docs(platform-objects): name the slotted page that declares the org record tab strip

[os-bill]

Part of #16270

• Clause-②: no — this PR puts no new key on any published payload.

The card's premise is false, and that is the deliverable

The card asked for a measurement first: three documents assert that the organization record page opens on a Members tab with Invitations and Teams beside it, while packages/platform-objects/src/identity/ declares no Field.relatedList prominence key. It framed a two-way disjunction — reading 1, the metadata is short three relatedList: 'primary' declarations; reading 2, the three documents are stale — and the triage admitted the card on the grounds that both branches are defects.

Neither branch holds. The tab strip is declared metadata. It is simply not declared by the key the measurement looked for, and not in the directory it looked in.

packages/platform-objects/src/pages/sys-organization.page.ts exports SysOrganizationDetailPage: a type: 'record', kind: 'slotted', isDefault: true page for sys_organization whose slots.tabs override carries exactly three record:related_list tabs — Members, Invitations, Teams, in that order. packages/plugins/plugin-auth/src/auth-plugin.ts hands it to the runtime as pages: [SysOrganizationDetailPage, SysUserDetailPage], and its own comment there already said so: "sys_organization gets a Members / Invitations / Teams tab strip" — a fourth document making the claim, and the only one that named the mechanism.

The renderer half, read from objectui at the sha this repo pins (.objectui-sha = 53ded82bf7):

• packages/plugin-detail/src/synth/buildDefaultPageSchema.ts — when the assigned page supplies a tabs slot, the synthesizer pushes that authored node and never calls buildDefaultTabs. The authored strip replaces the synthesized one outright, so Members is at index 0.
• The same file's buildDefaultTabs always seeds items[0] with Details. So even under reading 1 — three relatedList: 'primary' declarations added — tab 0 would have been Details, never Members. Reading 1 could not have produced the documented surface.
• packages/app-shell/src/utils/deriveRelatedLists.ts sets isPrimary from fieldDef.relatedList === 'primary', exactly as the card quoted. That key is a different mechanism, and while the tabs slot is present it would not move this page at all.

sys-organization.page.ts was already in the tree at 77781151d, the commit the card measured against (verified through the contents API at that ref: HTTP 200, 4606 bytes). The zero was real; the inference from it was not.

What this PR changes — prose only

No metadata. No behaviour. The three documents that assert the tab order now name the page that declares it, so the next reader does not repeat the measurement:

file	change
packages/platform-objects/src/identity/sys-member.object.ts	the invite_user mirror's rationale gains a note naming ../pages/sys-organization.page.ts as the declaring page, and stating that relatedList is a different mechanism absent from this directory
packages/platform-objects/src/identity/invite-entry-toolbar.test.ts	the file header — the stated premise of the whole pin — gains the same pointer
docs/qa/platform-checklist/areas/identity-auth.json	identity-auth.org-membership-team-management gains one source entry, plus the revision 2 to 3 and history bump its ledger requires

What a grader will now do differently (the checklist is what QA grades, so this is stated plainly): nothing about the grade changes. The item's steps, acceptance clauses, their oracles and verify text, and the negative list are byte-unchanged — the same screenshot, the same three tabs, the same FAIL condition for the Organization nav landing on the raw list. What is added is a source entry, which is a pointer a grader reads to locate the surface, and it tells them one new thing: if the page renders Details plus a stacked Related tab instead of the three, that means SysOrganizationDetailPage failed to load — it does not mean the clause was wrong. That reading was not available before, and it is the reading this card was filed for lack of.

