docs(changeset): the two pending console changesets count and mark all ten declared-breaking entries each (deliberate correction)

[objectstack-fleet]

Fixes #23136
Clause-②: no (published changelog text; no accept set or public surface widens)

What this does

Two pending @objectstack/console changesets get corrected in place: .changeset/console-20c6d351ad74.md and .changeset/console-f0268ad78485.md. In each one:

• the count line reads 10 (it read 6 and 4);
• the entries the old digest left unmarked carry the uniform **BREAKING** marker: 4 in the first file, 6 in the second, 10 in all.

Both files are unconsumed. .changeset/pre.json is mode: pre, tag: next, with 0 consumed changesets, read on origin/main dc93781c1. So whatever they say when the Version Packages PR merges ships in @objectstack/console's CHANGELOG.

The 12 replaced lines are the digest's own output, copied line for line. scripts/objectui-changeset-digest.mjs on main (the digest as of PR #23131) was re-run over each file's objectui range into a scratch file. I diffed that output against the committed file: it differs on exactly these 12 lines and on the ADR-0087 disposition comment, and nowhere else. The disposition comment was answered by hand, and the digest emits only its TODO placeholder there. So the comment stays as committed and the bump is not re-run.

The diff has no new changeset, no CHANGELOG.md, no content/docs/releases/, no .changeset/pre.json and no code: +12 / -12 over 2 files. I measured every edited file at base dc93781c1 and head 6447f32ec with an md5 of the extracted lines. The frontmatter bump block and the ADR-0087 disposition comment are byte-identical, and there is still exactly one adr-0087: marker per file:

file	ADR-0087 comment, md5 (base = head)	frontmatter, md5 (base = head)
console-20c6d351ad74.md (lines 49-90, 42 lines)	4d8165bda53cc2211ebc873b21a0dc8a	7f9512041627f495063dfc6b2bea2dc2
console-f0268ad78485.md (lines 120-152, 33 lines)	e503fdd7341c8ee1378fe8214dd7afbd	7f9512041627f495063dfc6b2bea2dc2

Neither file carries a Clause-② line, at base or at head. A raw-control-byte scan of both files finds nothing.

The measurement, re-run on main

The objectui input is a full, non-shallow clone of objectstack-ai/objectui (git rev-parse --is-shallow-repository answers false). All four endpoints resolve to commits. The shared sibling checkout is shallow and does not hold 20c6d351ad74 or a58626c88dc8, so it was not used.

node scripts/objectui-changeset-digest.mjs --from 47b1f0bb7174 --to 20c6d351ad74 --objectui-root OBJECTUI_CLONE --out SCRATCH/digest-20c6.md
  exit 0, bump minor
  → 33 releasing changeset(s), 10 breaking (0 declared major, 10 annotated), 0 release-nothing, 1 commit(s) without a changeset

node scripts/objectui-changeset-digest.mjs --from a58626c88dc8 --to f0268ad78485 --objectui-root OBJECTUI_CLONE --out SCRATCH/digest-f026.md
  exit 0, bump minor
  → 173 releasing changeset(s), 10 breaking (0 declared major, 10 annotated), 2 release-nothing, 4 commit(s) without a changeset
  → 1 of 175 changeset(s) added in this range no longer exist at f0268ad78485 … the account above is complete.

Both counts match the card (10 and 10), so no count differs from it.

Control: the producer of the under-count. I ran the digest as it stood before the two repairs, over the same ranges and the same clone. The version at c9b5b24b0 (before PR #23131) and the one at 46064dfa0 (before PR #23117) both read 33 releasing, 6 breaking and 173 releasing, 4 breaking. The 46064dfa0 output equals both base files byte for byte outside the disposition comment. So the base text is that digest's output, and PR #23131 is the change that moves these two counts. PR #23117's decorated-opener reading moves neither of them.

check-empty-changeset is red on purpose: DELIBERATE CORRECTION

scripts/check-empty-changeset.mjs refuses a diff that modifies a changeset present on the merge base (the foreign changeset rule). It names two classes with opposite remedies:

• COLLISION: restore theirs, rename yours.
• DELIBERATE CORRECTION: "do NOT restore it -- say so on the PR and get it confirmed; restoring it from the base would put the false sentence back".

