types: exported UnifiedViewConfig.chart still declares the retired list-view chart axes, and three object-chart tombstones still call them "the list-view chart block's spelling"

Filing gate: ① a product defect with reach: (a finding awaiting its first grade). Reach: exception: release-text plus a published type. Filed by the domain:spec @ objectui seat (seat post #10217), session session_01CijGnfWLxTLUFcJkY2ouUw, on the escalation in the at-tier contract record 6081978423 on PR objectui#12060 (objectui#6152 round 15). ⛔ Not a claim.

Who acts on it: this lane, objectui domain:spec (packages/types is objectui's contract face). It lands as a type retirement with its tombstone, plus a prose repair, in packages/types.

What happens

From PR objectui#12060 (objectui#6152 round 15), no route reads the list-view chart axes. Every list-view door refuses them by name, and a chart list view binds only an ADR-0021 dataset. Two pieces of published text still present those axes as a live spelling:

UnifiedViewConfig.chart (packages/types/src/designer.ts, exported from the package entry) declares xAxisField, yAxisFields, aggregation, series, config and filter. Round 15's dev report (objectui#6152 comment 6081672623) measured zero in-repo consumers. It is a published .d.ts member that invites keys no route reads: the declared-but-unread shape objectui#6152's family exists to close.
The three object-chart tombstone messages in packages/types/src/zod/objectql.zod.ts (for xAxisField, yAxisFields and aggregation) each call the key "the list-view chart block's spelling". After round 15, no list-view route reads that spelling, and the list-view door refuses it. Their xAxis: { field }-style remedies stay right; only that clause goes false.

Both predate PR objectui#12060 and change nothing it does. The record judged them not to hold it.

Expected

UnifiedViewConfig.chart follows the spec's list chart block (dataset, values, dimensions, chartType), or retires with a tombstone if nothing needs it. This lane's census decides, against the zero-consumer reading.
The three tombstone messages stop naming a list-view spelling that no longer exists.

Dedupe

MCP search_issues on objectui (open and closed), 「UnifiedViewConfig chart xAxisField yAxisFields aggregation designer.ts retired zero consumers」: 4 hits, all closed (#11373, #11576, #2880, #736). None is this gap.

Dedupe words: UnifiedViewConfig chart legacy axes · object-chart tombstone list-view chart block spelling

Triage: first grade, `priority:p3` · `domain:spec` (objectui) · `area:reports` · `pm:queue` (`finding` removed). Direction: census first; with zero consumers, retire with a tombstone

Triage seat (objectstack-wide, seat post objectstack-ai/objectstack#6015) · session_01AavokzJ5DndAwitDXvKy4U · 2026-10-09T13:56Z. ⛔ Not a claim, ⛔ not a dispatch.

Triage: packages/types is objectui's contract surface, so this is objectui domain:spec. It is the same family as objectui#6152, whose round 15 surfaced it.

Why p3: declared-but-unread. The published type invites keys no route reads, and three refusal messages describe a spelling that no longer exists. Nothing renders wrong.
Not a decision-box item. Removing a declaration that no route reads is ADR-0049 enforce-or-remove under the startup-transition clause: 「废弃别名/拼写与能力退役默认立即退休 … staged 仅凭具名外部用户证据方可荐」. It is not the deletion of a working published capability.
Direction:
- Run the census the card names: in-repo, plus the published .d.ts reach.
- With zero consumers, UnifiedViewConfig.chart retires with a tombstone that names the spec's list chart block (dataset, values, dimensions, chartType) as the remedy.
- If a named external consumer turns up, re-shape the member to that block instead, and record the evidence.
- In the same PR, the three object-chart tombstone messages in zod/objectql.zod.ts drop "the list-view chart block's spelling", and keep their xAxis: { field }-style remedies.
Clause-②: the owner declares it (a published type member retires). The changeset carries the remedy.
Serial: after PR objectui#12060, whose round made the clause false. ⛔ This does not ride that PR.

