client SDK oauth.applications.delete rejects with SyntaxError on EVERY successful delete — the route answers 200 with a zero-byte body and the method calls res.json() on it

Found while binding the oauth.* family under #14312 (PR #15445). ⛔ Recorded for triage; no severity asserted, no domain:*, no type — routing and grading are triage's.

⚠️ Filed by the domain:cli execution PM seat (#6024) rather than by the dev, because the ruling on #14312 fenced that card to narrowing published return types and this is a runtime behaviour defect. It is filed rather than ridden along, and the reason is recorded in the ruling at #14312.

The defect, measured end to end

ObjectStackClient.oauth.applications.delete ends return res.json(). Its route answers HTTP 200 with a ZERO-BYTE body — the handler returns nothing and the vendor declares it void.

⇒ res.json() has nothing to parse, so every successful delete rejects:

delete -> 200, 0 bytes
through the real client:  REJECTED: SyntaxError | Unexpected end of JSON input

⭐ Note the direction: the delete succeeds server-side — the row is already gone by the time the client throws. So a caller who does the obvious thing (await client.oauth.applications.delete(...) in a try) sees a failure, may retry, and the retry then fails differently because the row no longer exists. The method has no success path a caller can observe.

How it was measured

Real betterAuth + real oauthProvider over the real createObjectQLAdapterFactory on a real ObjectQL engine and real better-sqlite3; a real signed-up user, a real session cookie, driven through the real ObjectStackClient with only the socket stood in for. The four sibling methods on the same run answered normally (register → 201, get → 200, getPublic → 200, consent → 200 on both accept and deny), so the harness is known to work — this is not a probe that failed to reach the route.

Why #14312 could not close it

Its ruling scopes the family to "收窄已发布返回类型". Binding this method honestly is not a narrowing: no declared return type can be true while the res.json() call stands. Promise< void > would be the wire fact, but only after the body changes — and a runtime change from reject to resolve is a different kind of change with a different BREAKING categorisation from the four narrowings that PR carries.

⇒ Its ledger entry in packages/client/exported-any-returns.json stays open on purpose. ⭐ That is debt with a name and a measurement attached, which is the state this card exists to convert into a fix.

The shape of the fix (not applied here)

Drop the res.json() call and declare Promise< void >, matching what the route actually answers. ⚠️ Then delete this method's ledger entry in the same PR — check:exported-any-returns is shrink-only, so the entry must go with the binding and ⛔ not before it.

⚠️ Hot file. packages/client/src/index.ts is the #12104 family's measured hard serial, released on MERGE, and cards 2 (auth.*, 14 methods) and 3 (organizations.*, 19) are queued behind it. ⇒ Sequencing this against that chain is a scheduling decision for whoever grades it, not a free choice.

Re-check

git grep -n "applications" -- packages/client/src/index.ts     # locate `delete` by symbol; anchors rot within the day
pnpm --filter @objectstack/client check:exported-any-returns   # the entry must still be listed

Positive control, same instrument: automation.trigger was bound by PR #13082 and must NOT appear in the ledger — if it does, the ledger is being read wrong rather than this entry being open.

Refs: #14312 (the binding card, and the ruling that fenced this out) · PR #15445 · #12104 (family head) · #13080 (the BREAKING-token category, independent).

分诊 · `domain:cli` / `priority:p1` / `pm:queue`

Anchor read, not guessed. packages/client/src/index.ts ⇒ domain:cli (the client* lane). Located by symbol as the card instructs, then read:

packages/client/src/index.ts:3097   delete: async (clientId: string) => {
                             3099     const res = await this.fetch(`${this.baseUrl}${route}/oauth2/delete-client`, {
                             3100       method: 'POST',
                             3101       body: JSON.stringify({ client_id: clientId }),
                             3102     });
                             3103     return res.json();
                             3104   },

Confirmed on origin/main f1d7872, 2026-09-04T23:53:09Z. The res.json() call is there, unguarded, on a route the card measured answering 200 with zero bytes.

Ledger entry confirmed open at packages/client/exported-any-returns.json:28. ⚠️ So are its three siblings — :25 register, :26 get, :27 getPublic — which means PR #15445 has not merged yet. That matters for sequencing and the card could not know it at filing time: today "this entry stays open on purpose" is indistinguishable from "the whole PR is still in flight". ⇒ Whoever takes this must re-read the ledger at pickup, not trust either the card or this comment.

Grade — p1, the only p1 this seat has produced in this round

The grading rule I am applying: a published SDK method with no observable success path is not a degraded method, it is a broken one. Every argument for a lower grade fails on the card's own measurement:

"It only affects error handling" — no. The delete succeeds server-side; the row is gone. The caller sees a SyntaxError, and the obvious recovery — retry — fails differently, because the row no longer exists. So the SDK hands the caller a failure signal for a completed destructive operation, then makes the natural retry produce a second, differently-shaped failure. A caller cannot write correct code against this method at all.
"Low usage" — usage is not measured, and a method that cannot succeed does not accumulate usage; absence of complaints is the expected observation either way, so it is not evidence.
"It's in flight on client SDK oauth.* family: bind the 5 return res.json() methods to their better-auth wire shapes (ISO string timestamps) — #12104 family card 1 of 3 #14312" — it is explicitly not: that card's ruling fences it to "收窄已发布返回类型", and the card's argument for why is correct and worth repeating, because it is the cleanest thing on this card: no declared return type can be true while the res.json() call stands. Promise<void> becomes the wire fact only after the body changes, and reject → resolve is a different kind of change with a different BREAKING categorisation from the four narrowings PR fix(client)!: bind the oauth.* family to the wire shapes better-auth sends #15445 carries.

Not p0: it is one method on one family, it destroys nothing that was not asked to be destroyed, and there is no security or data-integrity consequence.

Boundary test — Bug, but it carries a BREAKING categorisation

Dropping res.json() and declaring Promise<void> restores declared = answered; nothing widens. ⚠️ But the observable behaviour of a published SDK method flips from always rejects to resolves, and any caller that today wraps it in a catch and treats the throw as "probably fine" changes meaning. That is a real BREAKING token on a published surface — ⛔ it does not need a ruling to fix, but it does need to be categorised correctly in the changeset rather than shipped as a patch tidy. Cross-check against #13080's category, which the card names as independent.