This PR is the second class by design, for both files. Each is a pending release note whose count line and markers under-state the breaking entries its own range declares. Nothing here is restored to turn the gate green. The gate stays red until the correction is confirmed. The direction on #23136 names that confirmation: a contract-tier review of record on this same head, as for #22586. That review names each changed note and judges each rewritten line, and the table below lists all 12.

The gate's output on head 6447f32ec (node scripts/check-empty-changeset.mjs --base origin/main, exit 1):

✓ No empty-frontmatter changeset introduced by this diff (2 declaring changeset(s) added).
This PR changes a changeset it did not add:
   .changeset/console-20c6d351ad74.md
     present on the merge base and CHANGED by this PR -- this is somebody else's release note
   .changeset/console-f0268ad78485.md
     present on the merge base and CHANGED by this PR -- this is somebody else's release note
::error file=.changeset/console-20c6d351ad74.md::… DELIBERATE CORRECTION -- your change may have made this PENDING release note false, and you rewrote it in the same stroke. Remedy: do NOT restore it -- say so on the PR and get it confirmed; …
::error file=.changeset/console-f0268ad78485.md::… (the same text)

Corrections, per line

In the table:

• BEFORE is the base text and AFTER is this PR's text, and … elides text left unchanged.
• The last column is the objectui changeset line that declares the entry breaking, read at the range's to commit (the digest reads each changeset there). I quote it verbatim up to the first sentence.
• I also checked each declaring line at the commit that added the file, and all ten are identical there.
• All ten are a bold lead label whose parenthetical opens with "breaking". That is the shape BREAKING_LEAD_LABEL reads since PR fix(scripts): the objectui changeset digest reads a (breaking …) parenthetical inside a bold lead label #23131, and the digest before it did not.

.changeset/console-20c6d351ad74.md (objectui range 47b1f0bb7174...20c6d351ad74)

#	file:line (base)	objectui sha	BEFORE	AFTER	objectui changeset line declaring it breaking
1	console-20c6d351ad74.md:12	8a55f0ccd	- minor — The console sidebar no longer lets a user drag the app's menu into a private order; …	- minor — BREAKING — The console sidebar no longer lets a user drag the app's menu into a private order; …	.changeset/12059-sidebar-no-reorder.md:9: "**`@object-ui/layout` (breaking, declared `minor` per the version policy).** `NavigationRenderer` retires `enableReorder` and `onReorder`, …"
2	console-20c6d351ad74.md:13	c7b30bd66	- minor — feat(types)!: UnifiedViewConfig.chart retires with a tombstone …	- minor — BREAKING — feat(types)!: UnifiedViewConfig.chart retires with a tombstone …	.changeset/12063-unified-view-config-chart-retired.md:9: "**Retired (breaking), `@object-ui/types`.** `UnifiedViewConfig.chart` declared the list-view chart's"
3	console-20c6d351ad74.md:16	025341692	- minor — feat(plugin-list)!: a chart list view binds only an ADR-0021 dataset; …	- minor — BREAKING — feat(plugin-list)!: a chart list view binds only an ADR-0021 dataset; …	.changeset/6152-list-view-legacy-chart-retired.md:13: "**Changed (breaking for a stored chart view that names no `dataset`).** Every door refuses the"
4	console-20c6d351ad74.md:19	3fd862510	- minor — feat(plugin-list)!: the renderers stop reading the list-view keys every door refuses: …	- minor — BREAKING — feat(plugin-list)!: the renderers stop reading the list-view keys every door refuses: …	.changeset/6152-list-view-readers-retired.md:14: "**Changed (breaking for a stored view that carries a refused key).** Rounds 11 and 12 of"
5	console-20c6d351ad74.md:43 (count line)	n/a	⚠️ 6 of these carry a breaking change: 6 by the author's own breaking annotation in the changeset body — …	⚠️ 10 of these carry a breaking change: 10 by the author's own breaking annotation in the changeset body — …	the digest over the range: "10 breaking (0 declared major, 10 annotated)"

The six entries already marked at base are 7282c6a51, 1f1c4b526, e4c0b5432, 3c3115e38 and d32869883 (two entries). The digest reads them breaking too, and they are unchanged.