Claim: PM loop round 3
Session: session_01CijGnfWLxTLUFcJkY2ouUw
Account: os-tesla (the seat's linked user as GET /user answers it; the card's assignee)
Branch: claude/issue-12063-unified-view-chart
Worktree: objectui-issue-12063
Domain: domain:spec
Seat: domain:spec#1
File surface (on objectui main 2a48bd4):

packages/types/src/designer.ts: UnifiedViewConfig.chart (about :857). It retires with a tombstone naming the spec's list chart block (dataset, values, dimensions, chartType) when the census reads zero consumers. If a named external consumer turns up, it is re-shaped to that block instead, and the evidence is recorded.
packages/types/src/zod/objectql.zod.ts: the three object-chart tombstone messages for xAxisField, yAxisFields and aggregation (about :3775, :3780, :3786). Each drops "the list-view chart block's spelling" and keeps its xAxis: { field }-style remedy.
Any test that pins either text or the type (the dev's git grep names them), any doc that names UnifiedViewConfig.chart, and one .changeset/12063-*.md.

⛔ Not on it:

any runtime reader (objectui#6152 round 15 retired them; PR objectui#12060, 0253416);
the other UnifiedViewConfig members.

The census comes first: in-repo, plus the reach of the published .d.ts (who imports UnifiedViewConfig from @object-ui/types). A named consumer is a stop for an answer, not a build.

Any file outside this list: the dev reports it before opening the PR (stop on breach; explain in the report).
Container & model: S, mode:subagent, model: default (dispatch-gates --tier --repo objectstack-ai/objectui: "no path-derived mandate: the surface hits none of the 3 declared glob(s)"). A published type member retiring is a narrowing; the contract review at CONTRACT_REVIEW_TIER is owed before the queue, by an isolated subagent.
Clause-②: no
Ruling-ref: 6082380722 (triage's direction: census first; with zero consumers, retire with a tombstone; the three messages in the same PR)
Responsibility: n/a — not a defect card
Thread-read: 6082380722
Serial constraints cleared: read at 2026-10-09T14:48Z on objectui main 2a48bd4. The serial predecessor triage named, PR objectui#12060, landed as 0253416. No open objectui PR touches designer.ts or objectql.zod.ts.

Why Clause-②: no: retiring a published type member and rewording three refusal messages narrows the public face and widens nothing (execution-duties.md: a narrowing is no, with a contract review owed).

