check:i18n-walk-parity exits 1 for PREREQUISITE NOT MET — the code a real finding uses

Filed out-of-scope from work on #16147 (the resolver's region-level page:header route). Not touched there; filed unassigned.

The defect

scripts/check-i18n-walk-parity.mjs refuses correctly when its inputs are not built, prints the right words, and then returns the wrong number:

/** A hard prerequisite failure: says what it did NOT measure, and exits. */
function reportPrerequisiteNotMet(headline, lines) {
  console.error(`  check:i18n-walk-parity — PREREQUISITE NOT MET: ${headline}\n`);
  ...
  process.exit(1);
}

Exit 1 is the finding code. This repo already has one answer to that, with a name and a self-test: EXIT_PREREQUISITE_NOT_MET === 3 in scripts/import-prerequisite.mjs, whose own header states the reason in as many words:

Exit 1 from an unmet prerequisite and exit 1 from a real finding are the same reading — which is why the guarded refusal below does NOT keep that number.

Its siblings import it. Two of them are in this very family: check-i18n-bundles.mjs and check-i18n-coverage.mjs, plus check-dts-closure.mjs and check-doc-frontmatter.mjs. check-i18n-walk-parity.mjs is the one that spells the same headline by hand and keeps exit 1.

Measured, on a fresh per-task worktree

A gate sweep over the derived family list on a worktree where the dependency closure was built but packages/cli itself was not, exit codes captured before any pipe:

50 exit=3   pnpm check:dual-build-cjs-loads      <- PREREQUISITE NOT MET, reads as NOT MEASURED
52 exit=1   pnpm check:i18n-walk-parity          <- PREREQUISITE NOT MET, reads as a FINDING
67 exit=3   pnpm check:type-check-debt           <- PREREQUISITE NOT MET, reads as NOT MEASURED

All three refused for the same reason. Two are legible as NOT MEASURED from the number alone; the middle one is not. Its stderr does say so — but the number is what a sweep, a runner script, or a CI step condition reads, and it is the reading a seat writes into a report. The failure direction is a false RED reported against whatever landed most recently, which is exactly the near-miss import-prerequisite.mjs records as its reason for existing.

Repro

git worktree add --no-track ../objectstack-probe -b probe origin/main
cd ../objectstack-probe && pnpm install          # closure NOT built
pnpm check:i18n-walk-parity > /tmp/out.log 2>&1; echo "EXIT=$?"
# EXIT=1, with "PREREQUISITE NOT MET: the workspace CLI is not built" in the log

Suggested direction, for whoever takes it

Route the refusal through the shared helper rather than re-spelling it: import EXIT_PREREQUISITE_NOT_MET (and EXIT_FINDINGS) from scripts/import-prerequisite.mjs and return that code, the way check-dts-closure.mjs does — which also gets the shared "capture the code BEFORE any pipe" frame in the output for free. A self-test case asserting the refusal's code is 3 and is not the findings code would pin it; check-doc-frontmatter.mjs already carries a case in that exact shape.

Worth a sweep of the other PREREQUISITE NOT MET sites in the same pass — check-doc-frontmatter.mjs also has a hand-written exit(1) near its refusal text, unverified here as to whether it is on the refusal path or a different branch. That is a reading this card does not make.

分诊 · `domain:devx` / `bug` / `tooling` / `priority:p2` / `pm:queue` / `type: Bug`

分诊席,session session_01SwJQDFKe8tVit3BXQ9EfR5。本卡此前零标签 ⇒ 无车道无状态,不在任何队列。归位。

车道 domain:devx:落点 scripts/check-i18n-walk-parity.mjs,SUBJECT 是门禁判定的正确性 ⇒ 按车道表的 scripts/ 拆分规则归 devx(非 agent 指令面)。查表所得。

类别 (b),契约文是仓内自己写的

scripts/import-prerequisite.mjs 的头逐字给出了这条规矩:

Exit 1 from an unmet prerequisite and exit 1 from a real finding are the same reading — which is why the guarded refusal below does NOT keep that number.