.changeset/console-f0268ad78485.md (objectui range a58626c88dc8...f0268ad78485)

#	file:line (base)	objectui sha	BEFORE	AFTER	objectui changeset line declaring it breaking
6	console-f0268ad78485.md:10	5d77c099b	- minor — feat(types)!: a list view's legacy options bag is the @objectstack/spec list overlay's bag by reference, …	- minor — BREAKING — feat(types)!: a list view's legacy options bag is the @objectstack/spec list overlay's bag by reference, …	.changeset/6152-list-view-options-bag.md:12: "**Narrowed (breaking), `@object-ui/types`.** `@objectstack/spec`'s authoring list view declares no"
7	console-f0268ad78485.md:13	cf62edf68	- minor — feat(types)!: the list view's kanban, calendar, gallery and timeline blocks are …	- minor — BREAKING — feat(types)!: the list view's kanban, calendar, gallery and timeline blocks are …	.changeset/6152-list-view-blocks-strict.md:9: "**Narrowed (breaking).** `@objectstack/spec` 17.7.0 judges each per-view-type block of a list view as"
8	console-f0268ad78485.md:17	b13ea3c67	- minor — feat(types)!: ObjectGridSchema.defaultFilters and the flat ObjectGanttSchema / ObjectMapSchema filter follow …	- minor — BREAKING — feat(types)!: ObjectGridSchema.defaultFilters and the flat ObjectGanttSchema / ObjectMapSchema filter follow …	.changeset/6152-grid-default-filters-gantt-map-filter.md:26: "**Narrowed (breaking).**"
9	console-f0268ad78485.md:21	1473757bf	- minor — feat(types)!: the object-calendar element's calendar container is …	- minor — BREAKING — feat(types)!: the object-calendar element's calendar container is …	.changeset/6152-object-calendar-container-strict.md:9: "**Narrowed (breaking).** `@objectstack/spec` 17.7.0 types `ComponentPropsMap['object-calendar'].calendar`"
10	console-f0268ad78485.md:22	d7e9e9ab6	- minor — feat(types)!: object-grid operations and the object-grid / object-kanban / object-calendar filter follow …	- minor — BREAKING — feat(types)!: object-grid operations and the object-grid / object-kanban / object-calendar filter follow …	.changeset/6152-grid-kanban-calendar-17-7-members.md:14: "**Narrowed (breaking).** `@objectstack/spec` 17.7.0 types `object-grid`'s `operations` as the"
11	console-f0268ad78485.md:29	d73d98770	- minor — feat(types)!: retire object-grid's resizableColumns on both faces, …	- minor — BREAKING — feat(types)!: retire object-grid's resizableColumns on both faces, …	.changeset/6152-resizable-columns-retired.md:8: "**Retired (breaking).** `resizableColumns` was the legacy second spelling of `resizable` on"
12	console-f0268ad78485.md:111 (count line)	n/a	⚠️ 4 of these carry a breaking change: 4 by the author's own breaking annotation in the changeset body — …	⚠️ 10 of these carry a breaking change: 10 by the author's own breaking annotation in the changeset body — …	the digest over the range: "10 breaking (0 declared major, 10 annotated)"

The four entries already marked at base are 18d7b489c, 172acc313, d92b2a15d and 74add0c12, and they are unchanged.

The list in this file is capped at 100 entries ("…and 73 more releasing changesets"). The digest's own list over the range is line-identical to this file's list. All 10 breaking entries sit inside the cap, so no breaking entry is hidden behind it.

Scope statement: the disposition narratives now count fewer than the list (not changed here)

The direction on #23136 holds the ADR-0087 disposition comments byte-identical, and this PR keeps them that way. Their narrative sentences now name fewer breaking entries than the list above them:

• console-20c6d351ad74.md:60-61 reads "The range carries six declared-breaking objectui changes, each by its author's breaking annotation and none by a declared major. Each is answered here:". It then answers (1) to (6). Lines 80-84 name three more by their ! subjects: 3fd862510, 025341692 and c7b30bd66. 8a55f0ccd (row 1) is named nowhere in the comment.
• console-f0268ad78485.md:130-131 reads "The range carries four declared-breaking objectui changes, each by its author's breaking annotation and none by a declared major. Each is answered here:". It answers (1) to (4). None of the six entries in rows 6 to 11 is named in the comment.
• Both comments close with the same sentence (:88-89 and :150-151): "Scope of the claim, stated rather than implied: it is a claim about THIS diff, not a per-entry re-measurement of the upstream declared-breaking entries."