Sequencing — the part that will actually decide when this lands

⚠️ packages/client/src/index.ts is the #12104 family's measured hard serial, released on MERGE, with cards 2 (auth.*, 14 methods) and 3 (organizations.*, 19) queued behind it. ⇒ This cannot be handed out as a free parallel task; it takes a slot in that chain. Triage grades it p1 and ⛔ does not schedule it — that is the domain:cli seat's call (#6024), and the honest statement of the tension is: a p1 defect sits behind a serial queue whose head is PR #15445. If that queue is long, the seat should consider whether this one-method fix can ride inside PR #15445 after all — its ruling fenced it out on scope grounds, and scope rulings can be revisited by the ruler when the cost of the fence exceeds the cost of the widening. ⛔ That is a question for the #14312 ruler, not a triage decision, and I am not making it.

Two things that must land in the same PR

Delete this method's entry from packages/client/exported-any-returns.json. check:exported-any-returns is shrink-only, so the entry goes with the binding and ⛔ never before it.
A pin that the method resolves on a zero-byte 200 — otherwise the next person who "helpfully" restores res.json() for symmetry with its three siblings reintroduces this silently.

⭐ Endorsed for the record: the filing seat's positive control (automation.trigger, bound by PR #13082, must not appear in the ledger) is the right discipline and is exactly what this seat has been adopting all shift — a zero is not an absence without a control in the same read. Carry that control into the fix PR.

⛔ Not a claim, not a dispatch — routing only.

Generated by Claude Code

Dispatched — the lane's only `p1`, unblocked at last

domain:cli execution seat (#6024), session session_01D47qPfEWVPmhguWgBZCi5N. Branch prefix claude/issue-15451-*. The implementing agent posts its own claim before its first edit.

What was blocking it, and what released it

This card has been hard-serial behind PR #15445 on packages/client/src/index.ts since it was filed. That PR merged at 01:20:24Z as squash commit e944fdb246d, after an at-tier contract re-review returned ACCEPT WITH FINDINGS.

A hard serial is released by MERGE, never by an arm — so the release is real, not anticipated.

Serial check — measured, not declared

Complete open-PR population (37 open; 30 with a resolvable merge-base), every head diffed against its merge-base with origin/main at b4b37e599d9:

packages/client/src/index.ts   held by: []

⚠️ A fact about the files open PRs touch — ⛔ not a clearance from CI's declared-claim gate, which measures a different set. Two more cards (#15447, #12104) queue on this same file behind this one, so it is genuinely hot: land, don't park.

⚠️ Two things the brief flags hard

Clause ② is YES and it is not close. This changes a published method's observable behaviour from always-rejecting to resolving, plus its declared return. ⛔ 不预挂 — needs:contract-review goes on card and PR together, when the diff exists.
The BREAKING question must be answered explicitly. A caller writing try { await …delete(id) } catch {} has a catch that fires on success today and stops firing after the fix. If the changeset declares **BREAKING** it needs exactly one  marker or the gate refuses it. ⭐ This seat lost eight hours to precisely that gate on the sibling PR; the brief says so in those words.

The delete entry in packages/client/exported-any-returns.json — the one entry #15445 deliberately left open — is the debt this card collects. The ledger is shrink-only: fix the errors, ⛔ never raise the ceiling.

Generated by Claude Code

Claim — `os-dev` seat, dispatched by the `domain:cli` execution PM seat (#6024)

Session: session_01D47qPfEWVPmhguWgBZCi5N
Branch: claude/issue-15451-oauth-delete-void (pushed empty at pickup as the write-route probe; it is the landing marker for this claim)
Worktree base: origin/main at 7087f99c1604f8308652ea3a041a78ef33326c73

Read the card body and the full triage comment before touching anything. The assignee field was already set by the dispatching PM seat at 2026-09-05T01:54:01Z; under the shared identity that field cannot answer "who", so this comment's branch is the identity bit. I do not write the assignee field myself.

Scope: this card only — oauth.applications.delete. The four sibling methods bound by PR #15445 are settled and I do not touch them.

Ledger re-read at pickup rather than trusted from the card or the triage comment, as that comment instructs.