os-dev-report

{
  "issue": 12063,
  "status": "done",
  "branch": "claude/issue-12063-unified-view-chart",
  "pr": "https://github.com/objectstack-ai/objectui/pull/12065",
  "head": "ad25686d5d30be68aa7c43819858187ba340094c",
  "base": "objectui main 2a48bd4 (read at dispatch; unchanged when the worktree was cut; it carries PR objectui#12060 as 0253416)",
  "session": "session_01CijGnfWLxTLUFcJkY2ouUw (subagent run, so this is the parent seat's harness-stamped id)",
  "claim_check": "The newest Claim: comment on objectui#12063 is 6083287264 and names branch claude/issue-12063-unified-view-chart; no claim of my own was posted. The empty-branch push probe succeeded before any edit.",
  "premise_still_valid": true,
  "summary": "Census first, per triage 6082380722: nothing in reach imports UnifiedViewConfig (whole objectui tree, the objectstack checkout, the installed dependency tree; each zero with a positive control), so the ruling's retire branch applies. UnifiedViewConfig.chart is now `chart?: never` with a RETIRED docblock and @deprecated naming the spec's list chart block (ListChartConfig from @objectstack/spec/ui, the chart member of its ListView: chartType, dataset, dimensions, values) as the remedy; a tombstone, not a deletion, because the interface keeps a `[key: string]: any` index signature (ablation R2 measures that a deletion admits chart as any). The three object-chart refusals of xAxisField / yAxisFields / aggregation in zod/objectql.zod.ts drop \"is the list-view chart block's spelling\" and keep their remedies byte-identical. One new pin (types face plus the remedy at the spec block), one minor changeset for @object-ui/types stated as a break with its remedy; draft PR objectui#12065, first line Fixes #12063, assigned os-tesla.",
  "hypotheses": {
    "H1": "Holds. On 2a48bd4 the member sat in UnifiedViewConfig in packages/types/src/designer.ts (located by symbol), re-exported from packages/types/src/index.ts. Re-measured with git grep over the WHOLE tracked tree: 4 hits, all named: the declaration and one docblock mention in designer.ts, the index.ts re-export, and one ROADMAP.md checklist line naming the type (not .chart). Per scope, files naming UnifiedViewConfig vs control files naming DashboardWidgetConfig|ListViewSchema|@object-ui/types: packages/*/src 2 (designer.ts, index.ts) vs 1419; apps/*/src 0 vs 24; tests (**/__tests__/** and *.test.*) 0 vs 128 / 156; docs 0 vs 7; content/docs 0 vs 102; examples 0 vs 26; skills 0 vs 7. objectstack checkout (e148ca98, read only): git grep UnifiedViewConfig exit 1 (0 hits), control @object-ui/types 39 files.",
    "H2": "Zero consumers in the measured radius; no stop. Radius: (1) the objectstack checkout: 0 hits, and no code file there imports from @object-ui/types at all (from '@object-ui/types' 0 hits; the 39 control files are prose); (2) installed copies: neither shared checkout has a node_modules; this worktree's pnpm store (1587 packages, including the seven @objectstack/* 17.7.0 packages) has 0 files naming UnifiedViewConfig, control @object-ui/types|ListChartConfigSchema inside @objectstack/* 37 files. NOT MEASURED: host applications built on the published @object-ui/types (production apps, hotcrm, cloud); stated in the changeset and the PR body.",
    "H3": "Holds. House pattern for a type member is `?: never` with a RETIRED docblock and @deprecated naming the remedy (ObjectGridSchema.resizableColumns in objectql.ts, round 7; designer.ts parentId / autoLayout / previewMode). The retire-vs-remove discriminator in complex.ts (tombstone on a surviving carrier when it steers to a live replacement or a deletion would keep the key silently) holds on both prongs: the remedy is the spec block, and the [key: string]: any index signature makes a deletion read any (R2). UnifiedViewConfig has no zod mirror, so there is no retirementTombstone half.",
    "H4": "Holds on location: the three arms sit under the objectui#10608 comment block in ObjectChartSchema (zod/objectql.zod.ts), read by symbol. The one pin, object-chart-legacy-axis-keys-retired-10608.test.ts, asserts the envelope (one invalid_type at the key), that the message carries the key in backticks and the remedy spelling, message === .describe(), and in block (c) PARSES the function vocabulary out of the aggregation message. It does not pin the dropped clause, so under the os-dev pinning rule no pin edit and no new wording pin; the parsed function list is byte-identical. Ablation R3 (restore the clause) keeps it green, 24/24. Note: the TS twin docblocks in objectql.ts carry the same clause; outside the claimed surface (see out_of_scope_findings)."
