-approval-display-title.test.ts\n- service-automation/src/sys-automation-run-display-title.test.ts\n- service-messaging/src/objects/http-delivery-display-title.test.ts\n\nPer object, each test asserts:\n\n- resolveDisplayField resolves display_title;\n- a seeded row read back with findOne carries the H1 the titleFormat intended (for example flow:deal_desk · opp_42), and that H1 is not the row's id;\n- the H1 equals the titleFormat rendering of the same row;\n- the server-side accessor resolveRecordTitle agrees;\n- the nullable-column leg titles the row by the other column;\n- display_title is absent from the synced table's columns.\n\nRuns at a02c8673d7:\n\n| Suite | Result |\n|:---|:---|\n| @objectstack/plugin-approvals, full (vitest run --maxWorkers=2) | 50 files, 785 tests passed |\n| @objectstack/service-automation, full | 145 files, 1729 tests passed |\n| @objectstack/service-messaging, full | 45 files, 483 tests passed |\n| typecheck, all three packages (incl. the check:test-typecheck ledgers) | exit 0 |\n\ntsc --listFiles confirms every new test file is inside a typecheck program: service-messaging tsconfig.json, plugin-approvals tsconfig.test.json, and service-automation tsconfig.test.json and tsconfig.json.\n\n### Ablations (committed state, scripts/ablation-replace.mjs, each restore proven blob == HEAD)\n\nThe object files are imported relatively from source, so no dist/ sits on the resolution path and no rebuild is involved.\n\n| Leg | Mutation | Result |\n|:---|:---|:---|\n| ptr-request | nameField: 'display_title' → 'id' | red 2/5: expected 'id' to be 'display_title'; expected '1noNVpCJW8_iN7Aq' to be 'flow:deal_desk · opp_42' |\n| ptr-action | same | red 2/5: H1 '9kMVUkyx9hCJHlrr', expected 'approve · manager_review' |\n| ptr-approver | same | red 2/5: H1 'CQJXqZmwf6e_0vl4', expected 'role:finance · N3bOxv5ejSt1FZj9' |\n| ptr-run | same | red 4/8, including the re-judged pointer-pair pin (expected 'id' to be 'display_title') |\n| ptr-http | same | red 3/4: H1 'JNhfRQXBccs860NR', expected 'order.created → https://hooks.example.com/in' |\n| col-request | Field.formula( → Field.text( | red 3/5, including expected [ 'id', 'created_at', …(20) ] to not include 'display_title' |\n| col-run | same | red 5/8, including the stored-column pin |\n| col-http | same | red 4/4, including the stored-column pin |\n\nEvery leg printed ok mutation landed before its run and ok restored: blob == HEAD after it. The post-batch tree was clean.\n\n## Gates\n\nnode scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --commands from the real diff derived 62 families. All 62 were run, and --ran reconciles: "62 derived, 62 run, 0 NOT-MEASURED, 0 UNRUN". check:dual-build-cjs-loads first answered PREREQUISITE NOT MET (six unrelated packages had no dist/). It was re-run green after building them: "104 published require entry point(s) across 67 package(s) load".\n\nAlso run:\n\n- check:i18n-coverage: OK, 621 baselined, none new;\n- GITHUB_TOKEN=\"$GH_TOKEN\" node scripts/check-issue-citations.mjs: exit 0;\n- narrowed eslint over the 17 changed .ts files (--no-inline-config --format json): 17 files, 0 errors, 0 warnings, 0 ignored. The config never enables type-aware linting (eslint.config.mjs, around line 328), so a file's verdict depends on that file alone. The repo-wide pnpm lint is CI's.\n\nThe derivation noted the tree is a few commits behind origin/main. The one stale input is scripts/engine-double-contract.pinned.json, and this diff adds no fake engine.\n\n## i18n\n\ndisplay_title label and help are extracted into the plugin-approvals (request, action) and service-messaging bundles, and translated by hand in zh-CN, ja-JP and es-ES. A second --write returned the source-hash companions to their prior bytes, because none of the new leaves is a fill. sys_approval_approver and sys_automation_run are outside any extract config, before and after this PR.\n\n## Acceptance notes\n\n- Finding for the seat (not filed from here): 19 objects declare titleFormat with no pointer, and the registry stamps nameField: 'id' on them. Measured by registering each declaration in a real ObjectQL registry at a02c8673d7. The designate-only pass (provisionPrimary(…, { synthesize: false })) derives id as the first title-eligible field. A body served by a /meta read carries that stamp (packages/spec/src/data/display-name.ts:331), so the renderer in objectstack-ai/objectui#9436 would draw the raw id as the H1 for these too, the same regression as this card. The 19 are, by lane:\n - platform-objects: sys_account, sys_business_unit_member, sys_invitation, sys_member, sys_scim_group_member, sys_scim_projection_grant, sys_scim_subject, sys_team_member, sys_two_factor, sys_verification;\n - plugin-approvals: sys_approval_delegation;\n - plugin-security: sys_position_permission_set, sys_user_permission_set, sys_user_position;\n - service-messaging: sys_notification_delivery, sys_notification_preference, sys_notification_receipt, sys_notification_subscription;\n - service-realtime: sys_presence.\n\n Not fixed here. They are outside the claim's file surface, and the remedy has two readings: per-object formula titles, or a designation pass that never stamps id.\n- Six more no-pointer objects get a non-id field designated, so their H1 moves to that field's value: sys_attachment → file_name, sys_report_schedule → name, sys_audience_binding_suggestion → permission_set_name, sys_record_share → object_name, sys_share_link → label, sys_notification_template → subject. This is like the card's eleven "moves to another field" objects, which may be intended. Noted only.\n- Clause-②: no is carried from the claim. The change adds a read-only virtual field to five objects' read shapes, and widens no accepted input.\n"
}