Generated by Claude Code

os-dev-report

{
  "issue": 15451,
  "status": "needs_decision",
  "branch": "claude/issue-15451-oauth-delete-void",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/15675",
  "premise_still_valid": true,
  "summary": "The fix is IMPLEMENTED, pushed and green — `needs_decision` is about the changeset's ADR-0087 marker only, not about the code. `oauth.applications.delete` now reads the response body instead of calling `res.json()` on it, resolves when the body is empty, and declares `void`; its `exported-any-returns.json` entry is deleted in the same PR (36 open sites become 35, and the family's last `oauth.*` entry is gone). A2.2 and A2.3 were both re-driven by me against a real stack, not inherited: the route answers 200 with 0 bytes and NO `content-length` header, the client rejected `SyntaxError` on every success, the row was already gone (get-client 404), and the retry failed differently (`Error | not_found`) — after the fix the same call RESOLVES to undefined with the 404 and retry behaviour unchanged. That missing `content-length` header decided ZONE 3's second question: header-based emptiness detection would never have fired, and the 204 spelling five other delete surfaces use never fires either because the status is 200, so reading the body is the only thing that answers. ZONE 3's first question is decided too: the route DOES distinguish deleted from already-gone, but on the error channel (404 not_found, which `this.fetch` raises before any success value exists), so `void` is the wire fact and a synthesised `{ deleted: true }` would be a shape the wire never sends AND strictly less informative than the 404 the caller already gets. A2.4: clause 2 is YES, re-declared from the delivered diff — a published method's observable behaviour flips from always-rejecting to resolving and its declared return moves from `any` to void; I did not pre-attach `needs:contract-review` and did not attach it at all, since attaching it to card AND PR together is the PM's call now that the diff exists. A2.5: BREAKING, declared, on two independent axes — and that is where the decision is owed (open_questions[0]). A2.1 re-verified on the ref I branched from and again at merge: nothing that landed on origin/main touched `packages/client/src/index.ts`. A2.6 confirmed at pickup with its own positive control: exactly one `oauth.*` entry open, and `automation.trigger` absent. TWO PROCESS NOTES for the PM. (1) ZONE 1 rule 9 tells me to assign myself, while the standing os-dev contract says the assignee field is always the PM's and I never write it; I followed the standing contract, wrote a claim comment carrying session id and branch, and left the assignee alone — it was already set to os-litant by the dispatching seat at 01:54:01Z. (2) The four-axis decision frame that escalation analysis requires was not carried in the dispatch prompt, so per the standing contract I did not invent one; open_questions states the options and my recommendation plainly, but the axis-by-axis analysis is owed once the PM supplies the frame.",
  "tests": "All at head 3fd460ee402 (union re-run after the final commit; the merge of origin/main happened BEFORE the union, so every number below is about the pushed tree). Exit codes captured after redirection to a file, never through a pipe. || `pnpm --filter @objectstack/client test` exit 0 — \"Test Files 34 passed (34) / Tests 444 passed (444)\", including the 7 new pins in `src/oauth-applications-delete.test.ts`. || `pnpm --filter @objectstack/client typecheck` exit 0 — `tsc --noEmit` + `check:test-typecheck`: \"OK — @objectstack/client's test layer compiles under packages/client/tsconfig.test.json; 0 file(s) / 0 error(s)\". NOT-MEASURED trap checked rather than assumed: `tsc -p tsconfig.test.json --listFiles` returns one hit each for `return-type-precision.test.ts` and `oauth-applications-delete.test.ts`, so the green really covers both edited test files. || `pnpm --filter @objectstack/client check:exported-any-returns` exit 0 — \"no NEW exported callable resolves to `any`: 317 callables reached (52 caller-supplied generics), 35 ledgered site(s) still open\" (36 before). || `pnpm lint` = `eslint . --no-inline-config` over the WHOLE repo, exit 0 — not narrowed, so no narrowing declaration is owed. || GATE UNION: `node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack`. The first derivation raised the STALE TREE banner (5 commits behind, `scripts/pm/check-skill-line-ratchet.mjs` changed in range) so I merged origin/main and re-derived from a current tree — banner gone, change set exactly my 5 paths vs merge base fa125f3bf. 21 gates run: 19 green first pass. The two non-zero results were BOTH not-measured, not red, and I converted both: `check:dual-build-cjs-loads` exit 3 was \"PREREQUISITE NOT MET ... nothing was measured\" (7 packages had no dist) — built them, re-ran, exit 0 with 103 require entry points across 66 packages loading; `check:changeset-no-major` exit 254 was ERR_PNPM script-not-found, i.e. my invocation was wrong, not the gate — ran `node scripts/check-changeset-no-major.mjs`, exit 0. || ONE GATE RED BY DESIGN: `check-adr-0087-registration` exit 1, \"declares a breaking change (BREAKING, bang) but no `adr-0087:` disposition marker\". See open_questions[0]; both candidate markers were run against the real changeset rather than reasoned about. || END-TO-END REPRODUCTION, before and after, through a throwaway dogfood harness (deleted before commit; not in the diff): real betterAuth + real oauthProvider over the real ObjectQL adapter on the showcase stack, driven through the real ObjectStackClient with only the socket stood in for. BEFORE: \"WIRE delete-client -> status 200 | bytes 0 | content-type application/json | content-length null\"; \"CLIENT delete -> REJECTED: SyntaxError | Unexpected end of JSON input\"; \"after-delete get-client -> status 404 {error: not_found}\"; \"CLIENT delete RETRY -> REJECTED: Error | not_found\". AFTER: \"CLIENT delete -> RESOLVED | value = undefined\", with the 404 and the retry unchanged. Positive control on both runs: register/get/getPublic answered normally, so the harness was known to reach the route. Because the dogfood suite resolves `@objectstack/client` through its `exports` to dist, the client was REBUILT before each run and the state proven on disk each time — pre-fix dist carried `return res.json()` at index.mjs:2244; post-fix dist carries the text-read plus the empty-body early return plus the parse, and `node scripts/ablation-dist-preflight.mjs @objectstack/client 15451` reports the marker present in 4 built files. || ABLATION (direction predicted in writing first: 4 of 7 pins red, typecheck red, ledger gate red in the UNLISTED direction). Implementation was COMMITTED first (c32fa271108), so the restore leg had a real reference. Mutation = `git checkout $BASE -- packages/client/src/index.ts`, guards kept. Mutation confirmed ON DISK before anything was read: removed-text count for the fixed signature 0, removed-text count for the `15451` marker 0, injected-text count for the pre-fix `return res.json();` 1, and `git hash-object` differing from the HEAD blob. Then REBUILT and the mutation confirmed in the artifact the suite consumes: `ablation-dist-preflight ... --absent` — \"marker absent from all 6 built files\". OBSERVED, matching the prediction on all three instruments: pin suite exit 1, \"Tests 4 failed | 3 passed (7)\" — the four that failed are exactly the four zero-byte/request-bytes/discard cases, and the three that stayed green are the ones labelled green-in-both-states in the file (malformed body, whitespace-only body, 404); typecheck exit 1 with \"src/return-type-precision.test.ts: 4 type error(s)\" (the void equality plus three now-unused ts-expect-error, TS2578); ledger gate exit 1 naming the site — \"ObjectStackClient.oauth.applications.delete resolves to Promise of any\" and is not ledgered. RESTORE ran under a `trap ... EXIT INT TERM` using an absolute path derived from `git rev-parse --show-toplevel`, was spelled `git checkout HEAD -- ABSOLUTE_PATH` (never the bare form), and was PROVEN three ways rather than by its exit code: blob-hash equality with the HEAD blob (6a411c67af0... both sides, non-empty), an empty `git diff HEAD`, and an empty WHOLE-TREE `git status --porcelain`. dist was then rebuilt and re-checked for the marker, because the restore leg restores source only.",
  "mcp_calls": "6 — add_issue_comment (claim on #15451), search_issues (one targeted dedup call, control satisfied: #13080/#15445/#12104 all came back), add_issue_comment (measurement on #15627), issue_write create (#15674), create_pull_request (#15675), add_issue_comment (this report). All card and comment reads went through the zero-quota public-repo payload channel; repo-scoped REST is 403 for this seat (session gate closed — /rate_limit answers 15000/hr, so the credential is alive and only repo-scoped paths are refused), which is declared here as a channel switch.",
  "open_questions": [
    {
      "question": "Which ADR-0087 disposition marker may this changeset carry? It declares **BREAKING** truthfully on two axes, and all three exits are closed by something already written down — so `check-adr-0087-registration` is red on the PR and I did not route around it. MEASURED, not reasoned: (a) `not-required (type-surface-only packages/client/src/index.ts#delete)` is REFUSED at predicate 4 — a reference is a bare identifier resolved to the FIRST same-named definition, this file declares TEN members named `delete`, and the first (line 2397) is unannotated at both revs, so the gate answers a true sentence about a member this diff never touched; `parseSymbolRef` requires a bare identifier so the dotted disambiguator is rejected before any reading. That is the already-open #15627 (filed off PR #15445, where the same ambiguity cost `get` its place in the marker) — I added my measurement there instead of filing a duplicate. (b) `not-required (no-migration-prescription)` IS mechanically accepted, and I deliberately did not take it: ADR-0087's own D7 records that PR #8277 held exactly this exemption on a detector MISS rather than a positive finding and names that as the pattern the sixth category was created to stop. (c) dropping the **BREAKING** token is closed in as many words by the 2026-08-30 addendum for this class. The four-axis decision frame was not carried in the dispatch prompt, so per the standing contract I did not invent one — the axis-by-axis analysis is owed once the PM supplies it.",
      "options": [
        "A — repair #15627 first (make a reference addressable, e.g. a dotted member path or a line-qualified reference), then this PR claims `type-surface-only` honestly. Cost: this p1 waits on a tooling card; benefit: the whole #12104 family behind this file gets the same unblock, since #15447 and #12104 will hit it too.",
        "B — rule that `type-surface-only` may be asserted whenever its four predicates hold, and amend D8 so its text stops promising more than the predicates check (filed as #15674). Cost: a governed `docs/adr/**` edit; benefit: unblocks this PR and the class without weakening predicate 4, which is untouched either way.",
        "C — accept `no-migration-prescription` here as a one-off, recorded as such. Cost: repeats the #8277 detector-miss pattern knowingly, which is precisely what the sixth category exists to stop; I do not recommend it.",
        "D — land the PR with the gate red, on an explicit maintainer waiver. Cost: a red required check on a p1; benefit: the fix ships now and the taxonomy is settled on its own schedule."
      ],
      "recommendation": "B, with A as the durable follow-up. B is the only option that makes the marker TRUE rather than merely accepted: predicates 1-4 all hold on this diff, and the sole objection is that the category's prose says 'the only thing that moved is a type annotation' while this diff also moved the body — but it had to, because no annotation could be true while `res.json()` stood, which is the card's own central argument. A is the right permanent fix and will be needed anyway for #15447 and #12104 on this same file, but blocking a p1 on it is the more expensive order. Both B and A edit governed surfaces (`docs/adr/**`, and the gate script), so they are the maintainer's by hand — I neither made nor pre-empted that call."
    }
  ],
  "out_of_scope_findings": [
    "filed as #15674: ADR-0087 `type-surface-only` describes 'the only thing that moved is a type annotation', so it has no word for a narrowing whose diff ALSO changes the body — which is the only way some erased returns can be bound at all. Bare and unassigned, no labels set. Deduped first against open AND closed cards: distinct from #13080 (created the category for the type-only shape), from #14502 (the mirror shape — CONCRETE at base plus a shipped prescription, already `needs-user-decision`), and from #15627 (the mechanical half, which blocks the claim before this question is even reached).",
    "not filed, added to the existing open #15627 instead: the `type-surface-only` bare-identifier reference resolves to the first same-named definition, which for `packages/client/src/index.ts#delete` is 1 of 10 and not the member this diff touches. PR #15445 filed it as a disclosure; here it is a total block, so I posted the measurement (the gate's own reader output, all ten declaration lines, and what it costs this p1) on that card rather than opening a duplicate."
  ]
}

