docs(qa): 18.0 sweep anchor round 6 — mark the objectui live specs stale at the console pin, and narrow or drop the pins that cover no full clause, in the UI-facing areas

[objectstack-fleet]

Refs #24579
Clause-②: no

Round 6 of the 18.0 sweep's wave anchor. This PR covers the anchor's §2 themes "objectui live e2e is not green at its own pin" and "pins that cover no full clause", for the items round 4's report (6278788942) names in the four UI-facing area files. Only ledger text changes: 18 items in studio-authoring.json (2) and records-forms.json (16) take one revision bump each and one history row. There is no product code, test, pin or objectui spec change in either repository. dashboards.json and search.json hold no item in either theme, so they are unchanged.

Each history row cites its run records: Tier 1 run #24598, and the area records #24572 / #24616 (records-forms) or #24571 / #24604 (studio-authoring). The rows were committed with the placeholder #PENDING-24579R6, because a history ref names this PR and the PR had no number yet. A follow-up commit on this branch swaps in this PR's number and changes nothing else.

The console pin moved after the sweep

The sweep ran its live specs at objectui 4016370af3c9, the .objectui-sha of the sweep subjects 5b14b7f3 and ba4afcc2. Before this branch was cut, #24660 moved the pin to 12dc7b47c340, so that is the shipped pin at main 92ce3af404 and at 045dae8f23, which this branch merges. I re-read every spec at the new pin:

• git diff 4016370af3c9 12dc7b47c340 -- e2e/live/ changes only e2e/live/ci/backend.env (OBJECTSTACK_VERSION next.3 → next.5). No spec byte changed. Nothing in e2e/live/ changed between 12dc7b47c340 and objectui main 90518cdb84 either.
• So a red spec could have turned green at the new pin only through a change on the surface it drives. For each red spec I checked the cause the round-1 records name against objectui source at 12dc7b47c340 (and at 90518cdb84), and against the showcase metadata at objectstack 92ce3af404. The metadata is unchanged since ba4afcc2. Every cause still holds, and no spec is green at 12dc7b47c340 that was red at 4016370af3c9. So each red spec gets a stale record. Its why cites the run at 4016370af3c9 and the re-reading at 12dc7b47c340.
• Specs that were green at 4016370af3c9 stay cited. Their bytes are identical at 12dc7b47c340, but their verdict there is NOT MEASURED, because no live run exists at the new pin.

spec (objectui e2e/live/)	at 4016370af3c9 (#24598)	at 12dc7b47c340 (re-read)	cause (round-1 record)
studio-object-designer.spec.ts	2 pass / 2 fail (F1, F4)	F1, F4 still stale	F4 expects no banner for a picklist that has no option source, which the spec choice-source rule refuses (packages/spec/src/data/field.zod.ts is unchanged between the objectstack builds the two pins consume, 2c9bf9430a11 and f4755a68a797). F1 opens /apps/com.objectstack.studio, which no stock boot loads (#24571)
studio-record-page.spec.ts	0 / 1	still stale	fills the Object control as an input. On the page create path, the only change between the pins is ResourceEditPage's declaration-only gate for sharing rules (already marked stale from #7695)
user-filters.spec.ts	2 / 4	4 cases still stale	it expects pre-ADR-0053 showcase metadata: in-view tabs, four visualizations, userFilters on tabular. It also expects a cell role where the grid exposes gridcell (#24572)
inline-edit-polish-2572.spec.ts	0 / 1	still stale	it expects an approval lock, but manager_review is deliberately lockRecord: false. The spec also consumes its own fixture (#24572)
field-conditional-rules.spec.ts	0 / 1	still stale	it expects a disabled input. The UI renders read-only text instead. objectui packages/fields is unchanged, and the form renderer's only change is objectui#13055's hide-on-fault group gate (#24572)
grid-file-upload.spec.ts, form-view-subforms.spec.ts	0 / 1 each	still stale	the 'Line Items' locator collides in strict mode with the dialog's screen-reader-only description, "Enter the record and its line items, then save." That text is unchanged at both shas (#24572)
cascading-options.spec.ts	1 / 1 (re-filter case red)	re-filter case still stale	it waits for md-form-submit, which only MasterDetailForm renders, on showcase_cascade. That object has no child object, so its form is a flat ModalForm (#24572)
detail-related-list.spec.ts	0 / 1	still stale	it waits for ?recordId=, but a row click opens /record/ID. The navigation code and packages/spec/src/data/record-surface.ts are unchanged (#24572)
record-history-display.spec.ts	0 / 0 / 1 skipped	still skips	it runs only against app-todo, on the premise that the showcase has no enable.trackHistory object. That premise is stale: showcase_account has one
saved-view-filter, required-when-submit, grid-conditional-rules, grid-parent-rules, master-detail, summary-rollup, list-row-action-cel, action-modal	pass	NOT MEASURED (bytes identical)	—