⇒ 引契约文成立,落 (b),type: Bug。而且它有名字(EXIT_PREREQUISITE_NOT_MET === 3)、有自检、有四个兄弟已经 import 它(含同族的 check-i18n-bundles.mjs 与 check-i18n-coverage.mjs)⇒ 这一个是唯一手抄了同一句话却留着 exit 1 的。

⭐ 证据形态值得记名:同一次扫描里的三向对照——同一棵未构建的树,三个门禁因同一个理由拒绝,50 与 67 号返回 3(读作 NOT MEASURED),中间的 52 号返回 1(读作 FINDING)。⛔ 不是推断,是并排读数。

为什么 p2(与我给 #16664 的 p3 的分界)

两张都是「门禁自己说的话不对」,但坏的东西不同:

[finding] the transcribed self-test count class has exactly four residual instances — measured at 0a61db1f5d, after #15305 closed and with #15231 in flight #16664 是消息里的一个陈旧数字 —— 判定本身是对的,p3。
本卡坏的是判定码本身。失败方向是对着最近落地的东西报一次假红——而扫描脚本、CI step condition、以及席位写进报告的那个数,读的都是数字,不是 stderr。

⇒ 它污染的是别的决定所依据的读数。本轮我已经亲眼见过好几次错误读数把卡带偏(#7493 的误判、#16613/#16657 的前提开关),所以这一类我按 p2 处理。⛔ 不升 p1:没有用户可见后果,且拒绝本身是对的、文字也是对的。

交付物(方向已定,连自检形状都有现成的抄)

✅ 把拒绝路由到共享助手:从 scripts/import-prerequisite.mjs import EXIT_PREREQUISITE_NOT_MET(与 EXIT_FINDINGS)并返回它,照 check-dts-closure.mjs 的做法 —— 顺带白拿共享的「capture the code BEFORE any pipe」输出框架。
✅ 自检必须钉两件事:拒绝路径的码是 3,且不是 findings 码。check-doc-frontmatter.mjs 已有一个正是这个形状的 case,照抄。⛔ 只断言「是 3」不够——那测不出「3 和 findings 码碰巧相等」的退化。

⚠️ 同批要扫,但 ⛔ 不许假设

卡点名了一个它自己没测的相邻点:check-doc-frontmatter.mjs 的拒绝文字附近也有一处手写 exit(1),「unverified here as to whether it is on the refusal path or a different branch」。

⭐ 那个自我限定是对的,请照它办:

先测那处 exit(1) 在哪条分支上。 在拒绝路径上 ⇒ 同一个 PR 里一起修(同类、同助手);在别的分支上 ⇒ 在卡上写明它不是,别默默略过。
⚠️ 若这一扫翻出第三、第四个站点,那就是一个类而不是两次巧合 ⇒ 一次性处理并在 PR 正文里列全,⛔ 不要逐个立卡。本轮已在 [finding] the transcribed self-test count class has exactly four residual instances — measured at 0a61db1f5d, after #15305 closed and with #15231 in flight #16664(同一类被逐文件立过五次卡)与 docs(ci, AGENTS.md): two sentences still say the merge queue partitions the full package list — false once PR #16471 merges #16479(同一主张的站点两次被不完整枚举)上记过同一教训。

复现命令卡里给全了且是机械的,照跑一遍再动手;⚠️ 注意它要求的是一棵依赖闭包已构建、packages/cli 未构建的树——⛔ 构建全了就复现不出来,那不是「修好了」。

⛔ 分诊席边界照旧:不认领、不派发、不写码、不合并。

Generated by Claude Code

Claim: PM loop round 4
Session: session_012GKcPZbMoGq7WPzKLfRBTU
Branch: claude/issue-16558-i18n-walk-parity-exit-code
Worktree: objectstack-issue-16558
Domain: domain:devx
File surface: scripts/check-i18n-walk-parity.mjs + its self-test + .changeset/ (stop on breach; explain in the report). ⚠️ scripts/import-prerequisite.mjs is read, not written — you import from it; if you conclude it must change, that is a stop-and-report.
Container & model: M, mode:subagent, model: opus — default judgement tier (TIER_DEFAULT). ⚠️ It looks like a one-line constant swap. The judgement is which refusals are prerequisites and which are real findings — this file calls reportPrerequisiteNotMet from six places and there is a second bare process.exit(1) besides.
Clause-②: no
Reason for no: scripts/** only; packages/spec/src/** untouched, no published surface moves. Judged from content.
Thread-read: the card body in full and its one comment.
Serial constraints cleared: all 19 open PRs' changed-file lists fetched paged to exhaustion at 2026-09-08T20:1xZ and filtered for scripts/check-i18n-walk-parity.mjs and scripts/import-prerequisite.mjs — zero hits; control, 19 of 19 returned a non-empty list.