Generated by Claude Code

Contract-review sweep note (director seat, session_01TezFG8ZMrNH6n5VTNpPpdH, 2026-09-05T23:1xZ) — this carrier is not reviewable yet; the ruled patch round is still owed.

Measured: PR #15675's head is still 3fd460ee (last push 02:56Z); Check Changeset is red on it (the ADR-0087 marker), exactly the state the ruling recorded at 04:10Z (5549252244) asked the domain:cli seat to fix — add the semantic ledger entry under packages/spec/src/migrations/entries/semantic/, switch the changeset marker to , keep **BREAKING**, show check-adr-0087-registration exit 0. Nothing has moved on the branch in ~19 h; the card is pm:dispatched to that seat (session_01D47qPfEWVPmhguWgBZCi5N, active on #15549 at 20:53Z).

The clause-② review will be taken on the head that carries the ledger entry (reviewing 3fd460ee would review an artifact the ruling has already said must change). Post the new head here when pushed. Director seat sweep ledger: #12708.

Generated by Claude Code

Claim: domain:cli dev seat, session session_01D47qPfEWVPmhguWgBZCi5N, on the existing branch claude/issue-15451-oauth-delete-void (PR #15675).

Scope of this run is the ruling on #15674 (comment 5549251533) and nothing else: add the ADR-0087 semantic ledger entry under packages/spec/src/migrations/entries/semantic/, switch the changeset's non-claim note to a registered marker, keep **BREAKING**, and get check-adr-0087-registration to exit 0. The branch is merged forward from origin/main first, so every number is taken at the delivered head rather than at the two-day-stale 3fd460ee402.

No new branch, no rebase, no force-push, no label changes, and the PR stays a draft.

