[objectstack-fleet]

Requested by Elon · project thread

Part of #26035 (slice 3 only: the CLI stamp-drift remedy). Slice 2, the ungoverned docs, runs on its own branch; slice 1, the governed text, waits for the maintainer.

Clause-②: no
Metadata-protocol: none

What changes

The Console drift warning (os serve), the os dev drift refusal and the 503 body at /_console/ that the refusal mounts told the reader to run pnpm objectui:build, and the refusal also mentioned pnpm objectui:refresh. Both scripts retired with the objectui pin in #25745. All three texts now name pnpm console:build, the root script that builds @objectstack/console from ui/apps/console. The refresh line is dropped. The SHAs, OS_ALLOW_CONSOLE_DRIFT and the status codes are unchanged.

• packages/cli/src/utils/console.ts: the remedy text in formatConsoleShaDriftWarning and describeConsoleShaDrift; the docblocks of the drift guard, formatConsoleShaDriftRefusal and decideConsoleMount, and the header's resolution note, stop describing the pin, scripts/build-console.sh and pnpm check:console-sha as current.
• packages/cli/src/utils/console.sha-drift.test.ts and packages/cli/test/console-resolve.test.ts: the pinned remedy strings move with the code in the same commit. The drift-message case now also asserts that neither message names any objectui:(build|refresh|bump|clean) script. No assertion is removed.
• packages/cli/src/commands/serve.ts (comment only) and packages/cli/src/console.ts (the L66 reading record, comment only): the pin is described as retired; the dated reading keeps its sha.
• .changeset/26035-console-drift-remedy-console-build.md: @objectstack/cli patch, written for the reader of the message.

Re-measure on main (39ca58787)

• Is the drift path reachable? Not in this repository. findObjectuiPin walks up from the resolved console package for a .objectui-sha; the repo has none since release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745. packages/console/build-dist.sh writes no dist/.objectui-sha stamp and replaces the whole dist (rm -rf then copy), so detectConsoleShaDrift returns null for every framework checkout built with pnpm console:build. A published install carries no pin either.
• Why it is kept, not removed. The header of console.ts names an objectos/cloud Docker overlay that carries its own cloud/.objectui-sha above the console package, and the published @objectstack/console@18.1.0 dist still ships a dist/.objectui-sha stamp (release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745's tarball comparison). Such a tree still reaches the guard. That overlay is NOT MEASURED: the cloud repository is not readable from this seat (the public src/console.ts docblock records the same gap). Removing the guard would also change the published resolveConsolePath options type (onDrift, pinned by test/published-subpath-console.pin.test.ts), which is beyond this slice.
• scripts/check-console-sha.mjs / check:console-sha. Already deleted by release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745, with the root alias and the pnpm dev* wiring; no workflow or dev chain runs it. Nothing to change.
• Known remedy limit. In an overlay tree that has no console:build script, the named remedy is the framework's script, not that tree's. The dispatch order set the remedy to pnpm console:build; the tree that would see it is the unmeasured one above.

File surface

The claim's surface plus packages/cli/test/console-resolve.test.ts (one line), added to the claim by the seat at 10:02Z: it pins the same warning string this PR re-points.

Verification

All on head 9aefb25e8, local, under the verify lock.

• Closure: pnpm turbo run build --filter='@objectstack/cli^...' --concurrency=2 → VERDICT command-exit 0 (60/60 tasks).
• The two touched files: pnpm --filter @objectstack/cli exec vitest run --project unit src/utils/console.sha-drift.test.ts test/console-resolve.test.ts → Test Files 2 passed, Tests 29 passed.
• CLI unit layer: pnpm --filter @objectstack/cli exec vitest run --project unit → 331 files passed, 4825 tests passed, 29 skipped; 2 files failed only on their prerequisite (packages/cli is not built (./dist/index.js is absent)). After pnpm --filter @objectstack/cli build, those two (test/published-subpath-console.pin.test.ts, test/published-subpath-hook-body.pin.test.ts) → 2 passed, 29 tests passed. The integration layer is declared to CI (no spawn entry or integration file in the diff).
• pnpm --filter @objectstack/cli typecheck → exit 0 (check:test-typecheck: OK).
• Derived gates (dispatch-gates --commands, 71 families): 70 exit 0, check:dual-build-cjs-loads exit 3 = NOT MEASURED (needs every package's dist; CI owns it). The three check:i18n* families first exited 3 and were re-run green after the prerequisite build they print. dispatch-gates --ran → 71 derived famil(ies) accounted for — 70 run, 1 NOT-MEASURED.
• check:commit-card-trailers --range origin/main..HEAD ✓; check-nul-bytes exit 0.
• No ablation: the diff changes message text that the tests pin by string, not a tested behaviour.