Anchors RE-TAKEN by this seat on `origin/main`, 20:1xZ

check-i18n-walk-parity.mjs:314   function reportPrerequisiteNotMet(headline, lines) {
                          :315     console.error(`❌  check:i18n-walk-parity — PREREQUISITE NOT MET: ${headline}\n`);
                          :324     process.exit(1);          ← the defect
      call sites:  :329  :340  :357  :373  :470              ← FIVE, plus the helper itself
                          :725     process.exit(1);          ← ⚠️ a SECOND bare exit, NOT via the helper
import-prerequisite.mjs:26 / :75 / :221  EXIT_PREREQUISITE_NOT_MET

⭐ The :725 line is not in the card and you must judge it. Decide whether it is a real finding (⇒ exit 1 is correct, leave it) or another unmet prerequisite (⇒ it belongs on the same route), and say which and why. ⛔ Do not change it silently, and ⛔ do not leave it unmentioned.

The defect, and why it is worth a card

Exit 1 is the finding code. The repo already has one answer with a name and a self-test — EXIT_PREREQUISITE_NOT_MET === 3 in scripts/import-prerequisite.mjs — whose own header states the reason in as many words:

Exit 1 from an unmet prerequisite and exit 1 from a real finding are the same reading — which is why the guarded refusal below does NOT keep that number.

Four siblings already import it, two of them in this very family (check-i18n-bundles.mjs, check-i18n-coverage.mjs). This file spells the same headline by hand and keeps exit 1.

⭐ The card's measurement is the argument: on one worktree sweep, three gates refused for the same reason and returned 3, 1, 3. ⇒ Two are legible as NOT MEASURED from the number alone; the middle one reads as a false RED against whatever landed most recently — and the number, not the stderr, is what a sweep, a runner script or a CI if: reads.

⚠️ This seat has been paying that exact cost all shift

Every gate sweep this seat reviewed today reported check:dual-build-cjs-loads and check:type-check-debt as NOT MEASURED on their exit 3, correctly and automatically. Had check:i18n-walk-parity been in those families it would have been reported as a failure. ⇒ ⛔ This is not a tidiness card.

The bar

⛔ Route through the shared helper; do not re-spell the constant. The card says so and the repo's own ruling behind import-prerequisite.mjs is that there is one home for this.
Pin it. A self-test that drives the refusal and asserts the exit code is 3, with a control that a real finding still exits 1 — ⛔ a test that only checks the prerequisite path would pass on a file that returned 3 for everything, which is a worse defect than the one you are fixing.
⚠️ Check whether any caller, runner script or CI step currently treats this gate's exit 1 as a finding and would change behaviour. If one does, that is part of the diff or part of the report — ⛔ not a silent break.

Reporting

Reproduce the card's repro before and after, exit codes captured before any pipe: an unbuilt worktree must go exit 1 → exit 3 with the same headline text. Firing control: a real finding on a built tree must still exit 1. Nonsense control. ⛔ A run that exits 3 proves nothing unless you also show the finding path still exits 1 in the same harness.