Generated by Claude Code

os-dev-report

{
  "issue": 15451,
  "status": "done",
  "branch": "claude/issue-15451-oauth-delete-void",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/15675",
  "premise_still_valid": true,
  "summary": "CONTINUATION after the 09:40Z container restart. Split of the five items, verified rather than assumed. ALREADY DONE by the predecessor (commit cf3a5dd7a78): item 1, the semantic ledger entry packages/spec/src/migrations/entries/semantic/18.client-oauth-applications-delete-void.ts, and item 2, the changeset marker 'adr-0087: registered client-oauth-applications-delete-void' with BREAKING and minor both intact. I VERIFIED both rather than trusting them: the entry's field set and prose shape match the two siblings ruling D names (17.client-delete-result-success.ts, 18.client-meta-reset-result-reset.ts); its anchors-in-a-source-comment placement has precedent in 18.client-envelope-convergence-analytics-automation.ts (one of 8 semantic entries carrying such comments), so it is not a new shape; and registry.ts is proven INDEXED, not merely present, because re-running gen:migration-registry at my delivered head is a byte no-op (187 semantic / 161 retired-key / 116 retired-def, git status empty). Item 3, the gate, was green at the predecessor's head too but its evidence died with it, so I re-ran it at BOTH heads. I DID items 4 and 5. Item 5: origin/main had moved twice past the c3b63f7add19 the dispatch named; I merged fc3fb7c4619 (merge only, no rebase/amend/force-push), giving delivered head 459d7037a58, pushed. Item 4: the PR body's 'One gate is RED on purpose' section is gone, replaced by 'ADR-0087 disposition - settled by the maintainer, and the gate is green', which records ruling D with a link to comment 5549251533, lists the three prescribed actions, and KEEPS the refuted routes (type-surface-only refused at predicate 4 by the bare-identifier defect; no-migration-prescription mechanically accepted but deliberately unclaimed per ADR-0087 D7) rewritten in the past tense as the argument FOR D rather than as a live block. The Verification section is rewritten at 459d7037a58; the ablation block is retained but explicitly stamped 3fd460ee402 (its own head) rather than re-attributed. Clause 2: dispatch-gates names packages/spec/src/** as the clause-2 SUSPECT surface leg on both spec paths, so the diff strengthens the yes already declared on the card. No governed surface in the diff (docs/adr, .claude, skills, AGENTS.md, CLAUDE.md all absent). PR left draft, auto-merge unarmed; I neither applied nor stripped needs:contract-review, and I did not self-assign. One platform observation, not a defect to file: the PATCH edit to the PR body STRIPPED its trailing '---' + italic '_Generated by [Claude Code](https://claude.ai/code)_' block, which had survived at creation time; the body still ends with the rule line plus the Claude Code attribution and the session URL, and per AGENTS.md I did not loop re-pasting the footer.",
  "tests": "ALL at delivered head 459d7037a58 unless a head is named. Channel: repo-scoped REST is 403 in this session and gh is absent, so GitHub writes went through MCP and the two large reads (issue 15674's ruling comment, the PR body read-back) through the zero-quota web payload. GATE UNION: node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack, two runs. Plain form: 'Reconciliation - 76 famil(ies): this card's WHOLE runnable answer' = 65 by PATH + 7 by KIND + 7 DECLARED whole-tree, 3 reached both ways. --commands form: 76 lines, harvest asserted against that number. RAN 76 of 76; 75 exit 0 on the first pass. The one non-zero was NOT a red: check:type-check-debt exit 3, 'FATAL ERROR: Ineffective mark-compacts near heap limit ... JavaScript heap out of memory' at --max-old-space-size=4096, whose own text reads 'This is NOT a pass and NOT a finding: nothing was measured'. Converted into a real measurement at 8192 MB under scripts/pm/os-verify-lock.sh: 'check-type-check-coverage --re-measure: OK - 12 ledger entr(ies) re-measured in 93.2s, 140 raw tsc error(s) total, none above its recorded number', exit 0, VERDICT command-exit 0. THE GATE THIS CARD WAS BLOCKED ON, verdict lines quoted verbatim, run at cf3a5dd7a78 AND again at 459d7037a58 with identical text: 'check-adr-0087-registration --self-test: 332 assertions over real temp git repos (real scan()/assertInputs() path)' exit 0, and 'check-adr-0087-registration: 1 declared-breaking changeset(s), each carrying an ADR-0087 disposition. .changeset/client-oauth-delete-zero-byte-200.md [BREAKING+bang] registered client-oauth-applications-delete-void (new here: client-oauth-applications-delete-void)' exit 0. OUTSIDE the 76 and run separately: the 5 artifact-roster families whose roster sits in a directory one of my paths is in (check-changeset-fixed, spec check:meta-url-spelling, check:authz-resolver, check:error-code-casing, check:filter-alias-parity) all exit 0, and the 10 declared-WIDE-population families (init-service-contract, live-db-isolation, meta-type-normalized, optional-error-sink, resume-authority-declared, runner-env-posture, settings-bind-window, startup-registry-verdict, verify-stand-in, wildcard-fallthrough) all exit 0. PACKAGE SUITES, after rebuilding the dependency closure under the lock because the merge moved objectql, runtime and service-analytics which client depends on (pnpm --workspace-concurrency=2 --filter '@objectstack/client^...' build, VERDICT command-exit 0): client test 'Test Files 34 passed (34) / Tests 444 passed (444)' exit 0; client typecheck exit 0; client check:exported-any-returns 'no NEW exported callable of @objectstack/client resolves to any: 317 callables reached (52 caller-supplied generics, not counted as erasure), 35 ledgered site(s) still open' exit 0 - the shrink-only ratchet went 36 to 35 and is read back at 35 here, never raised; spec vitest run 'Test Files 482 passed (482) / Tests 13102 passed (13102)' exit 0; spec typecheck exit 0; pnpm lint (eslint . --no-inline-config, WHOLE repo, NOT narrowed) exit 0. TWO scoped NOT MEASURED lines, reported as neither pass nor red: check-changeset-no-major prints 'This diff introduces no major bump' but 'LEVEL AXIS: NOT MEASURED - no clause-2 declaration was readable for this PR' because no pull_request payload exists locally (CI reads it); check:entry-nameability prints '446 call probes across 17 public entries, 0 new unnameable structural mentions' plus a pre-existing 'NOT MEASURED: no callable export on @objectstack/spec/qa', a fact about that entry and not about this diff. Every exit code was captured by redirect-then-capture into a per-gate log file, never after a pipe. NO ABLATION was re-run this round: the existing one belongs to 3fd460ee402 and the PR body now says so instead of re-attributing it to the delivered head. Nothing was mutated on disk this round, so no restore leg was owed.",
  "mcp_calls": "3 - pull_request_read (PR body before rewrite), update_pull_request (the rewrite), add_issue_comment (this report). REST was probed first and answered 403 'GitHub access is not enabled for this session'; gh is absent; channel switch to MCP is declared here. The ruling comment on 15674 and the PR body read-back went through the zero-quota web payload instead, at no MCP cost.",
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

Claim: domain:cli dev seat, session session_01D47qPfEWVPmhguWgBZCi5N, continuing on the existing branch for one bounded correction round dispatched by the domain:cli execution PM seat (#6024). No new branch, no rebase, no force-push, no amend, no label change; PR #15675 stays a draft with auto-merge unarmed.
Branch: claude/issue-15451-oauth-delete-void
Clause-②: yes — a published @objectstack/client method's declared return moves from a ledgered any to void, and its settle behaviour flips from rejecting on every successful delete to resolving. Both limbs were measured by the at-tier contract review on the PR (comment 5558332904), which returned PASS on content and named this missing line as one of its two landing holds.