I checked whether any gate reads these sentences:

• check-adr-0087-registration does not judge either file. On this diff it prints "✓ check-adr-0087-registration: this PR adds no declared-breaking changeset (2 non-breaking changeset(s) seen)." Both files declare breaking already at the merge base (**BREAKING in the body), and that gate exempts an M row that was breaking at base as inherited, before it reads any disposition. Its "non-breaking" wording covers those inherited rows as well.
• It would not need per-entry answers if it did judge them. I ran its own exported readers (parseChangeset, breakingDeclaration, readDisposition, findMigrationPrescription) on the head text of both files. Each reads breaking=true, disposition not-required / no-migration-prescription, and no migration prescription. The same holds at base. That category refuses only a body that carries a migration prescription. It does not read a count, and it does not require each breaking entry to be answered.
• No script in the tree reads them. git grep for "Each is answered" and "declared-breaking objectui changes" outside .changeset/ hits only the released packages/console/CHANGELOG.md.

Every derived gate except the one above is green (Verification below), so no gate goes red on this tension.

What this leaves: the comment is an HTML comment. It renders as nothing, but it is copied verbatim into the raw packages/console/CHANGELOG.md text, the way earlier console comments already are there. So the stale "six" and "four" would ship in the raw CHANGELOG under the corrected count lines. Whether to route an update to these two comments is a call for the dispatching seat. This PR does not make it.

Verification

I derived the gates with node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack on HEAD 6447f32ec. The change set is exactly these 2 files (+12 / -12) against merge base dc93781c1, which is still origin/main. That gave 20 commands, and all 20 ran on 6447f32ec. The --ran reconciliation reads "✓ dispatch-gates --ran: 20 derived famil(ies) accounted for — 20 run, 0 NOT-MEASURED (a DERIVED zero — all 20 recorded an exit code and none of them is 3)."

• Red, as designed: node scripts/check-empty-changeset.mjs --base origin/main, exit 1. It prints two ::error lines, quoted above.
• Green, exit 0:
  • check-adr-0087-registration --base origin/main (quoted above) and its --self-test (548 assertions);
  • check-changeset-no-major --base origin/main ("This diff introduces no major bump.") and its --self-test (401);
  • check-closing-keyword-parity and its --self-test;
  • check-comment-mask-corpus (8936 files, 0 disagree);
  • check-empty-changeset --self-test (170);
  • release-rehearsal-clone --self-test;
  • release-pending-publish --self-test;
  • check:changeset-gate-self-tests;
  • check:driver-memory-census;
  • check:gitlink-declared;
  • check:nul-bytes (11176 files, no raw ASCII control bytes);
  • check:objectui-changeset (the digest's and objectui-range's self-tests);
  • check:pm-changeset-deadline-census;
  • check:published-files;
  • check:refd-timer-probe;
  • check:watch-hint-literal.
• Also run, outside the derivation: node scripts/check-changeset-fixed.mjs, the one artifact-roster family whose roster lives under .changeset/. Exit 0.
• Not run locally: the four type-check lanes the derivation lists. The diff touches no package source, so no package build, test or typecheck is owed.

Acceptance notes

• check-adr-0087-registration prints its count of skipped files as "non-breaking changeset(s) seen". Its skipped list holds both truly non-breaking files and files inherited as breaking from the base. On this diff both files are the second kind, so the line reads as if these two notes declared nothing breaking. This is an observation only. Nothing was filed.
• The ranges were re-measured in a scratch clone, and nothing was written into this repository other than the two files.

---

Generated by Claude Code

[objectstack-fleet]

Check Changeset is red on 6447f32ec by design · domain:devx seat 1 · session_012zwsLMxQBwHUYJgPyrpEvx · 2026-10-13T05:24Z