Generated by Claude Code

Review — ACCEPT (landing gated on CI) · PR #20042 · head `a02c8673d7` · `domain:services` seat · 2026-09-24T23:04Z

_{Seat session_01Evb5jFDZGKQE9KG4jbMfMF. Checked against GitHub and origin/main, ⛔ not against the report's own account.}

Checklist

PR shape. Draft, targeting main. The first line is Fixes #20015, and Clause-②: no starts its own line. No other closing keyword.
Scope. 18 files, +580 / −17, all in the three claimed packages:
- five object definitions and the re-judged pointer pin;
- three new test files;
- generated translation bundles;
- .changeset/20015-name-field-id.md (patch for each package).
No platform-objects (sys_migration is reported, with the verdict "the template is the id itself, no regression"). No packages/spec. NOT governed, 597 lines, merges clean over origin/main.
Census re-run on main. 8 objects declare nameField: 'id', and 5 of them have a composite titleFormat. Those 5 changed. The other 3 are left with one-line verdicts.
Diff, read line by line.
- Each of the five objects gains display_title, a Field.formula with returnType: 'text' reproducing its titleFormat, null-guarded where a source column is nullable. nameField and the displayNameField mirror point at it.
- titleFormat is kept.
- This is the migration the titleFormat describe prescribes: 「a composite to a formula field designated as nameField」.
- A formula is virtual, and the tests read the synced columns to confirm it: no stored column, no migration.
- I checked the consumer paths myself. The engine plans formula projections: a projection naming display_title is rewritten and computed after the query. So lookup, audit and approval title reads that select the name field are served. $search excludes virtual types.
Tests. In three packages, a real ObjectQL + SQLite with the real declarations: the resolved field, the H1 equal to the titleFormat rendering (not the id), resolveRecordTitle agreeing, the nullable leg, and no stored column. Ablations: 5 pointer reverts and 3 stored-column mutations, all red, every restore blob-identical. plugin-approvals (785), service-automation (1729) and service-messaging (483) are green.

Open question, answered: A (Clause-②: no, patch). No accepted input widens: a formula is not writable. The field realises a title the objects already declared in titleFormat. No spec accept set moves.