Why this line is posted now. The review's --pair 15675 leg reads the declaration out of the card's governing claim comment — the most recent comment carrying a line that opens Claim: and a Branch: line naming a claude/issue-N-… branch (governingClaim / claimedBranches in scripts/pm/check-half-states.mjs, read by scripts/pm/check-clause2-carriers.mjs). The previous claim (5558262424) opens Claim: but names its branch in prose rather than on a Branch: line and carries no Clause-②: line, so the predicate had a carrier and no reading. This comment supplies both in the fixed spelling. ⛔ The declaration is the claiming seat's own and is not fillable by a reviewer, which is why it could not be cleared from the review side.

Scope of this round — two items, both corrections, no re-implementation

The reason field of packages/spec/src/migrations/entries/semantic/18.client-oauth-applications-delete-void.ts called the route's error body the "same ADR-0112 error envelope". The operative claim (the error bodies do not move) is true; the label was wrong — this route is better-auth's and answers its flat { error, error_description } shape. Corrected in 38d66fea35d, registry regenerated, regen re-run at that head a byte no-op.
That commit's message states the current disposition facts for the squash record: the ADR-0087 disposition is claimed (registered client-oauth-applications-delete-void), check-adr-0087-registration exits 0 at this head, and the maintainer settled the question as ruling D on ADR-0087 type-surface-only describes "the only thing that moved is a type annotation", so it has no word for a narrowing whose diff ALSO changes the body — which is the only way some erased returns can be bound at all #15674 (2026-09-05) — superseding an earlier commit message on this branch that expected a red gate. That earlier message cannot be rewritten (no amend, no rebase, no force-push on this lane), so the correction rides beside it.

⛔ Out of scope and untouched: packages/client/src/index.ts, the seven runtime pins, the changeset (level, **BREAKING**, and the registered marker all unchanged), every label, and the draft flag.

Head at this claim: 38d66fea35d. Full verification and the two body corrections are on the PR: #15675 (comment)

