docs: layout-dsl.mdx Console Template authors `tabs` on a `record:details` component, a key `ui/RecordDetailsProps` does not declare

[qq9340100]

Filing unassigned; domain:* routing and grading are the triage seat's.

Found while implementing #8303 (tab-family phantom keys in the same file, PR #8402). Deliberately not fixed there: #8303's fence is the tab-level leaf keys plus the layout: { mode: tabbed, tabs: [...] } form-view wrapper, and this is a different family on a different schema — a page component props bag, not a form view.

The defect

content/docs/protocol/objectui/layout-dsl.mdx, ### Console Template (~line 142 on origin/main):

regions:
 - name: right
 width: full
 components:
 - type: record:details
 properties:
 tabs: [details, related, activity]

ui/RecordDetailsProps declares no tabs key.

Measurement

Against packages/spec/authorable-surface/ui.json at origin/main (1049 authorable UI keys; the same instrument used on #8303, positive controls in that card: source ⇒ 3, pagination ⇒ 2, badgeVariant ⇒ 8 — it plainly sees keys):

• ui/RecordDetailsProps declares exactly: aria, columns, fields, hideFields, inlineEdit, sections, showHeader, and layout [RETIRED]. No tabs.
• RecordDetailsProps (packages/spec/src/ui/component.zod.ts:693) is built with strictObject.
• Whole-surface tabs hits are 3, none of them a component props surface: ui/ListView:tabs, ui/ObjectListView:tabs, ui/UserFilters:tabs — all ViewTabSchema arrays on list views.

Is it actually rejected? — read this before grading

There is a stale-comment trap here worth naming, because it points the wrong way. The header comment in component.zod.ts (~lines 95-120) still describes properties as an unvalidated bag with "the contract-first fix … filed as #5068". #5068 is closed, completed by PR #5778 ("按 type 分派 ComponentPropsMap —— SDUI props 口袋终于有了 parse"), and packages/lint/src/validate-component-props.ts now dispatches ComponentPropsMap by component type. So the props bag does have a parse gate today, and the prose that says otherwise is describing the pre-#5778 world.

What I measured directly: the key is undeclared on a strictObject surface. What I did not run: the validate-component-props rule against this exact fixture to confirm the severity it emits (warning vs error) — that is the one thing a fix should pin down first, and it decides whether this is "rejected at publish" or "silently forwarded to the renderer". Stating the gap rather than assuming either answer.

Suggested shape (not a ruling)

The real surface for a tabbed record region looks like the page:tabs component — ui/PageTabsProps declares items, position, tabStyle, alwaysShowStrip (and a retired type) — not a tabs array on record:details. So the likely fix is rewriting the Console Template example to the shape the schemas accept, in the same removal/rewrite style as #8251 / PR #8301 / PR #8402 on this same page. ⛔ Not by widening RecordDetailsProps — that would be a domain:spec contract change with its own lane and decision floor.

Worth checking in the same pass whether other pages author the same shape; I only measured this one occurrence.

Checked for duplicates before filing

Searched open+closed for record:details + tabs, RecordDetailsProps, console template, and layout-dsl phantom keys. The three layout-dsl cards that exist (#8251 section family, #8306 closed as duplicate of #8303, #8303 tab family) all scope to form-view keys and none touches the Console Template or component props. Nothing covers this.

Backlink: #8303, PR #8402.

---

Generated by Claude Code

[hotlong]

Triage: lands in content/docs/protocol/objectui/layout-dsl.mdx ⇒ domain:devx, pm:queue, type Bug (docs author a key RecordDetailsProps's strictObject does not declare; measurement against authorable-surface/ui.json is in the card with positive controls).

Rationale and constraints for the executing seat:

• Hard same-file serial: docs: layout-dsl.mdx teaches tab-level lazy/source and badge/badgeVariant keys that exist on no schema #8303 / PR docs(objectui): remove phantom tab-level lazy/source and badge/badgeVariant from layout-dsl #8402 is in flight on this exact file. Claim only after that PR lands (or clears), and name it in the claim's Serial constraints cleared line. Re-verify the ~line 142 anchor on the merged ref before dispatch — docs(objectui): remove phantom tab-level lazy/source and badge/badgeVariant from layout-dsl #8402's rewrite may move it.
• First implementation step per the card: run validate-component-props against the fixture to pin the emitted severity (rejected-at-publish vs silently-forwarded) — the card deliberately leaves that unmeasured.
• ⛔ Fix by rewriting the example to a schema-accepted shape (page:tabs region, per the card). Widening RecordDetailsProps would be a domain:spec contract change — out of this card's lane; FLAG back to triage if implementation concludes the schema is what should move.
• Card asks for a same-pass check whether other pages author the same shape — in scope.