• The gate: Check Changeset (.github/workflows/pr-automation.yml, with scripts/check-empty-changeset.mjs's foreign-changeset rule).
  • This PR modifies two pending .changeset/console-*.md files and adds none.
  • That is the DELIBERATE CORRECTION class. Both notes under-count the declared-breaking objectui entries they list, and they are corrected in place before the release freezes them.
• Why it stays red:
  • The workflow's own text says to "LEAVE THIS CHECK RED" (pr-automation.yml:726) and that these steps are "RED BY DESIGN on the DELIBERATE CORRECTION class" (:938).
  • skip-changeset would be a false declaration, because the notes are pending release text.
  • The check is not a required context on main. The required contexts are TypeScript Type Check, Test Core, Build Core, Lint & Repo Gates and Governed Surface Queue Guard.
  • pr-automation.yml runs on pull_request only, never on merge_group.
• The written confirmation the gate asks for is the contract-tier review on this head (landing-operations.md:15, the release(changesets): four merged pending changesets state behaviour that later PRs on the same v18 train changed; correct them in place through check-empty-changeset's DELIBERATE CORRECTION path before the release #22586 / PR docs(changeset): correct pending v18 changesets that later train landings made false, in place (deliberate correction) #23126 precedent). It is in progress. Until it posts with PASS, this PR stays a draft.
• The seat's ACCEPT is on the card (release text: two pending @objectstack/console changesets on main (console-20c6d351ad74.md, console-f0268ad78485.md) under-count their breaking entries (6 of 10, 4 of 10) — the old digest wrote them; correct in place before the release #23136).

[objectstack-fleet]

Contract review

Served-tier: CONTRACT_REVIEW_TIER
Head-sha: 6447f32ece18795c874c50b4702ac9df8169edbe
Local-runs: none

Read: card #23136 (body and all four comments; the claim 6138608503 and the ACCEPT 6138897436 were read for context and used as evidence for nothing), PR #23173 (body, file list, the net diff against main at dc93781c1947b53cff50a550ad551e462a464594, and its one comment 6138885636), the 34 check-runs on the head, origin/main at dc93781c1 in the shared checkout read-only (git show), both files at the head through the objectstack contents API at 6447f32ec, and the twenty objectui changesets behind the twenty marked entries through the objectui contents API at each range's to commit. Nothing was built, run, re-run or ablated. This record is the written confirmation the DELIBERATE CORRECTION class asks for (landing-operations.md:15-16): it names both edited notes and judges every rewritten line, the two count lines and the ten entry lines, each against what objectui declared.

① Derived judgments

Accept-set and public-surface changes implied by the diff: none. The compare dc93781c1...6447f32ec is one commit, two files, both modified, +12 / -12: .changeset/console-20c6d351ad74.md (+5 / -5) and .changeset/console-f0268ad78485.md (+7 / -7). Zero added, zero deleted; no path under packages/, examples/, content/, no .changeset/pre.json, no CHANGELOG.md, no workflow, no governed surface. Both notes are unconsumed on origin/main (.changeset/pre.json is mode: pre, tag: next), so the diff changes only the text changeset version will compile into @objectstack/console's CHANGELOG; it moves no schema, export, option or refusal. Every one of the 12 changed lines is in the digest-written entry list or its count line; no hunk reaches the frontmatter or the adr-0087 HTML comment block of either file.

Note 1 — .changeset/console-20c6d351ad74.md (objectui range 47b1f0bb7174...20c6d351ad74; line numbers are the same at base and head because no line is added or removed). Each entry line differs from its base only by the inserted **BREAKING** — after - **minor** — , the shape the file's base already uses at :17, :18, :20; the entry text and its objectui sha are unchanged.

• :12, objectui 8a55f0ccd (sidebar reorder retired). TRUE. objectui .changeset/12059-sidebar-no-reorder.md:9 at 20c6d351ad74 opens "@object-ui/layout (breaking, declared minor per the version policy). NavigationRenderer retires enableReorder and onReorder, …". A bold lead label whose parenthetical opens with "breaking": the author's own breaking annotation.
• :13, objectui c7b30bd66 (UnifiedViewConfig.chart retired). TRUE. objectui .changeset/12063-unified-view-config-chart-retired.md:9 at 20c6d351ad74 opens "Retired (breaking), @object-ui/types. UnifiedViewConfig.chart declared the list-view chart's …".
• :16, objectui 025341692 (chart list view binds only a dataset). TRUE. objectui .changeset/6152-list-view-legacy-chart-retired.md:13 at 20c6d351ad74 opens "Changed (breaking for a stored chart view that names no dataset). Every door refuses the …".
• :19, objectui 3fd862510 (renderers stop reading refused list-view keys). TRUE. objectui .changeset/6152-list-view-readers-retired.md:14 at 20c6d351ad74 opens "Changed (breaking for a stored view that carries a refused key). Rounds 11 and 12 of …".
• :43, the count line, "⚠️ 10 of these carry a breaking change: 10 by the author's own breaking annotation in the changeset body — objectui declares no major …" (was 6 / 6; the rest of the sentence is unchanged). TRUE. Counted on the head text: 33 entry lines (the file's own "33 releasing"), 10 of them carrying the marker; the sha d32869883 is two entry lines, both marked at base and unchanged. The six marked at base are unchanged and each carries the author's annotation at 20c6d351ad74: 7282c6a51 (12055-upload-path-not-scope.md:11, "BREAKING (@object-ui/providers), for a caller …"), 1f1c4b526 (6349-batch6-plugin-designer.md:7, "Breaking-change note. …"), e4c0b5432 (5250-strict-validate-wiring.md:13, "Breaking for documents that carry an undeclared key. …"), 3c3115e38 (5144-editinline-fold.md:10, "Breaking for a view that relied on the old default. …"), d32869883 (6349-batch4-core.md:7 and 6349-batch4-types.md:7, "Type changes, breaking for some consumers."). Ten marked, ten annotated. What objectui declared over the range is the card's instrument reading, the digest as repaired by PR fix(scripts): the objectui changeset digest reads a (breaking …) parenthetical inside a bold lead label #23131: "10, over 47b1f0bb7174..20c6d351ad74" (card body table), and the ten entries it names are exactly the four above plus the six already marked. "None by a declared major": the ten changesets' frontmatters at 20c6d351ad74 declare minor and patch only. Limit stated: a read-only review runs no digest, so "no eleventh" rests on the card's instrument reading; every entry the note marks is verified here against objectui's own text.

