[objectstack-fleet]

Requested by Steedos · project thread

Fixes #26849

Clause-②: no
Metadata-protocol: none

维护者速读

• 改了什么:控制台的动作键清单 SPEC_ACTION_KEYS(@object-ui/core)补上 spec ActionSchema 新声明的 disabledReason 一项,附一行注释和一条测试断言。
• 为什么改:feat(spec): a greyed action says why — disabledReason beside disabled #26786 给 spec 加了 disabledReason 后,这份「逐条复述 spec 声明键」的清单没跟上,main 的 Test UI 在 actionKeys.pin.test.ts 上一直红,并传染到每个合过 main 的 ui PR。
• 风险与回滚:极低。只多认一个键,开发态不再把它报成未知键;ActionDef、运行时读取、spec 都不动。回滚即还原这一行。
• 先看:ui/packages/core/src/actions/actionKeys.ts(新增条目与注释);ui/packages/core/src/actions/__tests__/actionKeys.pin.test.ts(新增一行断言)。
• 两分钟验证:pnpm --dir ui exec vitest run packages/core/src/actions/__tests__/actionKeys.pin.test.ts,预期 19 个测试全绿。
• 需要你做的:无,席位落地。

What changes

• SPEC_ACTION_KEYS gains 'disabledReason', after 'disabled', with a comment of the kind 'description' carries: the array restates the spec's declared keys, the entry follows feat(spec): a greyed action says why — disabledReason beside disabled #26786, and the key's readers ship with the console half of [Decision] A greyed-out record action can only say 对此记录不可用 — re-ask for an authored disabled reason now that a second app needs it (re-file per #24034 B′) #26619.
• ActionDef is not changed and nothing in the runner reads the key. No packages/spec/**, server or new-key change.
• One root changeset, @object-ui/core patch.

Tests

• actionKeys.pin.test.ts "lists every key the spec ActionSchema declares" was red on main (missing: ['disabledReason'], main's Test UI 9/16 at aeb9cf0d) and is green here.
• One line beside the existing classifyActionKeys case asserts that an action carrying disabledReason classifies with no unknown and no retired key.

Acceptance notes

• The order's acceptance command, pnpm --filter @object-ui/core exec vitest run src/actions/__tests__/actionKeys, runs vitest from the package directory, which ui/scripts/vitest-invocation-guard.mjs refuses. The run below is its ui/-rooted equivalent from ui/AGENTS.md.

Verification

All at head e80b8cda, in a worktree off origin/main 4702c468.

ui core type-check and the actionKeys suite

• pnpm exec turbo run type-check --filter=@object-ui/core --concurrency=2 under scripts/pm/os-verify-lock.sh: VERDICT command-exit 0, 4/4 tasks (type-check covers the test file).
• pnpm --dir ui exec vitest run packages/core/src/actions/__tests__/actionKeys.pin.test.ts: Test Files 1 passed (1), Tests 19 passed (19).

Derived dispatch gates (34/34)

• node scripts/pm/dispatch-gates.mjs --commands --repo objectstack-ai/objectstack derived 34 commands; every one exited 0, the PR-body gates (check-closing-keyword-parity, check-adr-0087-registration, check-changeset-no-major) against this body.
• pnpm --dir ui exec vitest run scripts/ under the lock: Test Files 178 passed | 2 skipped (180), Tests 5065 passed | 2 skipped (5067).
• pnpm check:type-check-debt under the lock: 0 raw tsc error(s) total, none above its recorded number.
• dispatch-gates --ran: 34 derived famil(ies) accounted for — 34 run, 0 NOT-MEASURED.
• pnpm check:new-line-citations (from ui/): 0 new citation(s).

No ablation: the diff adds one data entry to a list the pin test compares against the spec; the red-before reading is main's own Test UI (9/16) at aeb9cf0d.

---

Generated by Claude Code

[objectstack-fleet]

Seat note · domain:ui#4 · 2026-10-27T19:17Z

Test UI (10/16) on e80b8cda is red on PackageFormDialog.validationTiming.test.tsx:66 (Unable to find a label with the text of: /软件包 ID/). The red is not from this PR. #26815 moved the package-form strings into the lazily loaded studio i18n namespace, and the test reads the zh label before that catalogue loads. The same intermittent red hit main at 4b59358a and PR #26856. It is filed as #26863, and no fix exists yet. This PR's own assertion, actionKeys.pin, is in another shard. The seat re-runs the failed jobs once after the run completes.

---

Generated by Claude Code

[objectstack-fleet]

Seat close · domain:ui#4 · session_01QsR2jTfvjNTaR29ehLk26T

Superseded. #26859 (spec seat 1, part of #26619) merged at 19:19Z with the same SPEC_ACTION_KEYS entry, so this branch now conflicts with main and adds only a comment and one classifyActionKeys assertion. Those are not worth a PR of their own. CI here was green at e80b8cda (43 success, 9 skipped, after one re-run of the flaky #26863 shard). Closing without merging.

---

Generated by Claude Code