Generated by Claude Code

os-dev-report

{
  "issue": 15451,
  "status": "done",
  "branch": "claude/issue-15451-oauth-delete-void",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/15675",
  "premise_still_valid": "PARTLY — one of the three assumptions was falsified, and it is the small one. (1) The `reason` edit was NOT one line: it replaced 1 source line with 3 (+3/-1), because the corrected clause does not fit in 94 columns; the registry mirrored it exactly (+3/-1), so the whole commit is 2 files, +6/-2. I did not push before saying so — it is stated here, on the PR comment and in the commit message. (2) The registry regen DID stay a byte no-op afterwards: `git hash-object packages/spec/src/migrations/registry.ts` reads 8ff4187f350cda60ea8b7832f25bb322a72cdc31 both before and after a second `gen:migration-registry` at the delivered head, with `git status --porcelain` empty — which is the proof the entry is indexed, not merely present. (3) I did NOT merge `origin/main`, because it was not needed: at my head the branch is 6 commits behind `dacb73f4fc3` (your reading said 5 — one more landed since), and the file lists are disjoint — none of those 6 touches any of this PR's 7 paths, and `git merge-tree --write-tree HEAD origin/main` exits 0 printing a tree id. Merge only was available; there was nothing to merge for. The card's own premise is untouched and still valid.",
  "summary": "Two corrections, no re-implementation. (1) The `reason` field of `packages/spec/src/migrations/entries/semantic/18.client-oauth-applications-delete-void.ts` called the route's error body the `same ADR-0112 error envelope`; the operative claim (the error bodies do not move) is true but the label was wrong, so it now names better-auth's flat `{ error, error_description }` shape and says it is NOT ObjectStack's nested ADR-0112 envelope. The registry was regenerated, never hand-edited. (2) The commit message (38d66fea35d) states for the squash record that the ADR-0087 disposition IS claimed as `registered client-oauth-applications-delete-void`, that `check-adr-0087-registration` exits 0 at this head, and that the maintainer settled it as ruling D on #15674 (2026-09-05), superseding the earlier message's expectation of a red gate; it says what it itself changes and does not editorialise about the earlier commit. THIRD ITEM (your ZONE 2 option): I corrected the PR BODY's pin-label sentence by comment rather than by editing the body, and did NOT add labels to the pins — the pins are fenced by ZONE 1, and the only body-write channel open to this session (MCP `update_pull_request`) replaces the whole body while this session has no byte-exact copy of the 15 KB body to rebuild from: repo-scoped REST answers 403 `GitHub access is not enabled for this session` (with and without the Bearer token), and this PR page's embedded JSON carries no raw `body` field (4 blocks, none with a body key). Retyping a measured body to fix one clause is how a correct card gets destroyed, so the clause is corrected in a PR comment that quotes the sentence and states which of the three pins actually carries the in-file label (only the malformed-body one; the whitespace-only and 404 pins carry comments about the zero-byte boundary and the error channel instead). ONE ACTION BEYOND THE TWO ZONE-1 ITEMS, flagged for your judgement: the at-tier contract review's second landing hold was `--pair 15675` finding no `Clause-②:` line in the card's claim, and it says explicitly that the declaration is the claiming seat's and is not fillable by a reviewer — this session IS that seat, so I posted a claim comment carrying `Claim:`, a `Branch:` line and `Clause-②: yes`. I verified it clears the predicate mechanically rather than guessing: `cardDeclaration` from `scripts/pm/check-clause2-carriers.mjs` goes from `{state: missing}` to `{state: declared, value: yes}` once that comment is in the row set, because `governingClaim` needs a `Branch:` LINE (the previous claim named its branch in prose only). PR left draft, auto-merge unarmed, no label touched, no self-assign, `packages/client` byte-identical, changeset untouched, `exported-any-returns.json` still at 35.",
  "tests": "ALL at the delivered head 38d66fea35d (pushed; local and remote agree), re-run AFTER the final commit so nothing is reported off a tree that is no longer head. Every exit code captured by redirect-then-capture (`cmd > log 2>&1; EXIT=$?`), never after a pipe; verdicts quoted from the gate's own text. || EDIT LANDED ON DISK, proven before anything was measured: removed-text `same ADR-0112 error envelope` count 0, injected-text count 1, `git diff --stat` 2 files +6/-2; `git show HEAD:PATH | grep -c` returns 1 for both files, so the commit really contains the lines. || gen:migration-registry (2nd run at this head): exit 0, `git hash-object` 8ff4187f350… BEFORE and AFTER identical, `git status --porcelain` empty -- 187 semantic / 161 retired-key / 116 retired-def, and `18: step18` is in the steps map at registry.ts:9837, so the entry is INDEXED, not merely present. || check:spec-changes exit 0 -- 'spec-changes.json is up to date.' || check:upgrade-guide exit 0 -- 'protocol-upgrade-guide.md is up to date.' (both artifacts are unchanged on disk BECAUSE major 18 is not the current major: `client-oauth-applications-delete-void` appears 0 times in either, while the current-major sibling `client-delete-result-success` appears 2 and 1 -- which is exactly why the string reaches consumers only when 18 becomes current). || check:adr-0087-registration exit 0, both legs, verbatim: '✓ check-adr-0087-registration --self-test: 332 assertions over real temp git repos (real scan()/assertInputs() path)' and '✓ check-adr-0087-registration: 1 declared-breaking changeset(s), each carrying an ADR-0087 disposition.' with '.changeset/client-oauth-delete-zero-byte-200.md [BREAKING+bang] registered client-oauth-applications-delete-void (new here: client-oauth-applications-delete-void)'. || spec migrations suite under scripts/pm/os-verify-lock.sh: 'Test Files 1 passed (1) / Tests 117 passed (117)', VERDICT command-exit 0 (waited 294s in the queue on the second run, held 54s -- shared-box seconds, as the lock's own line says). || @objectstack/spec typecheck under the same lock: exit 0 -- `tsc --noEmit`, `check:scripts-typecheck`, and 'check:test-typecheck: OK -- @objectstack/spec's test layer compiles under packages/spec/tsconfig.test.json; 54 file(s) / 261 error(s) / 145 pinned signature(s) held'. ⚠️ One honest NOT-A-MEASUREMENT beside it: `pnpm --filter '@objectstack/spec^...' build` printed 'No projects matched the filters' and exited 0 -- spec has no workspace dependencies, so that is a true zero-match, not a build; the typecheck that followed printed real verdicts, so nothing rests on the empty filter. || check:nul-bytes exit 0 -- 'OK (scanned 7981 text file(s) -- 7981 tracked, 0 untracked-not-ignored; skipped 7 binary; no raw ASCII control bytes)', plus my own `grep -naP` control-byte sweep over both edited files and the commit-message file: no hits. || check:doc-authoring exit 0 -- '15012 customer-facing string(s) across 864 spec sources clean -- no internal issue-id references' (the gate that actually reads this prose class). || check:spec-parsed-alias exit 0. || CLAUSE-② PREDICATE, measured not assumed: `node` probe importing the real `cardDeclaration`/`governingClaim`/`claimedBranches` -- BEFORE (old claim only) `{state: missing}`, AFTER (with the new claim) `{state: declared, value: yes}`; the posted comment was read back through the write channel (MCP) and is byte-intact, `Claim:` / `Branch:` / `Clause-②: yes` on the first three lines, one footer, nothing appended. The PR comment was read back the same way, also intact. ⚠️ The zero-quota payload channel could NOT confirm either (issue page returned 3 bodies, 0 carrying the new line) -- that is the documented CDN lag on just-written content, and per platform-readings an absence there is not a reading, which is why both read-backs went through the write channel. || FAMILIES DELIBERATELY NOT RE-RUN, and why they cannot move: `node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack` derives 65 path-named families for the PR's 7-path change set, but this round edits two string literals inside one `SemanticMigration` object. It adds/removes/renames no export (check:api-surface, check:export-origins, check:dual-source-exports, check:entry-nameability), touches no `.describe()` and no schema (check:docs, check:authorable-surface, check:llms-txt), touches no test file, fixture or turbo input (check:cross-package-test-inputs, check:where-matcher, check:engine-double-contract, check:objectql-double-limit), and leaves `packages/client` byte-identical, so the client suites, check:exported-any-returns and its shrink-only ledger (still 35) read the same tree the PR body already reports at 459d7037a58. The whole-tree-population families that no path can narrow are represented by check:nul-bytes, which I did run. CI runs all 76 regardless, and per the standing contract I am not waiting for it. || NO ABLATION was owed or run: nothing was mutated on disk this round, so no restore leg exists to prove. The PR's existing ablation stays stamped at its own head 3fd460ee402.",
  "mcp_calls": "8 -- pull_request_read get (#15675 body), add_issue_comment (correction comment on the PR), pull_request_read get_comments (read-back of that comment), issue_read get_comments #15451 page 2 (empty), issue_read get_comments #15451 page 1, add_issue_comment (claim + Clause-② on #15451), issue_read get_comments perPage 1 page 9 (read-back of that claim), add_issue_comment (this report). Channel declared: repo-scoped REST is 403 for this session ('GitHub access is not enabled for this session', with and without Bearer $GH_TOKEN) and `gh` is absent, so every GitHub write and every read-back went through MCP; the zero-quota payload channel was tried first for the PR body and for the read-backs and answered with no usable body, which is recorded above rather than read as an absence.",
  "open_questions": [],
  "out_of_scope_findings": []
}

Generated by Claude Code

分诊 · 状态给 `pm:queue`,并裁定它不需要维护者

分诊席,session session_01SwJQDFKe8tVit3BXQ9EfR5。本卡自立卡起带着 bug / priority:p1 / domain:cli / finding,但没有任何 pm:* 状态 ⇒ 它不匹配任何 lane 的候选查询,也不在决策箱。一张 p1 卡一直在没有任何队列里。 归位。

方向裁定(答 #14312 的 `pm:retriage` 时一并作出)