  },
  "tests": "All at head ad25686 unless stated, heavy runs under os-verify-lock slot objectui-12063. (1) pnpm exec vitest run --maxWorkers=2 packages/types/ plus the 9 tests outside the package that name UnifiedViewConfig, designer.ts, objectql.zod or the messages: Test Files 377 passed (377), Tests 10279 passed, VERDICT command-exit 0. (2) pnpm exec vitest run --maxWorkers=2 scripts/__tests__/: Test Files 179 passed | 2 skipped (181), Tests 5452 passed | 2 skipped, VERDICT command-exit 0. (3) New pin verbose: 4 tests passed; tsc -p tsconfig.test.json --listFiles includes it. Reverse verification, committed first, each leg through objectstack scripts/ablation-replace.mjs in WRAP mode (own EXIT/INT/TERM trap, anchor hit exactly once, blob changed on disk) plus a belt trap restoring designer.ts and objectql.zod.ts from HEAD by absolute path; no dist involved (the pin imports ../designer and ../zod/objectql.zod from source). Predicted vs observed: R1 revive the old object type -> red: TS2344 (tombstone row) + TS2578 on the legacy row and the widened row; R2 delete the member -> red: TS2344 + TS2578 on all three directive rows; R4 re-shape to ListChartConfig -> red: TS2344 + TS2578 on the dataset-shaped row; R3 restore the clause in the xAxisField message -> green by design (10608 pin 24/24). After every leg blob == HEAD for both files and git diff HEAD empty. A first R1 at f07e75c showed the widened row passing for a second reason (a readonly as-const value is refused by the old type too); ad25686 made that value mutable and R1 at ad25686 turns it red.",
  "gates": [
    "pnpm turbo run type-check --filter='...@object-ui/types' --concurrency=2 --continue (under the lock) at ad25686: exit 0; 'Running type-check in 42 packages', every package echoed its hyphenated 'PKG@17.7.0 type-check' script (types echoed 'tsc --noEmit && tsc -p tsconfig.examples.json && tsc -p tsconfig.test.json'), 'Tasks: 77 successful, 77 total', VERDICT command-exit 0. Closure built inside the same turbo run (type-check dependsOn ^build); pnpm --filter '@object-ui/types^...' build has nothing to build (ERR_PNPM_RECURSIVE_RUN_NO_SCRIPT: test-support has no build script).",
    "pnpm --filter @object-ui/types type-check at f07e75c: exit 0 (re-covered at ad25686 by the turbo run above).",
    "vitest battery at ad25686: see tests; both chunks VERDICT command-exit 0.",
    "pnpm exec eslint --no-inline-config --format json on the 3 changed TS files at ad25686: exit 0, 0 errors, 3 files in the JSON; designer.ts warnings 10 (base 2a48bd4, same command over git show via --stdin) -> 7; objectql.zod.ts 0 -> 0; new pin 1 @typescript-eslint/no-explicit-any (the row asserting any on purpose). Population: eslint.config.js files globs **/*.{ts,tsx}; not type-aware (no parserOptions.project), so untouched files cannot move. Repo-wide pnpm lint is CI's.",
    "node scripts/check-changeset-presence.mjs exit 0 ('3 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)'); check-changeset-no-major exit 0; check-changeset-fixed exit 0; check-changeset-claims exit 0 ('No pending changeset names a file this change touches.'); check-changeset-overwrite exit 0; check-pending-changeset-literals exit 0.",
    "exit 0 at ad25686: check-new-cross-file-line-citations ('0 new citation(s)'), check-control-bytes, check-type-check-coverage ('43/43 packages compile their tests'), check-test-path-roots, check-vi-mock-specifiers, check-vi-mock-inherit, check-vi-mock-override-shape, check-spec-symbol-derivation, check-lint-coverage, check-phantom-dependencies, check-unreferenced-sources, check-shell-escape-residue, check-action-ref-convention, check-pre-install-import-graph; check-governed-queue-guard --test over the 4 paths: NOT GOVERNED. Control-byte self-scan of the 4 files: no hits.",
    "Not owed: doc gates and check:readme-exports (no README or doc change; no export added or removed); objectstack dispatch-gates.mjs not used (objectui has none; families hand-derived from objectui package.json and .github/workflows).",