screen-flow.spec.ts (red at 4016370af3c9) belongs to automation.json, which is not in this claim.

Theme A — objectui live e2e is not green at its own pin

I used the shape studio-authoring.record-page-roundtrip already carries: automated.stale with since / observedIn / ref / why / runnerRule / ownedBy. Where green refs sit beside the stale one, the record also takes stale.partial, which is platform-core.metadata-authoring-roundtrip's precedent. Each runnerRule says which green specs a runner may still cite, and which clauses must be driven by hand.

item	field(s) edited	revision	run record	spec	verdict at pin
studio-authoring.object-designer-roundtrip	automated.stale	5 → 6	#24598; #24571 / #24604	studio-object-designer.spec.ts F1, F4	red at 4016370af3c9, stale at 12dc7b47c340; F2 / F3 green
studio-authoring.record-page-roundtrip	automated.stale.observedIn, automated.stale.why	5 → 6	#24598; #24571 / #24604	studio-record-page.spec.ts	red / stale at both (observedIn #7695 → #24598)
records-forms.list-view-capabilities	automated.stale	4 → 5	#24598; #24572 / #24616	user-filters.spec.ts (4 of 6), inline-edit-polish-2572.spec.ts	red / stale at both; saved-view-filter green
records-forms.conditional-rules-grid	automated.stale	5 → 6	#24598; #24572 / #24616	grid-file-upload.spec.ts	red / stale at both; the other two green
records-forms.cascading-options	automated.stale	3 → 4	#24598; #24572 / #24616	cascading-options.spec.ts re-filter case	red / stale at both; server case green
records-forms.cascading-multilevel-and-clear	automated.stale	3 → 4	#24598; #24572 / #24616	cascading-options.spec.ts re-filter case	red / stale at both; objectql pin green
records-forms.master-detail-atomic-save	automated.stale	3 → 4	#24598; #24572 / #24616	form-view-subforms.spec.ts	red / stale at both; the other two green
records-forms.field-history-tracking	automated.stale	2 → 3	#24598; #24572 / #24616	record-history-display.spec.ts	skipped at both; the plugin-audit pins green

Theme B — pins that cover no full clause

I applied one rule to each item round 4 named. The test is what the pin's own test bodies assert, read at 4016370af3c9. Every one of these pins is byte-identical at 12dc7b47c340.

• Narrow where a pin asserts at least one full clause. In that case automated.ref says which clauses each pin covers, in the round-2 scope-depth-asymmetry shape. A pin that scores none of the item's clauses is dropped from the ref, and the ref names it.
• Drop automated where no pin asserts a full clause, even when green. The item's fixtures.knownGaps line that said "AUTOMATION IS PINNED IN THE objectui REPO" now says that no pin covers a full clause, what each adjacent pin does assert, and when automated may come back.