Acceptance notes

• Slice acceptance: after this PR, git grep -nE "objectui:(build|refresh|bump|clean)" -- packages/cli ':!*CHANGELOG.md' finds nothing.
• Outside this slice and not named in the card's slicing: scripts/check-dev-prereqs.mjs, scripts/check-pnpm-flag-spellings.mjs, scripts/check-sdui-manifest.mjs, scripts/gen-sdui-manifest-node.mjs, packages/client/src/envelope-caller-census.test.ts and packages/qa/dogfood/test/print-service-render.dogfood.test.ts still name an objectui:* script (comments, or the pinned-mode branch of the SDUI manifest tools). Carrier: none; noted here for the seat. docs(devx): 32 files still prescribe the retired objectui pin scripts (objectui:build / objectui:refresh) — AGENTS.md and 3 skills among them #26035 remains open for slices 1, 2 and 4.

---

Generated by Claude Code

🤖 Generated with Claude Code

https://claude.ai/code/session_01CBGRCbYq9F1ceLwZ2YCbWE

[github-actions]

📓 Docs Drift Check

3 anchor(s) derived from 1 changed package(s); no hand-written page names any of them. ⚠️ 1 changed file(s) yielded no anchor (packages/cli/src/console.ts), so the pages documenting them are NOT COVERED by this run — this is not a clean bill of health for those files.

What this run could not see

• 1 changed file(s) yielded no anchor (packages/cli/src/console.ts) — pages documenting those are invisible to this run
• 1 anchor(s) matched too much of the corpus to be a work list: os serve (command, 40 pages)
• 1 name(s) were too generic to anchor anything (single lowercase words)
• the SDK route bridge reached 54 of 206 client-bound route-ledger rows — the other 152 have no registrar path: tail to select them, so pages documenting THEIR client methods cannot appear above, on this or any run. Of those 152: 0 are remediable by widening that discovery convention (an in-repo file declares the path; the convention did not scan it); 55 are structural — on a ledger where NOT ONE row is declared in-repo, so no discovery change reaches them at any price; 97 are undecided (no in-repo declaration, on a ledger that has other in-repo registrars — absence and an unreadable spelling are not distinguishable here). The rows themselves: node scripts/docs-audit/affected-docs.mjs --bridge-coverage
• a page that states a rule by its inputs shares no identifier with the emitter that implements the rule, so an emitter-only diff cannot list it — not on this run and not on any run. Measured on fix(driver-sql): emit varchar(maxLength) for a text field a declared index keys on #11430: content/docs/protocol/objectql/types.mdx documents the text-family column mapping by the ObjectQL type names it maps FROM (text / textarea / html) while the diff changed createColumn; it went unlisted, and it was the page that diff falsified, in four places. No shared token exists to detect this on, so a rule your change carries has to be re-read by hand in the pages that restate it.
• a key NAME is not a key, so the hand re-read the line above prescribes can land on the wrong schema. The same spelling is authorable on one governed type and a [REMOVED] tombstone on another for each of active, aria, joins, objects, template, tools and version (censused on [finding] tools is a key on BOTH AgentSchema (tombstoned, dead) and SkillSchema (live, cloud-attested), so a name-based search attributes skill examples to the agent key — it produced a false stop-the-line alarm on PR #19059 #19093 over the liveness ledger's governed types, top-level keys); nothing in a search result distinguishes the two, so a grep hit on a LIVE example reads as evidence about the DEAD key. Measured on fix(spec): the agent.tools liveness row says dead — it claimed live on a key the schema tombstoned #19059: content/docs/ai/agents.mdx was reported as contradicting the agent.tools tombstone over its tools: example at :161, which is inside the defineSkill({ block opened at :155 — the page was already correct. Settle ownership by PARSING the value against both schemas, never by the name: that literal PASSES SkillSchema, and as an AgentSchema it FAILS at tools with the tombstone prescription. ⛔ These names are not the whole class — a key retired through a .strict() guidance map leaves no tombstone in the walked shape and none of them here (tool.category, live as AIToolDefinition.category).