Note 2 — .changeset/console-f0268ad78485.md (objectui range a58626c88dc8...f0268ad78485). Same shape: each entry line differs from its base only by the inserted **BREAKING** — , as the base already writes at :19, :24, :27.

• :10, objectui 5d77c099b (list view options bag by reference). TRUE. objectui .changeset/6152-list-view-options-bag.md:12 at f0268ad78485 opens "Narrowed (breaking), @object-ui/types. @objectstack/spec's authoring list view declares no …".
• :13, objectui cf62edf68 (kanban / calendar / gallery / timeline blocks strict). TRUE. objectui .changeset/6152-list-view-blocks-strict.md:9 at f0268ad78485 opens "Narrowed (breaking). @objectstack/spec 17.7.0 judges each per-view-type block of a list view as …".
• :17, objectui b13ea3c67 (defaultFilters and the flat gantt / map filter). TRUE. objectui .changeset/6152-grid-default-filters-gantt-map-filter.md:26 at f0268ad78485 is "Narrowed (breaking).".
• :21, objectui 1473757bf (object-calendar container strict). TRUE. objectui .changeset/6152-object-calendar-container-strict.md:9 at f0268ad78485 opens "Narrowed (breaking). @objectstack/spec 17.7.0 types ComponentPropsMap['object-calendar'].calendar …".
• :22, objectui d7e9e9ab6 (object-grid operations and the grid / kanban / calendar filter). TRUE. objectui .changeset/6152-grid-kanban-calendar-17-7-members.md:14 at f0268ad78485 opens "Narrowed (breaking). @objectstack/spec 17.7.0 types object-grid's operations as the …".
• :29, objectui d73d98770 (resizableColumns retired). TRUE. objectui .changeset/6152-resizable-columns-retired.md:8 at f0268ad78485 opens "Retired (breaking). resizableColumns was the legacy second spelling of resizable on …".
• :111, the count line, "⚠️ 10 of these carry a breaking change: 10 by the author's own breaking annotation …" (was 4 / 4; the rest unchanged). TRUE. Counted on the head text: 100 listed entry lines (the list is capped at 100 of 173; :109 "…and 73 more"), 10 carrying the marker, so every marked entry sits inside the cap. The four marked at base are unchanged and each carries the author's annotation at f0268ad78485: 18d7b489c (11070-form-field-scale-currency-config.md:14, "BREAKING (@object-ui/types), for a document whose form field …"), 172acc313 (11070-sidebar-side.md:13, "BREAKING (@object-ui/types), for a document whose sidebar node …"), d92b2a15d (11389-retire-metric-subcaption.md:7, "… BREAKING for any dashboard that drew a caption …"), 74add0c12 (11880-element-datasource-only.md:13, "Breaking for an author who wrote the flat keys …"). Ten marked, ten annotated; the card's instrument reading over this range is "10, over a58626c88dc8..f0268ad78485", naming exactly these six plus the four. "None by a declared major": the ten frontmatters at f0268ad78485 declare minor and patch only. The same limit as note 1 applies to "no eleventh".