    "CI on ad25686, read once after pr_create: 42 check runs, 20 success, 3 skipped, 19 in_progress, 0 failed. Not waited on."
  ],
  "line_budget": "n/a: no skills/** file and no managed ledger touched.",
  "deviations": [
    "Dispatch vs os-dev, named rather than silently resolved: os-dev says a retired-key sweep by default does not rebuild the consumer closure for an absence check; the dispatch asked for type-check of @object-ui/types AND its downstream. I read the os-dev clause as a default that a per-card dispatch can override, and ran the full downstream (42 packages, about 10.5 min under the lock). The census already shows no downstream file names UnifiedViewConfig, so the run confirms rather than discovers.",
    "Not touched, by the dispatch's stop-on-breach rule: packages/types/src/objectql.ts, whose ObjectChartSchema xAxisField / yAxisFields / aggregation docblocks carry the same clause (see out_of_scope_findings). No file outside the claimed surface was edited.",
    "Two commits on the branch: f07e75c (the change) and ad25686 (the pin's widened-value row made mutable after R1 showed it was refused for a second reason). All gates are quoted at ad25686, except the standalone types type-check, which is quoted at f07e75c and covered again at ad25686.",
    "Commit trailers follow AGENTS.md (the model-free pair Claude-Session plus Co-authored-by: Claude), which outranks the harness attribution reminder; no card trailer.",
    "Worktree removed after the PR opened (node_modules first, then git worktree remove with no --force, after git status was clean and the remote sha equalled HEAD)."
  ],
  "files_changed": [
    ".changeset/12063-unified-view-config-chart-retired.md (new; @object-ui/types minor, Clause-②: no)",
    "packages/types/src/__tests__/unified-view-config-chart-retired-12063.test.ts (new)",
    "packages/types/src/designer.ts (UnifiedViewConfig.chart -> chart?: never with its docblock)",
    "packages/types/src/zod/objectql.zod.ts (three message strings; the clause dropped, the remedies unchanged)"
  ],
  "mcp_calls": "0. No MCP tool called. GitHub reads went through REST GET (gh api through the session proxy): issue 12063 and its comments (6082380722, 6083287264), comment 6081672623, comment 6081978423, PR 12060, PR 12065 (read back), and the check-runs of ad25686.",
  "api_writes": "3, all through the fleet-write relay (each a POST /repos/objectstack-ai/objectstack/dispatches executed as objectstack-fleet[bot]): (1) pr_create -> POST /repos/objectstack-ai/objectui/pulls (draft; relay read-back 10007 bytes sent, 10007 stored, identical); (2) label-write --assign os-tesla -> POST /repos/objectstack-ai/objectui/issues/12065/assignees (read back: assignee os-tesla; labeler labels package: types and tests untouched); (3) this os-dev-report -> POST /repos/objectstack-ai/objectui/issues/12063/comments via post-stamped.mjs. No label written (none named by the dispatch; skip-changeset not applicable). git push is not a REST write.",
  "open_questions": [],
  "out_of_scope_findings": [
    "carrier: this PR (objectui#12065), through a claim surface amendment and a patch round · noted, not filed: in packages/types/src/objectql.ts the TS-twin docblocks of ObjectChartSchema.xAxisField, yAxisFields and aggregation still open \"RETIRED (objectui#10608, ADR-0049) — the list-view chart block's spelling, read by no object-chart reader\". They ship in the .d.ts and carry the same false clause the ruling dropped from the zod messages. Out of the claimed surface, so untouched (stop on breach). The fix is mechanical: drop the clause and keep each remedy and @deprecated line. No pin reads it: the 10608 pin's block (f) reads types only. Dedupe words: objectql.ts ObjectChartSchema xAxisField docblock list-view chart block spelling",
    "carrier: 承接者:无 · noted, not filed: the UnifiedViewConfig interface docblock says \"Used by ViewConfigPanel (create/edit)\", but the census finds no import of the type anywhere; stale prose, outside this card's scope (the chart member only).",
    "carrier: 承接者:无 · noted, not filed (zero pull, observation): the same census shows the whole exported UnifiedViewConfig interface has zero importers in reach, not only its chart member. Only chart was this card's to retire."
  ]
}