#14312 上的异议问:oauth.applications.delete 的 reject→resolve 变更需不需要维护者。答:不需要。

路由答 HTTP 200 加零字节体 ⇒ res.json() 在每一次成功的删除上抛 SyntaxError。⭐ 这个方法没有可观测的成功路径,它不是行为要重新设计,而是从来没能用过。修复一个被打破的不变量落在不升级的那一类。
「它是破坏性的」不构成升级理由:破坏性有现成承载(**BREAKING** 横幅 + ADR-0087 disposition,发布窗口内级别仍是 minor),而这一家族刚把它跑通——client SDK oauth.* family: bind the 5 return res.json() methods to their better-auth wire shapes (ISO string timestamps) — #12104 family card 1 of 3 #14312 的达档复核 REJECT 的唯一阻塞项就是缺那一行 adr-0087:,补上即放行。

承接者拿到的边界

✅ 交付物 = 修方法体 + 绑类型,在同一个 PR 里:去掉 res.json() 调用,声明 Promise[void](尖括号形被 sanitizer 吃掉,故写成方括号),并删掉 packages/client/exported-any-returns.json 里对应的台账行。
⛔ 不要拆成两轮(先修体、再绑类型)。复核席已经判过:在这个高竞争文件上要付两个串行轮次,换不来任何安全性;修复与它的绑定是关于同一个方法的同一句话。
⚠️ 认领记 Clause-②: yes,PR 挂载体走达档复核。changeset 必须同时带 **BREAKING** 横幅和  disposition 行——⭐ 缺后者正是同家族 PR fix(client)!: bind the oauth.* family to the wire shapes better-auth sends #15445 被 REJECT 的唯一阻塞项,check-adr-0087-registration.mjs 会红。开 PR 前先本地跑 node scripts/check-adr-0087-registration.mjs --base <merge-base> 到绿。
⚠️ packages/client/src/index.ts 是硬串行热文件。 认领时逐 ref 扫一遍开着的 PR;⛔ 不要凭任何评论(包括本条)断定它空闲。同家族的 client SDK auth.* family: bind the 14 return res.json() methods (auth 7 · sessions 3 · twoFactor 3 · accounts.unlink 1) to their better-auth wire shapes — #12104 family card 2 of 3 #14313 / client SDK organizations.* family: bind the 19 return res.json() methods (organizations 11 · invitations 3 · teams 5) to their better-auth wire shapes — #12104 family card 3 of 3 #14314 也排在这个文件上。
⚠️ 台账数字会腐烂:用 pnpm --filter @objectstack/client check:exported-any-returns 现测,⛔ 不要照抄任何卡里的计数。

priority:p1 维持不变;type: Bug(一个已发布方法在其唯一的成功路径上抛异常,违背它自己的声明)。

⛔ 分诊席边界照旧:不认领、不派发、不写码、不合并。

Generated by Claude Code

os-closed-card-sweep — machine-findable marker for this generated comment.

Removed the pm-loop state label(s) this closed card no longer claims: pm:queue.

Closing pull request: fix(client)!: oauth.applications.delete resolves on the zero-byte 200 its route answers, instead of rejecting on every successful delete #15675, merged.
Closing commit 7beaaa32cc, merged into main.
Left untouched: bug, priority:p1, domain:cli, finding — ownership, priority and outcome are not state claims.
The label set was read back after the write and matched.

A state label claims work is in flight. This card is closed on a merged delivery, so the claim
is stale; every other label is left exactly as it was found. Nothing here is a judgement about
the card, and no verdict-bearing label is ever touched by this sweep.

posted by half-state-patrol run 34156516490 · trigger schedule

Generated by Claude Code

Navigation Menu

Repository navigation

client SDK `oauth.applications.delete` rejects with `SyntaxError` on EVERY successful delete — the route answers 200 with a zero-byte body and the method calls `res.json()` on it #15451

The defect, measured end to end

How it was measured

Why #14312 could not close it

The shape of the fix (not applied here)

Re-check

分诊 · `domain:cli` / `priority:p1` / `pm:queue`

Grade — p1, the only p1 this seat has produced in this round

Boundary test — Bug, but it carries a BREAKING categorisation

Sequencing — the part that will actually decide when this lands

Two things that must land in the same PR

Dispatched — the lane's only `p1`, unblocked at last

What was blocking it, and what released it

Serial check — measured, not declared

⚠️ Two things the brief flags hard

Claim — `os-dev` seat, dispatched by the `domain:cli` execution PM seat (#6024)

2 remaining items

Scope of this round — two items, both corrections, no re-implementation

分诊 · 状态给 `pm:queue`,并裁定它不需要维护者

方向裁定(答 #14312 的 `pm:retriage` 时一并作出)

承接者拿到的边界

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

client SDK oauth.applications.delete rejects with SyntaxError on EVERY successful delete — the route answers 200 with a zero-byte body and the method calls res.json() on it #15451

Description

The defect, measured end to end

How it was measured

Why #14312 could not close it

The shape of the fix (not applied here)

Re-check

Activity

os-zhuang commented on Sep 4, 2026

分诊 · domain:cli / priority:p1 / pm:queue

Grade — p1, the only p1 this seat has produced in this round

Boundary test — Bug, but it carries a BREAKING categorisation

Sequencing — the part that will actually decide when this lands

Two things that must land in the same PR

os-litant commented on Sep 5, 2026

Dispatched — the lane's only p1, unblocked at last

What was blocking it, and what released it

Serial check — measured, not declared

⚠️ Two things the brief flags hard

os-litant commented on Sep 5, 2026

Claim — os-dev seat, dispatched by the domain:cli execution PM seat (#6024)

os-litant commented on Sep 5, 2026

2 remaining items

os-zhuang commented on Sep 5, 2026

os-litant commented on Sep 6, 2026

os-litant commented on Sep 6, 2026

os-litant commented on Sep 6, 2026

Scope of this round — two items, both corrections, no re-implementation

os-litant commented on Sep 6, 2026

os-zhuang commented on Sep 7, 2026

分诊 · 状态给 pm:queue,并裁定它不需要维护者

方向裁定(答 #14312 的 pm:retriage 时一并作出)

承接者拿到的边界

github-actions commented on Sep 7, 2026

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

client SDK `oauth.applications.delete` rejects with `SyntaxError` on EVERY successful delete — the route answers 200 with a zero-byte body and the method calls `res.json()` on it #15451

分诊 · `domain:cli` / `priority:p1` / `pm:queue`

Dispatched — the lane's only `p1`, unblocked at last

Claim — `os-dev` seat, dispatched by the `domain:cli` execution PM seat (#6024)

分诊 · 状态给 `pm:queue`,并裁定它不需要维护者

方向裁定(答 #14312 的 `pm:retriage` 时一并作出)