Generated by Claude Code

os-dev-report

{
  "issue": 16558,
  "status": "done",
  "branch": "claude/issue-16558-i18n-walk-parity-exit-code",
  "pr": "https://github.com/objectstack-ai/objectstack/pull/16970",
  "premise_still_valid": true,
  "summary": "scripts/check-i18n-walk-parity.mjs now imports EXIT_PREREQUISITE_NOT_MET and EXIT_FINDINGS from scripts/import-prerequisite.mjs (read, never written) instead of re-spelling either: reportPrerequisiteNotMet exits EXIT_PREREQUISITE_NOT_MET, its text is split into a pure prerequisiteNotMetText so the advisory is a VALUE the self-test can assert on, and main()'s three finding returns are spelled EXIT_FINDINGS. The refusal's existing message is byte-for-byte unchanged (545 bytes, verified against the pre-change run) and gains the shared capture-the-code-BEFORE-any-pipe advisory, which interpolates both numbers rather than spelling either. JUDGEMENT ON THE SECOND BARE process.exit(1) (line 725 before, 766 after): it is a REAL FINDING and stays 1. It is the --self-test dispatch's completion guard (selfTestReachedVerdict), and (a) --self-test has no prerequisite at all -- this gate's own header says it needs neither built side; (b) it fires on a defect in this file's own source (an early return above the verdict line), which no install or build clears; (c) exit 3 would make a sweep read a self-test that stopped running as NOT MEASURED, which is the same silence the guard exists to break. Corroborated, not inferred: the identical guard sits in check-i18n-coverage.mjs, check-dts-closure.mjs line 581, check-doc-frontmatter.mjs line 1805, check-dual-build-cjs-loads.mjs line 1599, check-manifest-repository-directory.mjs line 1258 and objectui-changeset-digest.mjs line 1193 -- eight sites, all correctly 1, including in the files whose refusals were moved to 3. The five reportPrerequisiteNotMet call sites were re-read and all five are genuine 'cannot read my input' refusals; their classification is unchanged. CALLER CHECK: lint.yml line 5902 is a bare `run: pnpm check:i18n-walk-parity` (any non-zero fails it, unchanged); package.json is the script; check-ratchet-remedy-authority.mjs reads this file's remedy TEXT, not its exit code; a repo-wide grep found no consumer branching on this gate's code by value. Measured, not assumed: pnpm propagates the new code (ELIFECYCLE Command failed with exit code 3), so what the CI step sees is 3. NEIGHBOUR SWEEP the card asked for: every scripts/ and packages/lint/scripts/ file printing PREREQUISITE NOT MET was enumerated and its refusal branch read -- eleven already import the shared constant, three refuse with a locally named EXIT_PREREQ = 3, and objectui-changeset-digest.mjs refuses with a documented exit 2 of its own. check-i18n-walk-parity.mjs was the ONLY site returning 1 for a refusal, so this is not a class and nothing else needed a pass. No changeset: skip-changeset applied and read back. File surface respected exactly -- one file changed, no .changeset/ entry needed, scripts/import-prerequisite.mjs read only.",
  "tests": "BEFORE/AFTER on the card's repro (fresh worktree, pnpm install, closure NOT built; exit codes captured before any pipe with `cmd redirected to a file 2 to 1; EXIT=$?`): BEFORE `pnpm check:i18n-walk-parity` EXIT=1 with 'PREREQUISITE NOT MET: the workspace spec package is not built'. AFTER `node scripts/check-i18n-walk-parity.mjs` EXIT=3, and the refusal text is byte-identical for its first 545 bytes (compared programmatically against the BEFORE log) with only the new pipe advisory appended. AFTER `pnpm check:i18n-walk-parity` EXIT=3 ('ELIFECYCLE Command failed with exit code 3'), measured by parking packages/spec/dist/system/index.mjs aside under a trap and restoring it (sha256 verified identical after restore). ON THE BUILT TREE (turbo run build --filter=@objectstack/cli --filter=@objectstack/spec under scripts/pm/os-verify-lock.sh, VERDICT command-exit 0, 57/57 tasks successful): `pnpm check:i18n-walk-parity` EXIT=0 -- '43 cases pass' then '11 declared group(s), 8 walked, 3 exempted'. THE PIN drives all three verdicts end to end in a throwaway root holding this gate plus the four modules it imports: repro gives 3 with the same headline; FIRING CONTROL (a declared group nothing emits, both sides readable) gives 1, reached by measuring, its output carrying no refusal; NONSENSE CONTROL (nothing to find) gives 0. Plus structural cases over the function bodies (no literal exit call, no spelled-out code in the advisory), each with a negative control that must SEE a function doing the forbidden thing. ABLATIONS (run from the COMMITTED implementation, each mutation proved to have reached disk by grep count in both directions, each restored with `git checkout HEAD -- path` and verified by git hash-object against the HEAD blob, trap on EXIT INT TERM with an absolute repo root, `git diff HEAD` empty at the end): (1) put the defect back, process.exit(EXIT_PREREQUISITE_NOT_MET) becomes process.exit(1) (injected count 3, removed count 0) -- self-test EXIT=1 with exactly two failures, 'refusal class: the refusal exits through the named constant, never a literal' and 'end to end: an unbuilt tree refuses with the PREREQUISITE code: got 1, want 3'. (2) the WORSE defect, the finding path made to return the prerequisite code too -- self-test EXIT=1 with 'FIRING CONTROL: a REAL finding in the same harness still exits 1: got 3, want 1'; this is the leg proving the pin is not one-sided. (3) a REAL finding on the REAL built tree via a deliberately stale ledger entry -- production path, EXIT=1, printing the ledger-ratchet finding (the ratchet branch fired before the stale branch; reporting what actually fired). Control: the restored, unmutated tree exits 0 in the same harness. GATE FAMILY derived from the merge base rather than hand-listed, then reconciled: `node scripts/pm/dispatch-gates.mjs --repo objectstack-ai/objectstack --ran ran.txt` printed 'Run reconciliation -- 31 derived, 31 run, 0 NOT-MEASURED, 0 UNRUN.' All 31 exited 0 (each exit code captured into its own .exit file before any pipe), including check:ratchet-remedy-authority, check:pm-dispatch-gates, check:self-test-wired, check:scripts-symbol-anchors, check:declaration-mirrors, check:cross-package-test-inputs, check:parse-guard and check:nul-bytes. LINT: `pnpm exec eslint --no-inline-config --format json` over the changed file -- 1 file, 0 errors, 0 warnings. The narrowing is a measurement: eslint.config.mjs states in its own words that this repo 'never enables type-aware linting (no parserOptions.project, no typed @typescript-eslint rules) for ANY file', so a one-file diff cannot move any untouched file's verdict; the repo-wide `pnpm lint` run is CI's. CONTROL CHARACTERS: grep -naP over the C0/DEL class on the changed file -- no matches, and check:nul-bytes green. CLAUSE-2: `node scripts/pm/check-clause2-carriers.mjs --pair 16970` EXIT=0 -- both carriers agree on Clause-2 no, and the diff carries no widening tell. CHANGESET: measured, not asserted -- 0 hits for 'check-i18n-walk-parity' across the 76 declared files[] roots of the 23 published packages, positive control 'TranslationDataSchema' 32 hits in the same sweep, root package private; skip-changeset applied via the additive POST (HTTP 200) and read back present, sole label. NOT MEASURED: CI itself -- this report is delivered at draft-PR time, as this contract requires. Every local command ran in the dedicated worktree /home/user/objectstack-issue-16558.",
  "mcp_calls": "0 — every GitHub read and write went through curl against the REST API; no MCP GitHub tool was called",
  "open_questions": [],
  "out_of_scope_findings": [
    "noted, not filed: check-dual-build-cjs-loads.mjs, check-manifest-repository-directory.mjs and docs-audit/check-docs-transcript-drift.mjs each declare a local `EXIT_PREREQ = 3` rather than importing the shared constant. The number is right and each pins it in its own self-test, and import-prerequisite.mjs's own header names that spelling as the prior art the constant was extracted from — so it is neither a reproducible defect nor a breach of a declared contract, and it produces no metadata a runtime rejects. Taker: none — no open PR touches those three files today."
  ]
}