item	field(s) edited	revision	run record	pin	verdict at pin	choice
records-forms.adhoc-filter-sort-builder	automated.ref, automated.stale	4 → 5	#24598; #24572 / #24616	user-filters.spec.ts; saved-view-filter.spec.ts	user-filters 2 / 4 (stale at both), saved-view-filter green	narrow: user-filters → acceptance[4] only (the uf_FIELD half green, the uf__tab half stale). saved-view-filter pins a saved filter alone, not the merge in acceptance[0], so it is dropped
records-forms.conditional-rules-header	automated.ref, automated.stale, acceptance[4].verify	2 → 3	#24598; #24572 / #24616	list-row-action-cel.spec.ts (+ theme A field-conditional-rules)	green; field-conditional-rules stale at both	narrow: required-when-submit → acceptance[0]. list-row-action-cel gates the task list's Mark Done, not the invoice's Edit / Delete, so it is dropped, and acceptance[4].verify no longer says it pins the clause
records-forms.related-list-server-pagination	automated.ref, automated.stale	5 → 6	#24598; #24572 / #24616	detail-related-list.spec.ts	stale at both	narrow + stale: when green it pins acceptance[2] alone
records-forms.concurrent-edit-conflict	automated.ref, acceptance[4].verify	3 → 4	#24598; #24572 / #24616	occSave.test.tsx; inline-edit-polish-2572.spec.ts	occSave green; inline-edit-polish stale	narrow: occSave → acceptance[3] and acceptance[1]'s modal half. inline-edit-polish saves one field with no If-Match and no conflict, so it scores no clause here even when green. It is dropped, not marked stale
records-forms.import-job-undo-cancel	automated.ref, acceptance[0].verify	5 → 6	#24598 (not run); #24572 / #24616	import-console-undo.spec.ts, import-undo.spec.ts	not run (self-gated, own configs)	narrow: acceptance[1], acceptance[3] and acceptance[0]'s undoable-job half. Neither spec reads the per-row results, and nothing covers acceptance[2] (cancel)
records-forms.action-param-widgets	automated.ref	2 → 3	#24598; #24572 / #24616	action-modal.spec.ts (objectui half)	green	narrow: the objectui half only opens a task's Edit modal and dismisses it, so it is dropped. The dogfood pin → acceptance[2]
records-forms.action-location-matrix	automated removed, fixtures.knownGaps[0]	12 → 13	#24598; #24572 / #24616	list-row-action-cel.spec.ts, action-modal.spec.ts	green, adjacent	drop → Tier 2 (P1)
records-forms.calendar-interactions	automated removed, fixtures.knownGaps[0]	3 → 4	#24598; #24572 / #24616	CalendarView.dnd.test.tsx	green, geometry only	drop → Tier 2 (P1)
records-forms.grid-personalization	automated removed, fixtures.knownGaps[0]	5 → 6	#24598; #24572 / #24616	groupedPagination, groupedBooleanLabel, inlineEditPersistence tests	green, adjacent	drop → Tier 3 (P2)
records-forms.lookup-picker-create-new	automated removed, fixtures.knownGaps[0]	3 → 4	#24598; #24572 / #24616	LookupField.dependsOn, RecordPickerDialog.filterOptions tests	green, adjacent	drop → Tier 3 (P2)

Why this rule, on the four axes:

• Business need. The reader is the next sweep. QA run · tier1:automated-pins (149/166) · ba4afcc2 · 2026-10-19 · 60 PASS / 89 PARTIAL / 0 FAIL / 0 BLOCKED / 17 NOT-RUN #24598 measured these pins and they produced no clause verdict: every clause was not-run. QA run · records-forms (38/44) · ba4afcc2 · 2026-10-19 · 34 PASS / 2 PARTIAL / 2 FAIL / 0 BLOCKED / 6 NOT-RUN #24616 then had to drive the same items by hand.
• Long-term fit. The ledger claims only what a pin asserts. Extending a pin is objectui's test change and is outside this claim. When a pin is extended, re-adding automated is the reverse edit.
• AI-error prevention. A Tier 1 label on a pin that proves nothing is declared-but-not-enforced coverage. Naming the clause per pin makes an over-claim visible. Dropping the label removes the claim instead of tolerating it.
• No scope creep. No new key, gate or pin was added. Both shapes already exist in the ledger.

One consequence differs from the dispatch's hypothesis. A dropped P2 item becomes Tier 3 (deferred unless a change touches it, checklist-test skill §1), not Tier 2. That applies to grid-personalization and lookup-picker-create-new.