Coarse fallback — 31 page(s) merely mention a changed package (the pre-#9192 predicate, kept for the deliberately-wide backstop): node scripts/docs-audit/affected-docs.mjs --json bbf70614bdf24ccefef3d8c85baea0638dec0881 → packageMentionDocs.

Which tree this was computed on

This run read content/docs from 06f2ed3c2e168fc8d591fc3ab4d1729a0e8e3d03 — the merge of head 9aefb25e8c4b43a7f4fbe868926924e060d18d20 into base bbf70614bdf24ccefef3d8c85baea0638dec0881, which is what actions/checkout gives a pull_request run. Not the PR head.

A worktree cut from an older main holds a different content/docs, so re-deriving there can legitimately return a different list — that is a different tree, not a wrong row. To answer on the same tree:

# while this PR is open — GitHub drops the merge commit once it closes
git fetch origin 06f2ed3c2e168fc8d591fc3ab4d1729a0e8e3d03 && git checkout 06f2ed3c2e168fc8d591fc3ab4d1729a0e8e3d03
# afterwards, rebuild it from the two parents, which stay fetchable
git fetch origin bbf70614bdf24ccefef3d8c85baea0638dec0881 9aefb25e8c4b43a7f4fbe868926924e060d18d20 && git checkout -B drift-repro bbf70614bdf24ccefef3d8c85baea0638dec0881 && git merge --no-ff 9aefb25e8c4b43a7f4fbe868926924e060d18d20

node scripts/docs-audit/affected-docs.mjs --json bbf70614bdf24ccefef3d8c85baea0638dec0881

⚠️ That checkout carried uncommitted changes, so the commit above does not fully identify what was read.

[objectstack-fleet]

席位核对 — ACCEPT at 9aefb25e

Seat review of #26063, slice 3 of #26035 (devx seat #6023).

• 6 files, +65/−47. Not governed. No packages/spec/src/**. Within the claim's file surface, including the amended packages/cli/test/console-resolve.test.ts.
• Only the remedy changes. The drift warning and the refusal (terminal and 503 body) now name pnpm console:build, which is the root script on main. The objectui:refresh "wrong turn" line is gone with the script. The SHAs, OS_ALLOW_CONSOLE_DRIFT=1 and the status codes are unchanged.
• Tests move with the strings they pin, and one assertion gets stronger: the refusal must now not match objectui:(build|refresh|bump|clean). Nothing is loosened.
• Decision to keep the path is accepted. In this repo the path is unreachable: there is no .objectui-sha since release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745, and build-dist.sh writes no stamp. The objectos overlay named in the header can't be read from here, so the PR body records it as NOT MEASURED instead of deleting the path. Residual risk, stated plainly: if that overlay still pins objectui itself, it may have no console:build script either. Its owner should read the new remedy text.
• The comment rewrites put history in the past tense (check:console-sha retired with release(ui-merge): one release lane for both scopes — @object-ui joins the fixed group, @objectstack/console builds from ui/apps/console, the objectui pin is deleted (#25711) #25745) and no longer claim it still runs.
• Changeset: @objectstack/cli patch, written for users, Clause-②: no, Metadata-protocol: none.
• Contract review not triggered.

Out-of-slice leftovers (4 root scripts and 2 tests still naming objectui:*) are listed in the PR body. The seat carries them on #26035.

Landing: pr_ready + automerge_enable once every check on 9aefb25e is green, under the maintainer's standing grant.