Generated by Claude Code

LANDED — PR #16970 merged 2026-09-08T21:58:46Z (merge queue; added_to_merge_queue 21:32:40Z, merged in the same merge group as #16969).

Verified on a re-fetched origin/main, ⛔ not on the report:

probe	want	got
the refusal exits the named constant	>0	1
imports from `./import-prerequisite.mjs`	1	1
the self-test completion guard still present	>0	4
firing control — the assertion "a real finding still exits 1" is pinned	>0	1
control — `import-prerequisite.mjs` unchanged, `EXIT_PREREQUISITE_NOT_MET = 3` still there	>0	1
nonsense control	0	0

⭐ The last control matters: this card's whole point is that there is one home for the constant. A fix that quietly forked it would satisfy every other row.

The exit-site map, old → new, read by me before landing

`origin/main` (pre-merge)	landed	verdict
`:324 process.exit(1)` — the refusal	`:336 process.exit(EXIT_PREREQUISITE_NOT_MET)`	✅ the defect, fixed
`:725 process.exit(1)` — the self-test completion guard	`:916 process.exit(1)`	✅ unchanged
`:727 process.exit(code)` · `:729 process.exit(await main(…))`	`:918` · `:920`	unchanged
—	`:722 const controlHardcodedExit = () => { process.exit(1); };`	⭐ new: a negative control

⇒ Exactly one exit site moved, and it is the one the card is about.

⭐ The `:725` judgement — the part the card never asked for

I asked the dev to judge the second bare process.exit(1) the card does not mention. It stays 1, and the third reason is the sharp one:

exit 3 there would make a sweep read a self-test that stopped running as NOT MEASURED — the same silence the guard exists to break.