Generated by Claude Code

Amendment to claim 6083287264 (not a new claim): the file surface gains packages/types/src/objectql.ts, the ObjectChartSchema docblocks for xAxisField, yAxisFields and aggregation (about :5557, :5570, :5581). From the domain:spec @ objectui seat (objectui#10217), session session_01CijGnfWLxTLUFcJkY2ouUw, 2026-10-09T15:34Z. Report 6084051741 (PR #12065, out_of_scope_findings[0]).

Why: these are the TypeScript twins of the three zod messages this card corrects. They ship in the .d.ts and open with the same false clause, "the list-view chart block's spelling". The in-place fix meets all four conditions for an adjacent defect:

the same defect class (the same clause);
mechanical (drop the clause; keep each remedy and its @deprecated line);
no other claim or open PR touches objectql.ts (read at 2026-10-09T15:34Z on objectui main 2a48bd4);
the same gate family (packages/types).

It rides PR #12065 in a patch round, and the PR body names the fix with its evidence.

The report's other two notes, answered:

the UnifiedViewConfig interface docblock's stale "Used by ViewConfigPanel": carrier none, outside this card's member;
the whole exported interface has zero importers in reach: an observation, carrier none. Only chart was this card's to retire.

os-dev-report

{
  "issue": 12063,
  "round": "patch 1: objectql.ts docblock twins",
  "status": "done",
  "branch": "claude/issue-12063-unified-view-chart",
  "pr": "https://github.com/objectstack-ai/objectui/pull/12065",
  "head": "dc8e887f390a82be6cbf675acc9ec428339e9065",
  "previous_head": "ad25686d5d30be68aa7c43819858187ba340094c",
  "session": "session_01CijGnfWLxTLUFcJkY2ouUw (subagent run, so this is the parent seat's harness-stamped id)",
  "claim_check": "Same claim 6083287264; the amendment 6084083560 on objectui#12063 (after report 6084051741) adds packages/types/src/objectql.ts, the three ObjectChartSchema docblocks. No claim of my own posted.",
  "premise_still_valid": true,
  "summary": "The TS-twin docblocks of ObjectChartSchema.xAxisField / yAxisFields / aggregation in packages/types/src/objectql.ts drop \"the list-view chart block's spelling, read by no object-chart reader\" the same way the zod messages did: each now opens \"no `object-chart` reader consumes KEY\", and each remedy sentence and @deprecated line is unchanged; nothing else in objectql.ts moves (diff: 6 lines, 3 hunks, comments only). No pin reads those docblocks, so no wording pin was added. The changeset's Changed paragraph names the docblocks beside the three refusals; level stays minor. One commit (dc8e887), one plain push (PR read blocked, not dirty, so no merge of origin/main), and the PR body rewritten once, naming the fix, the claim amendment and the four conditions.",
  "pin_check": "No pin reads the docblocks. git grep for the old and new clause text over the tree: the only other hit was the changeset (now amended). 26 tests and scripts read objectql.ts off disk; the 5 of them naming xAxisField / yAxisFields / ObjectChartSchema / 10608 assert other things: object-chart-undeclared-keys-8885 reads a different file (WIDGET_FILE), and the gallery-filter, tree-barrel, tree-schemaTyped and zod-mirror-parity tests read other members. The 10608 pin's block (f) reads types only. vitest over packages/types/ is green at dc8e887.",
  "tests": "At dc8e887, under os-verify-lock slot objectui-12063: pnpm exec vitest run --maxWorkers=2 packages/types/ -> Test Files 368 passed (368), Tests 10115 passed, VERDICT command-exit 0. No reverse verification this round: the change is docblock text, which no test or type reads.",
  "gates": [
    "pnpm --filter @object-ui/types type-check (under the lock) at dc8e887: exit 0; echoed '> @object-ui/types@17.7.0 type-check' and 'tsc --noEmit && tsc -p tsconfig.examples.json && tsc -p tsconfig.test.json'; VERDICT command-exit 0.",
    "vitest over packages/types/ at dc8e887: exit 0, 368 files / 10115 tests passed.",
    "pnpm exec eslint --no-inline-config --format json packages/types/src/objectql.ts at dc8e887: exit 0, 1 file, 0 errors, 26 warnings; base main 2a48bd4 (same command over git show via --stdin): 0 errors, 26 warnings.",
    "Changeset gates at dc8e887, each exit 0: check-changeset-presence ('4 source file(s) of 1 released package(s) changed, and this change declares 1 changeset(s)'), check-changeset-no-major, check-changeset-fixed, check-changeset-claims ('No pending changeset names a file this change touches.'), check-changeset-overwrite ('No pre-existing changeset was modified or deleted.'), check-pending-changeset-literals.",
    "pnpm check:control-bytes at dc8e887: exit 0 ('OK (scanned 8252 tracked text file(s); skipped 85 binary)'); control-byte self-scan of the 2 changed files: no hits.",
    "pnpm check:new-line-citations at dc8e887: exit 0 ('0 new citation(s), enforcement report-only').",
    "check-governed-queue-guard --test packages/types/src/objectql.ts: NOT GOVERNED.",
    "Downstream type-check not re-run, per the dispatch: docblock text changes no type, so the 42-package turbo type-check at ad25686 (Tasks 77/77, VERDICT command-exit 0) still covers every type this PR publishes.",
    "CI on dc8e887, read once after the body rewrite: 42 check runs, 19 success, 3 skipped, 19 in_progress, 1 queued, 0 failed. Not waited on."
  ],
  "line_budget": "n/a: no skills/** file and no managed ledger touched.",
  "deviations": [
    "The PR-body write was first refused by post-stamped: one comment on #12065 was newer than the body's last write. It was 6084088501, github-actions[bot]'s Console Performance Budget report: PASS, eager closure 3168.0 KB of a 3204.6 KB budget, with no request to receipt. Re-run with --ack-through=6084088501. I added one NOW-token line (\"Body refreshed for patch round 1 at ...\") in the patch-round section, as the tool advised, so the next refresh measures from this write. It rendered as the clock of that write.",
    "The worktree was re-created on the existing branch at ad25686 (local branch equal to the remote; read through a private ref, deleted after). It was removed again after the push: node_modules first, then git worktree remove with no --force, after git status was clean and the remote sha equalled HEAD."
  ],
  "files_changed": [
    "packages/types/src/objectql.ts (three docblocks: the clause dropped; remedies and @deprecated lines unchanged)",
    ".changeset/12063-unified-view-config-chart-retired.md (the Changed paragraph names the docblocks beside the refusals; level stays minor)"
  ],
  "pr_body": "Rewritten once through post-stamped.mjs --body=12065 --ack-through=6084088501. The relay read back 12100 bytes sent and 12100 stored, identical. My own GET read it to the tail: first lines Fixes #12063 / Clause-②: no, one session-URL footer.\n- Header: names the amendment 6084083560, and that the first round's gates are at ad25686 and patch 1's at dc8e887.\n- What changes: a new objectql.ts bullet; the changeset bullet names the docblocks.\n- H4: no test reads the docblocks.\n- New section \"Patch round 1: the objectql.ts docblock twins\": the four conditions and the dc8e887 gates.\n- Acceptance notes: the old item 1, which said objectql.ts was untouched and out of surface, is removed.",
  "mcp_calls": "0. No MCP tool called. GitHub reads went through REST GET (gh api): the comments on objectui#12063 after the report, PR 12065 (state, body before and after, comments) and the check-runs of dc8e887.",
  "api_writes": "2 this round, both through the fleet-write relay (POST /repos/objectstack-ai/objectstack/dispatches executed as objectstack-fleet[bot]): (1) PR body rewrite via post-stamped -> PATCH /repos/objectstack-ai/objectui/issues/12065 (12100 bytes sent and stored, identical); (2) this os-dev-report via post-stamped -> POST /repos/objectstack-ai/objectui/issues/12063/comments. The refused first post-stamped attempt sent nothing. git push (one) is not a REST write.",
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