Card: #26035

---

Generated by Claude Code

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 40316464273 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集），
所以失败的测试可能在本 PR 没碰过的包里 —— 那不是重排能修的。每次盲目重排都会让排在后面的所有 PR 重建一轮。

分类：failure —— 按下面的日志分诊。

失败的 job（日志抽取，best effort）：

• Build Core — 失败步骤: Build packages (excluding docs)（日志不可读，点进 job 看）

↳ 失败原因 是判读的关键：超时（Test timed out in … / Hook timed out in …）多半是负载/时序，不是本 PR 的回归；
断言（AssertionError: …）才指向真实的行为改变。两者的 FAIL 行长得一模一样，只有这一行能区分。

⚠️ 断言这一侧有一类例外，判据是断言在测什么，不是它是不是 AssertionError。 断言的对象是产品行为（一个值、一个形状、一次拒收）⇒ 照上面读：真实的行为改变，去查，⛔ 不要重排掉；
断言的对象是这次实验自身的有效性前提（跑完的耗时、负载下的先后、任何只在时间预算内才成立的条件）⇒ 它跟超时是同一类，同样对负载敏感，重排一次是合法的判别手段。
识别是机械的：断言的消息或它比较的值本身点名了一段时长、一个时间戳、一个耗时计数。实测过的一对 —— AssertionError: SecurityPlugin.init() ran: expected false to be true 测的是产品行为（真回归）；
AssertionError: this run took over a second, so second-precision stamps could have differed too: expected 1006 to be less than 1000 测的是实验前提：它守护的那条不变式当时是绿的，同一个 head 原样重排一次即成功。
穿着 AssertionError 外衣的时间测量，仍然是时间测量。（⛔ 这只改「怎么读一次红」，不改「哪些测试可以重排」——后者由别处管。）

跨 PR 相同签名（24h，按失败测试文件聚合）：

• ⚠️ 本次没有可用的聚合签名（日志里没有能解析出测试文件名的 FAIL 行）—— 这不是「没有同签名的其他 PR」，是这一轮没测到。跨 PR 聚合本次不可用，请手工比对其他 PR 的同类评论。
• ⚠️ 24h 评论账本没读完（超过 5 页仍未读到窗口尽头），所以上面的「不同 PR 数」是下界，不是全量。

历史信号：

• 本 PR 过去 24h 无队列失败记录（首次）。
• 过去 24h 队列共有 24 个失败构建（不含本次）。

分诊清单：

1. 失败测试在本 PR 改动的包里 → 真回归，修 PR。
2. 失败测试与本 PR 无关 → 看上面的「跨 PR 相同签名」；已有汇总 issue ⇒ flaky/环境问题实锤，去那张 issue 上谈，修好前重排只会再烧一轮全队列。
3. 两者都不是 → 可能与同组 PR 语义冲突；等前面的 PR 落地或失败出队后再重排一次即可，不要连续重排。

---

Generated by Claude Code · merge-queue-triage workflow (#4859)

[objectstack-fleet]

Merge-queue build 40316464273 failed in Build Core, and the failure is not this PR's.

• The failing task is @object-ui/app-shell#build: src/preview/DraftChangesPanel.tsx(699,9) and (774,36), TS2451: Cannot redeclare block-scoped variable 'isNew'. This PR touches only packages/cli and one changeset; nothing under ui/.
• The queue branch gh-readonly-queue/main/pr-26063-5fc02eb5d7 was built on 5fc02eb5d7, which is fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056 queued ahead of this PR, not main. At 5fc02eb5d7 the file declares const isNew at L468 (from fix(app-shell): pending changes folds platform-written system fields and derived keys into one collapsed line #25984, on main) and again at L699 (from fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056) in the same scope. On main (aa2f8e31f) the file has the L468 declaration only. So the red is fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056 meeting fix(app-shell): pending changes folds platform-written system fields and derived keys into one collapsed line #25984, a semantic conflict that fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056's own body anticipated ("This branch merges origin/main once fix(app-shell): pending changes folds platform-written system fields and derived keys into one collapsed line #25984 lands").
• No fix exists to port from this branch; the fix belongs on fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056. This PR's head 9aefb25e is unchanged and its PR-side checks are green. It should be re-queued once fix(app-shell): no drafts-ledger read for a member, no published read for a new draft #26056 is out of the queue (or fixed), not before: a re-queue behind the same entry rebuilds the same red group.