Items left alone

• dashboards.json, search.json: no item in either theme. They cite no objectui live spec, and QA run · tier1:automated-pins (149/166) · ba4afcc2 · 2026-10-19 · 60 PASS / 89 PARTIAL / 0 FAIL / 0 BLOCKED / 17 NOT-RUN #24598 names none of their pins as clause-free.
• platform-core.metadata-authoring-roundtrip cites the same stale studio-record-page.spec.ts. It is in round 5's file (landed in docs(qa): 18.0 sweep anchor round 5 — the walled-posture and second-organization themes, runner environment facts, and two carried notes #24687) and still reads observedIn QA run · studio-authoring (FULL area) · 92f26f75 · 2026-08-11 · 3 PASS / 2 PARTIAL / 7 FAIL #7695.
• attachments-storage.inline-grid-receipt-cells (grid-file-upload.spec.ts) and automation.screen-flow-roundtrip (screen-flow.spec.ts) are in the same theme, but they are in round 5's files.
• studio-authoring.draft-publish-lifecycle: QA run · tier1:automated-pins (149/166) · ba4afcc2 · 2026-10-19 · 60 PASS / 89 PARTIAL / 0 FAIL / 0 BLOCKED / 17 NOT-RUN #24598 scores it NOT-RUN because its pins cover no full clause, but round 4's §2 list does not name it, so it is outside this claim.
• Green specs that cover a clause only in part, or on another carrier, stay cited in theme-A-only items, per the ruling that a spec green at the pin stays cited: summary-rollup.spec.ts on master-detail-atomic-save (it rolls up showcase_project, not invoice.total, though acceptance[2].verify says it pins the clause), and grid-conditional-rules.spec.ts on conditional-rules-grid (the inline flag only). These are candidates for the next theme-B pass.
• list-view-capabilities' dogfood half keeps round 4's "scores NO clause" annotation as it landed.
• integration-system.json (open PRs) is excluded. §4 fixture gaps are not this round's work.

Acceptance notes

• RUNNER.md says "23 items carry an automated.ref that lives ENTIRELY in the objectui repo". I measured 28 on main before this PR, and 24 after it. RUNNER.md is outside this claim, so the count is recorded here and not edited.
• grid-parent-rules.spec.ts stays cited as green. The surface it drives, MasterDetailForm's header-record read, was rewritten between the pins (objectui#13062's repair, bc98626e), so its verdict at 12dc7b47c340 is NOT MEASURED.
• No objectui card tracks any of the stale specs. I scanned all 5257 objectui issues (numbers 7–13097, every state) for each spec name and for the causes above, and found none that covers them. The anchor carries the finding for the PM to file.

Verification

Every area-file edit went through a scripted span editor that never re-serialises a file. It splices only the declared value spans, then re-parses and asserts four things: the document equals the old one with the declared ops applied; every other item and every top-level key is value-equal; each edited item's changed leaf paths are a subset of the declared fields plus revision and the one new history row; and the text outside the 18 edited items' spans is byte-identical.

Gates come from node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack, which derived 13 families for the 2 changed paths, against merge base 045dae8f2. All 13 ran at head 70963fb0c2, each exit captured before any pipe, all exit 0:

• node scripts/check-ci-filter-parity.mjs
• node scripts/check-closing-keyword-parity.mjs
• node scripts/check-closing-keyword-parity.mjs --self-test
• node scripts/check-comment-mask-corpus.mjs
• pnpm --filter @objectstack/lint run check:doc-formula-expressions
• pnpm check:cross-package-test-inputs
• pnpm check:doc-authoring
• pnpm check:driver-memory-census
• pnpm check:gitlink-declared
• pnpm check:nul-bytes
• pnpm check:platform-checklist
• pnpm check:refd-timer-probe
• pnpm check:watch-hint-literal

--ran reconciles: "13 derived famil(ies) accounted for — 13 run, 0 NOT-MEASURED (a DERIVED zero — all 13 recorded an exit code and none of them is 3)". check:doc-formula-expressions first refused with exit 3, a prerequisite: @objectstack/formula and @objectstack/lint were not built. I built both under the shared verify lock (VERDICT command-exit 0) and the rerun exited 0.

check:platform-checklist runs in no per-PR CI, so this local reading is its evidence:

check-platform-checklist: OK — 16 areas, 327 items (322 active, 1 planned)

Symbol anchors are 886/896 at 045dae8f23 and 886/896 at head, against 413 cited sources, with 18 file floors held both times. At the original base 92ce3af404 they were 875/885 before and after this diff. So 0 anchors were lost and 0 added.