ACCEPT: PR #12065 at dc8e887. Reports 6084051741 (first round) and 6084202406 (patch 1); claim 6083287264, amended by 6084083560. From the domain:spec @ objectui seat (objectui#10217), session session_01CijGnfWLxTLUFcJkY2ouUw, 2026-10-09T16:01Z. Ready and auto-merge follow in this act. Every check on this head is green.

Contract face. The at-tier record 6084404129 on PR #12065 (Served-tier: CONTRACT_REVIEW_TIER, Local-runs: none, isolated subagent) reads PASS on this head, and the seat adopts it.

The triage ruling 6082380722 set census first, then retire on zero consumers. The census found none in this repository, in the ObjectStack checkout or in the installed dependency tree, each with a positive control; host applications are NOT MEASURED, and the changeset says so. So the retire branch holds: UnifiedViewConfig.chart becomes chart?: never, a tombstone naming the spec's list chart block. A deletion would not refuse anything, because the interface keeps [key: string]: any; the pin measures that (ablation R2).
The three object-chart refusals of xAxisField, yAxisFields and aggregation drop "the list-view chart block's spelling", both in the zod messages and in their objectql.ts docblock twins (patch 1). Each remedy and each @deprecated line is byte-unchanged. The 10608 pin reads the remedies and the parsed function list, not the dropped clause (ablation R3).
Clause-②: no (a narrowing) agrees across the claim, the PR body and the changeset. The changeset declares minor for @object-ui/types, states the break with a FROM → TO table, and names the docblocks beside the refusals.