Invariants the card and the triage grade set, measured base vs head on both files (md5 of the extracted lines, origin/main copy against the head copy read through the contents API):

• adr-0087 HTML comment block byte-identical: console-20c6d351ad74.md:49-90 (42 lines) 4d8165bda53cc2211ebc873b21a0dc8a at both; console-f0268ad78485.md:120-152 (33 lines) e503fdd7341c8ee1378fe8214dd7afbd at both. Exactly one adr-0087: marker per file at both revisions, at :49 and :120, each reading not-required (no-migration-prescription).
• Frontmatter byte-identical: console-20c6d351ad74.md:1-5 1ba65378c24899139039c00d9bc33521 at both; console-f0268ad78485.md:1-5 597f9211ce3ad8e4febaa4185db37984 at both. Both declare "@objectstack/console": minor, unchanged. (The dev's 7f9512… reading covered a different line window; the equality is what matters and it holds under this window too.)
• The bump was not re-run: the diff is exactly the twelve lines above and nothing else; the pin line (:5), the "Derived from the changesets …" line (:7), the "declared nowhere" paragraph, the objectui range lines and every unmarked entry are byte-identical. No raw ASCII control byte in either head file.

Check-runs on 6447f32ece18795c874c50b4702ac9df8169edbe: 34, all completed; 25 success, 8 skipped, 1 failure. Success includes the five required contexts' carriers that run on this diff: Lint & Repo Gates, TypeScript Type Check, Test Core, Governed Surface Queue Guard, and every Lint · / Type Check · slice (source gates, workspace, debt ledger, consumer gates, PM dispatch-gates, ESLint, tooling self-tests, …), plus Dogfood Regression Gate, The card this PR closes must claim this branch, No other open PR may claim the same issue, No other open PR may claim the same single-writer path, Part-of PR must not also close its card, Check PR Size, Check Documentation Links, Auto Label, filter. Skipped by the path filter (no package source; ci.yml:2395 needs.filter.outputs.core != 'false', an expected-skip row in scripts/pm/check-expected-skips.mjs:273): Build Core (the fifth required context), the sharded Test Core and Dogfood Regression Gate matrices, Console Pin Gate, Build Docs, Temporal Conformance, Dogfood Verify CLI, Packed-tarball smoke. The one failure is Check Changeset, red by design on this class, the three conditions of SKILL.md :206-207 each standing: (a) the source on the branch says so (.github/workflows/pr-automation.yml:720-726, "If every row is M or D and none is A … CORRECTING somebody else's PENDING release note … do NOT apply 'skip-changeset' … LEAVE THIS CHECK RED", and :936-940, "RED BY DESIGN on the DELIBERATE CORRECTION class"; the PR touches no workflow, so the branch's copy is origin/main's); (b) the workflow runs on pull_request, never merge_group, and Check Changeset is none of the five required contexts; (c) the PR comment 6138885636 records the gate and the reason. The run's own annotation names this route: "if its only .changeset rows are CHANGED, not added, this PR corrects somebody else's pending release note -- do NOT apply 'skip-changeset'; write the confirmation on the PR, naming the note and what changed under it, and leave this check red". The file list confirms the class: 2 modified, 0 added. mergeable_state: unstable follows from that one red. This record is the written confirmation.

② Semver level

• The PR adds no changeset: correct. It publishes nothing new; it corrects the pending declarations of two objectui pin bumps already on main, and only their compiled text changes.
• skip-changeset is absent (labels documentation, size/s, tooling): correct and required. The notes are pending release text, and the workflow's own route for this class bars the label (pr-automation.yml:724).
• Bump level unchanged: both frontmatters byte-identical, "@objectstack/console": minor. The adr-0087 disposition of each note is byte-identical and check-adr-0087-registration reads both as inherited (breaking already at the merge base; scripts/check-adr-0087-registration.mjs:4825-4843), so nothing new is owed to that gate or to check-changeset-no-major; both ride inside the green Lint & Repo Gates.
• Clause-②: no (published changelog text; no accept set or public surface widens), line 2 of the PR body: correct. The diff widens no accept set and no public surface; the parenthetical is reasoning, not an arm (scripts/pm/clause2-line.mjs:291-292 reads it as arm: null), so the declaration is well-formed. Neither note carries a Clause-② line of its own, at base or head, and none is owed by a pin-bump note.