Verification

check	result
pnpm lint (repo-wide, eslint . --no-inline-config)	exit 0, 2m05s — full run, no narrowing claimed
pnpm --filter @objectstack/platform-objects test	exit 0 — 39 files, 561 tests passed
pnpm --filter @objectstack/platform-objects typecheck	exit 0
pnpm --filter "@objectstack/platform-objects..." build	exit 0 (dependency closure)
pnpm check:platform-checklist	OK — 15 areas, 264 items; symbol anchors: 633/633 resolved
pnpm check:nul-bytes · check:doc-authoring · check:page-declaration-shape · check:published-files · check:test-source-alias · check:cross-package-test-inputs · check:type-check-coverage · check:comment-mask-adoption · check:comment-mask-corpus · check:keyed-text-bounds · check:closing-keyword-parity	each exit 0
check-empty-changeset · check-changeset-no-major · check-adr-0087-registration, each --base origin/main	each exit 0
check-partof-closing-keyword with this body and this branch's commit list	exit 0

Every exit code above was captured before any pipe (cmd > file 2>&1; EXIT=$?), and the two lock-held runs are quoted from the wrapper's own VERDICT command-exit line.

Ablation — the checklist gate's new anchor really is checked. With the new source entry's symbol renamed on disk to SysOrganizationDetailPageXX (injected text present 1, original present 0, both counted with grep -oF), check-platform-checklist went to exit 1 with ABSENT SYMBOL ... is not in packages/platform-objects/src/pages/sys-organization.page.ts outside its comments. Restored from a pristine copy and proved byte-identical by git hash-object (168dca14 before and after). The green above is therefore a reading, not a gate that judged nothing.

Why a changeset and not the publishes-nothing exemption. @objectstack/platform-objects ships files: ["dist", ...], and its src comments reach dist: after the build, the new comment text appears 4 times under packages/platform-objects/dist, with an exported symbol (SysMember, 16 occurrences) as the positive control and the edited test file's header absent at 0. Published bytes move, so a patch changeset is owed — .changeset/16270-org-record-tab-strip-provenance.md, in the tree's issue-scoped spelling.

Declared narrowing. scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 62 runnable commands for this change set; the table above is a subset chosen by what this diff can move, and the rest is CI's farm. The derivation also warned that this branch's base is 3 commits behind origin/main and that one file it derives from (scripts/check-regen-pending.mjs) changed across that range; re-deriving after git fetch origin main produced a byte-identical command list.

Why Part of and not a closing keyword

This PR carries the measurement and repairs the discoverability gap, but it does not discharge the card. docs/adr/0133-org-management-open-basics.md, under "What this record does not decide", now states something false — that the tab claim is "declared by none of its metadata" and that the ordering is "either an emergent property of the renderer or a claim that has gone stale". docs/adr/** is a governed surface, outside this dispatch's declared file face, and its correction is the maintainer's call. Recorded as a finding in the round report rather than ridden in here.

验收备注

• objectui is read-only and was read, not assumed. All renderer readings are from the sha this repo pins (53ded82bf7), taken with git show 53ded82bf7:path in the sibling checkout. That checkout's own HEAD is 3fbdd4a2d on an unrelated branch; nothing there was edited.
• A second, genuinely different surface exists and is easy to confuse with this one. objectui's console carries a hand-written organization management area at /organizations/:slug (packages/app-shell/src/console/organizations/manage/OrganizationLayout.tsx), whose tabs are Members / Invitations / Settings with an index redirect to members. It is reached from a "Manage" button on the organizations list, not from Setup's nav_organization, and it has no Teams tab. Noted so the next reader does not measure that surface and grade this one. Not filed: no defect, just a collision of names.
• The checklist item has never been run. docs/qa/platform-checklist/runs/ contains only README.md, so identity-auth.org-membership-team-management has no run record and its tab assertion had never been executed against a live console. Noted, not filed.
• The running-app measurement the triage asked for was not performed. It is no longer decisive — the static evidence now names the declaring page, the registration site and the renderer branch that honours it — but it remains the only way to observe the rendered strip. Stated here rather than implied.

---

Generated by Claude Code

[github-actions]

📓 Docs Drift Check

Nothing in this diff resolved to a documentable surface (no symbol, route or SDK anchor derived from 1 changed package(s)), so this run has no opinion about the docs.

What this run could not see

• 1 name(s) were too generic to anchor anything (single lowercase words)
• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.

Coarse fallback — 3 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json e9efc403a9aeae715e0a56f8628e79e5ea25b4eb → packageMentionDocs.