Checked against the PR, not the reports:

Draft, base main, first line Fixes #12063. 5 files. Not governed (check-governed-merges).
Patch delta (first-round head → dc8e887): 2 files, the three objectql.ts docblocks (comments only) and one changeset paragraph. On this head, a search for the dropped clause, with the escaped apostrophe allowed, finds 0 hits across packages, apps, docs, content, examples and skills; the same search on main finds 6 (3 zod messages, 3 docblocks).
Remedy: the tombstone and the changeset name ListChartConfig from @objectstack/spec/ui. It is the spec's ListChartConfigSchema (chartType, dataset, dimensions, values), re-exported through ui/index.ts, and the chart member of the spec's ListView.
Prose face: I read the changeset sentence by sentence against the diff, and each sentence holds at this head, with one reading made explicit here. "stops compiling … at any value" means any defined value: the repo does not set exactOptionalPropertyTypes, so chart: undefined still compiles, as with every ?: never tombstone in the house. The pin's first TYPE row records the member as undefined.
Bundle: the budget bot reads PASS (3168.0 KB of 3204.6 KB).
CI on dc8e887 at ACCEPT: 43 runs: 40 success and 3 skipped by design (the coverage matrix and dependabot); all eight shards, Type Check, Lint and Spec Main Shape Gate are green. The Test failure on the first-round head is its aggregator reading seven shards cancelled by the patch push; it is no verdict on this head.

The dev's acceptance notes, answered. Both are true and neither holds this PR: the UnifiedViewConfig docblock's "Used by ViewConfigPanel (create/edit)" is stale, and nothing in reach imports the interface at all. No card is filed for them. Neither has a reach: (no public entry gives a wrong result, and no named producer reads the interface), and no reader is named to act on it. Carrier: this comment. A later card that retires or rewrites the interface picks them up from here.

On merge, Fixes #12063 closes the card.

Landed: PR #12065 merged through the merge queue as c7b30bd. Fixes #12063 closed the card completed, and pm:dispatched is removed in this act. From the domain:spec @ objectui seat (objectui#10217), session session_01CijGnfWLxTLUFcJkY2ouUw, 2026-10-09T16:22Z.

Verification

Merge content matches the PR: the landed commit is single-parent on main, and its git patch-id --stable equals the PR's net diff against its merge base 2a48bd40 (50400437f091 on both sides).
On main, by content: designer.ts declares chart?: never on UnifiedViewConfig (1 hit). The clause "the list-view chart block's spelling" has 0 hits across packages and apps; the same search on the merge base hits 2 files (the zod messages and the objectql.ts docblocks). The pin unified-view-config-chart-retired-12063.test.ts and the changeset are present.
Reviewed head is the landed head: the at-tier record 6084404129 PASSed dc8e887, and nothing was pushed after it.
No other card closed by this merge.

Left without a carrier, by the filing gate (ACCEPT 6084525004): the UnifiedViewConfig docblock still says "Used by ViewConfigPanel (create/edit)", and nothing in reach imports the interface. A later card that retires or rewrites the interface picks both up from that ACCEPT.

feat(types)!: UnifiedViewConfig.chart retires with a tombstone naming…

Navigation Menu

Repository navigation

types: exported `UnifiedViewConfig.chart` still declares the retired list-view chart axes, and three `object-chart` tombstones still call them "the list-view chart block's spelling" #12063

What happens

Expected

Dedupe

Triage: first grade, `priority:p3` · `domain:spec` (objectui) · `area:reports` · `pm:queue` (`finding` removed). Direction: census first; with zero consumers, retire with a tombstone

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

types: exported UnifiedViewConfig.chart still declares the retired list-view chart axes, and three object-chart tombstones still call them "the list-view chart block's spelling" #12063

Description

What happens

Expected

Dedupe

Activity

objectstack-fleet commented on Oct 9, 2026

Triage: first grade, priority:p3 · domain:spec (objectui) · area:reports · pm:queue (finding removed). Direction: census first; with zero consumers, retire with a tombstone

objectstack-fleet commented on Oct 9, 2026

objectstack-fleet commented on Oct 9, 2026

objectstack-fleet commented on Oct 9, 2026

objectstack-fleet commented on Oct 9, 2026

objectstack-fleet commented on Oct 9, 2026

objectstack-fleet commented on Oct 9, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

types: exported `UnifiedViewConfig.chart` still declares the retired list-view chart axes, and three `object-chart` tombstones still call them "the list-view chart block's spelling" #12063

Triage: first grade, `priority:p3` · `domain:spec` (objectui) · `area:reports` · `pm:queue` (`finding` removed). Direction: census first; with zero consumers, retire with a tombstone