Out-of-scope findings, one line each:

19 objects that declare titleFormat and no pointer get nameField: 'id' stamped by the registry's designation pass, and would show the raw id under objectui#9436: filed #20044 (class b, cross-lane; the remedy has two readings, routed by triage).
Six no-pointer objects whose title moves to another field: carried in ADR-0079 title: 19 objects that declare titleFormat and NO pointer get nameField: id stamped by the registry designation pass, so a renderer honouring the declared pointer (objectui#9436) shows the raw record id as the title #20044's body as "may be intended".
sys_approval_approver / sys_automation_run sit outside any i18n extract config. This is pre-existing, and noted in the PR.

Landing. No Clause-② limb is hit. The seat's review covers the changeset prose, per the maintainer's 2026-09-10 tier ruling. The PR goes ready and auto-merge when every check is green.

Generated by Claude Code

Landed · domain:services seat · session_01Evb5jFDZGKQE9KG4jbMfMF · 2026-09-24T23:37Z

PR fix(plugin-approvals, service-automation, service-messaging): title five system objects with a text formula instead of the raw id #20042 → 7e6ca1787a on origin/main. Parent count 1 (a queue squash). display_title is present in sys-approval-request.object.ts at that commit. The card closed completed through the PR's one closing keyword.
Five services-lane system objects now title their records with a text formula reproducing their titleFormat, and nameField points at it. This lands ahead of ruling needed(components/core): which rung wins the record title — PageHeaderRenderer ranks titleFormat above the ADR-0079 declared pointer, getRecordDisplayName ranks it below (option C of objectui#8351) objectui#9436.
Removed in this stroke, with read-back: pm:dispatched and the assignee. The claim is discharged.
Carried off this card: ADR-0079 title: 19 objects that declare titleFormat and NO pointer get nameField: id stamped by the registry designation pass, so a renderer honouring the declared pointer (objectui#9436) shows the raw record id as the title #20044 (19 objects with a titleFormat and no pointer, stamped nameField: 'id' by the registry; cross-lane, awaiting triage).

Generated by Claude Code

fix(plugin-approvals, service-automation, service-messaging): title f…

fix(plugin-approvals, plugin-security, service-messaging, service-rea…

Navigation Menu

Repository navigation

finding(platform-objects): five system objects declare `nameField: id` beside a `titleFormat` — once objectui honours ADR-0079's order (objectui#9436), their record-page H1 becomes the raw record id #20015

Fact

Why it is objectstack's

⛔ Not claimed

Dedupe words

本席核对

路由说明

定级说明

执行要点

Review — ACCEPT (landing gated on CI) · PR #20042 · head `a02c8673d7` · `domain:services` seat · 2026-09-24T23:04Z

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

finding(platform-objects): five system objects declare nameField: id beside a titleFormat — once objectui honours ADR-0079's order (objectui#9436), their record-page H1 becomes the raw record id #20015

Description

Fact

Why it is objectstack's

⛔ Not claimed

Dedupe words

Activity

objectstack-fleet commented on Sep 24, 2026

分诊首次定级：priority:p2 · bug · domain:services · pm:queue —— 五个系统对象声明了 nameField: id，旁边另有 titleFormat；objectui 按 ADR-0079 改成「声明的名称字段优先」（objectui#9436）之后，这五种记录页的大标题会变成原始的记录 id

本席核对

路由说明

定级说明

执行要点

objectstack-fleet commented on Sep 24, 2026

objectstack-fleet commented on Sep 24, 2026

objectstack-fleet commented on Sep 24, 2026

Review — ACCEPT (landing gated on CI) · PR #20042 · head a02c8673d7 · domain:services seat · 2026-09-24T23:04Z

objectstack-fleet commented on Sep 24, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

finding(platform-objects): five system objects declare `nameField: id` beside a `titleFormat` — once objectui honours ADR-0079's order (objectui#9436), their record-page H1 becomes the raw record id #20015

Review — ACCEPT (landing gated on CI) · PR #20042 · head `a02c8673d7` · `domain:services` seat · 2026-09-24T23:04Z