There is no ablation, because no tested behaviour changed. skip-changeset: docs/qa/** ships in none of the 86 package manifests' files[], and the repo root is private.

---

Generated by Claude Code

[objectstack-fleet]

Seat review

ACCEPT · domain:devx seat 1 · os-warren · session_01FM29ZqLTrnPNXxijioZwMp · 2026-10-20T00:38Z · head 4a516d601d · card #24579 round 6 (claim 6279626364)

• Form.
  • Draft, base main, first lines Refs #24579 / Clause-②: no.
  • Assignee os-warren; labels size/m, skip-changeset.
  • node scripts/pm/check-governed-merges.mjs --pr 24697 → NOT governed, 293 changed lines.
  • git merge-tree against main d1173825b7 is clean.
• Scope. records-forms.json (+211 / −51) and studio-authoring.json (+26 / −5) only. No product code, test, pin or objectui spec. dashboards.json and search.json hold no item in either theme and are unchanged.
• Ledger invariants, by parse against 045dae8f23.
  • 18 items differ: studio-authoring 2, records-forms 16. Each took one revision bump and one history row with ref docs(qa): 18.0 sweep anchor round 6 — mark the objectui live specs stale at the console pin, and narrow or drop the pins that cover no full clause, in the UI-facing areas #24697.
  • No item was added or removed, and no area-level key moved.
• Theme A (stale objectui live specs). 11 items carry automated.stale.
  • 9 use stale.partial, naming the stale spec or case while the green refs stay cited. platform-core.metadata-authoring-roundtrip already carries that shape, and the gate validates only automated.ref, so H1 is falsified and its alternative was not needed.
  • Each why cites Tier 1 run QA run · tier1:automated-pins (149/166) · ba4afcc2 · 2026-10-19 · 60 PASS / 89 PARTIAL / 0 FAIL / 0 BLOCKED / 17 NOT-RUN #24598 at the sweep's pin 4016370af3c9 and a re-read at the shipped pin 12dc7b47c340 (PR chore(console): move the objectui pin 4016370af3c9 → 12dc7b47c340, regime C for sharing rules (objectui#13050) #24660).
  • git diff 4016370af3c9 12dc7b47c340 -- e2e/live/ touches only ci/backend.env. No spec was repaired by the pin move.
• Theme B (pins covering no full clause).
  • 6 refs are narrowed per clause.
  • 4 items lose automated: action-location-matrix and calendar-interactions go to Tier 2; grid-personalization and lookup-picker-create-new go to Tier 3 as P2. Each knownGaps[0] now says what the adjacent pins do assert and when automated may return.
  • The per-item choices are argued on the four axes. Extending a pin is objectui's test change, outside this card.
• Gates.
  • The dev's dispatch-gates union: 13 derived, 13 run, all exit 0 at 4a516d601d.
  • check:platform-checklist holds 886/896 anchors, none lost.
  • Check-runs at 4a516d601d: 39, all completed, 31 success and 8 skipped. node scripts/pm/check-expected-skips.mjs --pr 24697 → OK — 8 skipped check-run(s), every one in the roster (exit 0).
  • No review, no comment.
• Filed from this round: one objectui family card for the stale live specs (the dev scanned all 5257 objectui issues and found none). Its number is in this seat's landing record on qa(checklist): 18.0 release sweep anchor at 5b14b7f3 — re-point 5 red-row clauses + the wave's checklist-accuracy findings, fixture gaps and not-run handover (18 records) #24579.
• Carried on the card.
  • studio-authoring.draft-publish-lifecycle's clause-free pins.
  • Two green specs that over-claim a clause: summary-rollup on master-detail-atomic-save, and grid-conditional-rules on conditional-rules-grid.
  • platform-core.metadata-authoring-roundtrip's stale observedIn #7695.
  • RUNNER.md's "23 items" count (28 before this PR, 24 after).
  • grid-parent-rules.spec.ts is NOT MEASURED at the new pin.

This seat readies the PR and queues it with auto-merge.

---

Generated by Claude Code