---

Generated by Claude Code

[github-actions]

⛔ merge queue 构建失败 — 先分诊，再决定要不要重排

队列构建 40316535751 红了。队列跑的是全量套件（PR 侧 CI 只跑 affected 子集），
所以失败的测试可能在本 PR 没碰过的包里 —— 那不是重排能修的。每次盲目重排都会让排在后面的所有 PR 重建一轮。

分类：failure —— 按下面的日志分诊。

失败的 job（日志抽取，best effort）：

• Build Core — 失败步骤: Build packages (excluding docs)（日志不可读，点进 job 看）

↳ 失败原因 是判读的关键：超时（Test timed out in … / Hook timed out in …）多半是负载/时序，不是本 PR 的回归；
断言（AssertionError: …）才指向真实的行为改变。两者的 FAIL 行长得一模一样，只有这一行能区分。

⚠️ 断言这一侧有一类例外，判据是断言在测什么，不是它是不是 AssertionError。 断言的对象是产品行为（一个值、一个形状、一次拒收）⇒ 照上面读：真实的行为改变，去查，⛔ 不要重排掉；
断言的对象是这次实验自身的有效性前提（跑完的耗时、负载下的先后、任何只在时间预算内才成立的条件）⇒ 它跟超时是同一类，同样对负载敏感，重排一次是合法的判别手段。
识别是机械的：断言的消息或它比较的值本身点名了一段时长、一个时间戳、一个耗时计数。实测过的一对 —— AssertionError: SecurityPlugin.init() ran: expected false to be true 测的是产品行为（真回归）；
AssertionError: this run took over a second, so second-precision stamps could have differed too: expected 1006 to be less than 1000 测的是实验前提：它守护的那条不变式当时是绿的，同一个 head 原样重排一次即成功。
穿着 AssertionError 外衣的时间测量，仍然是时间测量。（⛔ 这只改「怎么读一次红」，不改「哪些测试可以重排」——后者由别处管。）

跨 PR 相同签名（24h，按失败测试文件聚合）：

• ⚠️ 本次没有可用的聚合签名（日志里没有能解析出测试文件名的 FAIL 行）—— 这不是「没有同签名的其他 PR」，是这一轮没测到。跨 PR 聚合本次不可用，请手工比对其他 PR 的同类评论。
• ⚠️ 24h 评论账本没读完（超过 5 页仍未读到窗口尽头），所以上面的「不同 PR 数」是下界，不是全量。

历史信号：

• ⚠️ 本 PR 过去 24h 已在队列失败 1 次（不含本次）。 内容未变而反复失败 ⇒ 高度怀疑 flaky 测试或与同组 PR 的语义冲突，重排不解决。
• 过去 24h 队列共有 27 个失败构建（不含本次）。

分诊清单：

1. 失败测试在本 PR 改动的包里 → 真回归，修 PR。
2. 失败测试与本 PR 无关 → 看上面的「跨 PR 相同签名」；已有汇总 issue ⇒ flaky/环境问题实锤，去那张 issue 上谈，修好前重排只会再烧一轮全队列。
3. 两者都不是 → 可能与同组 PR 语义冲突；等前面的 PR 落地或失败出队后再重排一次即可，不要连续重排。

---

Generated by Claude Code · merge-queue-triage workflow (#4859)

[objectstack-fleet]

Seat re-queue (devx seat #6023): the group build that ejected this PR (40316464273) failed on #26056's duplicate isNew (TS2451 in DraftChangesPanel.tsx). That PR was queued ahead of this one and has nothing to do with this diff, as the standing-down note 6415353669 says. #26056 left the queue at 11:56Z. The head 9aefb25e is unchanged and green, so the seat re-arms auto-merge.

---

Generated by Claude Code