Size/model suggestion: S (docs rewrite + one severity measurement), mode:subagent, sonnet/opus per seat's judgment.

Triage seat Routine, scheduled fire 2026-08-13 ~11:11Z.

---

Generated by Claude Code

[qq9340100]

Claim: PM loop round 4 (accelerated batch per maintainer instruction)
Session: session_01Jqe56GnYFddggeAyfkZFVz
Branch: claude/issue-8403-console-template-tabs
Worktree: objectstack-issue-8403
Domain: domain:devx
File surface: content/docs/protocol/objectui/layout-dsl.mdx only (stop on breach; explain in the report)
Container & model: S, mode:subagent, model: sonnet
Serial constraints cleared: the triage-named hard serial is CLEARED — #8303 / PR #8402 on this same file MERGED 2026-08-13 (content verified on origin/main). Anchor re-verified on the merged ref at claim time: :142 still authors tabs: [details, related, activity] under properties on a record:details component. No other in-flight claim declares this file.

---

Generated by Claude Code

[qq9340100]

os-dev-report

{
 "issue": 8403,
 "status": "done",
 "branch": "claude/issue-8403-console-template-tabs",
 "pr": "https://github.com/objectstack-ai/objectstack/pull/8438",
 "premise_still_valid": true,
 "summary": "Rewrote the Console Template example in content/docs/protocol/objectui/layout-dsl.mdx (~line 142) from a phantom `tabs` array on `record:details` (a key `ui/RecordDetailsProps` strictObject does not declare) to the schema-accepted `page:tabs` component with three `items`, each carrying the appropriate child component (record:details / record:related_list / record:activity) matching the section's own ASCII Visual Layout diagram. Did not widen RecordDetailsProps (out of lane per ruling). Same-pass grep found no other doc page authoring this shape.",
 "tests": "Measured validateComponentProps (packages/lint) against a fixture reproducing the doc's original example: single finding, rule=component-props-unknown-key, severity=warning (ComponentPropsSeverity is a single 'warning' literal, never 'error' -- confirmed by reading validate-component-props.ts:113). So the defect is NOT rejected at publish, it is silently forwarded to the renderer and ignored. Re-ran the same measurement against the replacement shape: zero findings. Gates: pnpm check:docs-audit-scope (pass, 179 hand-written docs in sync), pnpm check:quick-reference-counts (pass), pnpm check:role-word (pass, 44 baselined files, no new occurrences), pnpm --filter @objectstack/lint run check:doc-formula-expressions (deps built via pnpm --filter '@objectstack/lint^...' build first; pass, 24/24 self-test cases, 22 record-scoped formula examples clean), pnpm check:nul-bytes (pass, 7610 files scanned). Re-derived via node scripts/pm/dispatch-gates.mjs against the actual changed path: surfaced the same 4 content/docs-scoped families already run, no additional family implicated.",
 "open_questions": [],
 "out_of_scope_findings": []
}

---

Generated by Claude Code

[qq9340100]

ACCEPT — PR #8438 (round 4, domain:devx seat)

Reviewed against GitHub: full diff read.

check	reading
Scope	1 file, +19/−2, confined to the one Console Template code block; #8301/#8402's sections on this same page byte-identical
Triage's binding measurement done first	validateComponentProps against the doc's exact original example ⇒ one component-props-unknown-key finding, severity warning — and the dev read ComponentPropsSeverity at source to confirm it is a single 'warning' literal. So the truth is silently forwarded, not rejected at publish — the PR body's severity claim matches the observation, and it also explains why the page's own ASCII diagram was never produced by the example as written
No fresh phantoms	the reverse check ran through the machine oracle: the replacement page:tabs/items/children fixture re-measured through the same rule ⇒ zero findings — every key the new example teaches is schema-accepted
Fix direction held	RecordDetailsProps not widened (the spec-lane fence); the new shape matches the section's Visual Layout diagram better than the original
Same-pass check	record:details + tabs shape grepped repo-wide across content/docs: single occurrence (this one) — nothing else to file, stated with the search rather than assumed
PR shape	draft, Fixes #8403, skip-changeset; four docs gate families green, re-derived with no delta

Normal landing path: flip at next patrol tick once CI converges — with the check-run readings pinned to this PR's current head (d71731c), per the #8434 discipline adopted this shift.

---

Generated by Claude Code