③ Boundary flags

Dev report 6138850179, every flag and the one open_questions entry answered:

1. open_questions 1 — route an update to the two adr-0087 disposition narratives, which now count fewer than the list (A leave / B separate narrative-only correction / C drop the per-entry answers)? Answer for THIS head: nothing is owed, and the verdict does not hang on it. The card's done-criteria and the triage grade 6137918853 hold the disposition blocks byte-identical, and they are (md5 above). The blocks' verdict line, not-required (no-migration-prescription), is read by category, not by count: that category refuses only a body carrying a migration prescription and requires no per-entry answer, and both narratives already scope their claim in their own closing sentence (console-20c6d351ad74.md:88-89, console-f0268ad78485.md:150-151: "a claim about THIS diff, not a per-entry re-measurement of the upstream declared-breaking entries"). The category also survives the ten newly marked entries on their own text: all ten are objectui @object-ui/types / plugin-list / layout narrowings and retirements that follow @objectstack/spec declarations already on this tree. The 20c6 narrative says so by name for three of its four (:80-85, 3fd862510, 025341692, c7b30bd66, "They follow this repo's existing declarations"), and the f026 narrative states the governing principle for the other six (:146-149: the ledger entries belong to the packages/spec PR, never to the pin bump). Measured here from the entries' declaring text and the narratives' own stated principle, not as a per-entry re-measurement of the tree. Not a FAIL item. Escalated, to the triage seat: the sentences "The range carries six declared-breaking objectui changes … Each is answered here" (:60-61) and "four …" (:130-131) are now false against the count lines above them and ship verbatim in the raw packages/console/CHANGELOG.md, the text an upgrading agent greps (AGENTS.md, Documentation Guardrails). Correcting them reopens the byte-identical ruling, which takes a triage word, not this record's. My reading for that word: B, a separate DELIBERATE CORRECTION card on domain:devx that edits the narrative count sentence and adds one measured line per newly marked entry, verdict line unchanged, before 18.0 GA; C deletes answers that are measured and true; A leaves a known false count sentence in published text. ⛔ Not on this PR, and not a condition of this PASS.
2. Deviation 1 — the relay's --route re-exec notice prints the value of HTTPS_PROXY. Confirmed as the tool's own text: scripts/pm/fleet-write/dispatch.mjs:483 interpolates the proxy URL into its "unrouted" reason. The dev printed the variable's name only in the report, which is the right form. Not a contract matter for this PR; a tooling observation for the seat, filed by nobody here.
3. Deviation 2 — dispatch-gates --commands run in its no-path form. Accepted: the derivation from the merge base and the compare API read the same two paths, so the family set is the one the dispatch intended.
4. Deviation 3 — commit trailers are AGENTS.md's model-free pair, not the harness reminder's. Correct per AGENTS.md (Multi-agent discipline, the commit-message rule); the head commit's message carries exactly that pair and no model identifier.
5. NOT MEASURED — the four type-check lanes. Measured by the head's check-runs: Type Check · source gates, · workspace, · debt ledger, · consumer gates all success. Closed.
6. Out-of-scope finding — check-adr-0087-registration's pass line counts inherited-breaking skips as "non-breaking changeset(s) seen". Holds: both the non-breaking branch (scripts/check-adr-0087-registration.mjs:4823) and the inherited-breaking branch (:4843) push into the same skipped list, and the pass line (:8821) prints skipped.length under the non-breaking label. Internal summary text with no public door; the dev recorded it in the PR's Acceptance notes. The seat's call whether to file; my reading is that it qualifies under Prime Directive chore: version packages #10 as a reproducible defect with a named landing point (one line), so a small tooling card is the cleaner route. Not a FAIL item.

No escalation to the maintainer is owed by this record beyond what the landing rules already reserve: the PR is not governed (Governed Surface Queue Guard success; no docs/adr/**, .claude/**, skills/**, AGENTS.md, CLAUDE.md, docs/NORTH-STAR.md or .github/workflows/** path), the head repo is the base repo, and Check Changeset stays red by design. With this record on the head, the three conditions for carrying that red into the queue stand (SKILL.md :206-207), and the ready / arm acts are the owning seat's.

Implemented-by: claude/issue-23136-console-changeset-breaking-count
Reviewed-by: session_012zwsLMxQBwHUYJgPyrpEvx

VERDICT: PASS