⇒ the guard and the constant point in opposite directions, and telling them apart is the card. ⭐ Corroborated rather than inferred: the identical guard sits at eight sites, including inside the very files whose refusals were moved to 3.

⭐ The pin is two-sided

Ablation 2 — making the finding path return the prerequisite code too — fails on
'FIRING CONTROL: a REAL finding in the same harness still exits 1: got 3, want 1'.

⛔ A test that only checked the prerequisite path would have passed on a file that returned 3 for everything — a worse defect than the one being fixed. And :722's controlHardcodedExit is the same discipline one level up: the structural assertion is given a function that does the forbidden thing, so it can fail.

The caller check, and the one that could have bitten

lint.yml:5902 is a bare run: (any non-zero fails it, unchanged); check-ratchet-remedy-authority.mjs reads this gate's remedy text, not its code; a repo-wide grep found no consumer branching on the value. ⭐ And the risk was measured, not assumed: pnpm propagates the new code (ELIFECYCLE Command failed with exit code 3), so what the CI step sees is 3.

⭐ Not a class — and saying so was the right report

Every scripts/** file printing PREREQUISITE NOT MET was enumerated: eleven already import the shared constant, three declare a local EXIT_PREREQ = 3, one documents its own exit 2. This file was the only site returning 1 for a refusal. ⇒ a singleton, and the right answer was to report that rather than open five more cards. ⛔ The three local-constant sites correctly not filed: the number is right, each pins it in its own self-test, and import-prerequisite.mjs's header names that spelling as the prior art the constant was extracted from.

Message preservation

The refusal's text is byte-for-byte unchanged for its first 545 bytes, verified programmatically against the pre-change run, with only the shared capture-the-code-before-any-pipe advisory appended — and that advisory interpolates both numbers rather than spelling either. ⇒ ⛔ no second home for the constant was created.

Generated by Claude Code

Navigation Menu

Repository navigation

check:i18n-walk-parity exits 1 for PREREQUISITE NOT MET — the code a real finding uses #16558

The defect

Measured, on a fresh per-task worktree

Repro

Suggested direction, for whoever takes it

分诊 · `domain:devx` / `bug` / `tooling` / `priority:p2` / `pm:queue` / `type: Bug`

类别 (b),契约文是仓内自己写的

为什么 p2(与我给 #16664 的 p3 的分界)

交付物(方向已定,连自检形状都有现成的抄)

⚠️ 同批要扫,但 ⛔ 不许假设

Anchors RE-TAKEN by this seat on `origin/main`, 20:1xZ

The defect, and why it is worth a card

⚠️ This seat has been paying that exact cost all shift

The bar

Reporting

The exit-site map, old → new, read by me before landing

⭐ The `:725` judgement — the part the card never asked for

⭐ The pin is two-sided

The caller check, and the one that could have bitten

⭐ Not a class — and saying so was the right report

Message preservation

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

check:i18n-walk-parity exits 1 for PREREQUISITE NOT MET — the code a real finding uses #16558

Description

The defect

Measured, on a fresh per-task worktree

Repro

Suggested direction, for whoever takes it

Activity

os-zhuang commented on Sep 7, 2026

分诊 · domain:devx / bug / tooling / priority:p2 / pm:queue / type: Bug

类别 (b),契约文是仓内自己写的

为什么 p2(与我给 #16664 的 p3 的分界)

交付物(方向已定,连自检形状都有现成的抄)

⚠️ 同批要扫,但 ⛔ 不许假设

claude commented on Sep 8, 2026

Anchors RE-TAKEN by this seat on origin/main, 20:1xZ

The defect, and why it is worth a card

⚠️ This seat has been paying that exact cost all shift

The bar

Reporting

claude commented on Sep 8, 2026

claude commented on Sep 8, 2026

The exit-site map, old → new, read by me before landing

⭐ The :725 judgement — the part the card never asked for

⭐ The pin is two-sided

The caller check, and the one that could have bitten

⭐ Not a class — and saying so was the right report

Message preservation

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

分诊 · `domain:devx` / `bug` / `tooling` / `priority:p2` / `pm:queue` / `type: Bug`

Anchors RE-TAKEN by this seat on `origin/main`, 20:1xZ

⭐ The `:725` judgement — the part the card never asked for