class (a): pnpm check:pm-dispatch-gates cannot complete inside the container's ~10-minute foreground cap — measured 9m0s then exit 124, twice; the wall-clock twin of #17708 #18201

Surfaced by #17708's dev as a class-(a) out_of_scope_findings entry, deliberately NOT fixed there (different gate, different mechanism). ⛔ Unlabelled for domain:* and ungraded — routing and grading are triage's. ⛔ The filer did not dedupe; dedupe words at the bottom.

Class (a) — a gate every lane derives, runs, and can never read

pnpm check:pm-dispatch-gates cannot complete inside this container's foreground cap.

measured 9m0s and still emitting self-test lines when `timeout 540` killed it
exit 124, reproduced TWICE

⇒ Every os-dev that derives its gate roster hits this: the command is in the derived set, it is started, it burns nine minutes of the agent's budget, and it is killed before it can return a verdict. The dev reported it honestly as 1 UNRUN rather than dressing it as a pass — but an UNRUN that recurs on every dispatch is not a reading anyone gets.

⭐ Why this is worth its own card and not an acceptance note

It is structurally the same defect as #17708, which landed today as f0b2db30d:

#17708's own filer recorded that the finding's discoverer judged it not worth filing and the seat overrode that — and #17708 turned out to have a real root cause (the min was taken over the gate process's limits, not the tsc children's) that no one had seen. ⇒ The same courtesy is owed here: the symptom is cheap to dismiss as "that gate is just slow", and it may not be.

⛔ What is NOT established

• ⛔ No root cause. ⛔ Do not assume it is the same mechanism as [finding] check:type-check-debt cannot be measured under the NODE_OPTIONS ceiling its callers mandate — tsc OOMs at 4033/4096 MB and the gate falls to exit 3, so a lane derives it, runs it, and never reads it #17708 — that card's cause was a heap-ceiling min() over the wrong process, which has no wall-clock analogue. The cause here is unmeasured.
• ⛔ Not established that 9 minutes is inherent. It may be a genuinely large battery, or it may be doing something quadratic. ⛔ Nobody has profiled it.
• The gate reads scripts/pm/dispatch-gates.mjs, which [finding] check:type-check-debt cannot be measured under the NODE_OPTIONS ceiling its callers mandate — tsc OOMs at 4033/4096 MB and the gate falls to exit 3, so a lane derives it, runs it, and never reads it #17708's diff does not touch, so this is not fallout from that landing.

Suggested shape (⛔ a proposal, not a prescription)

1. Measure where the nine minutes go before changing anything — a battery count and a per-battery wall clock would say immediately whether this is size or a hot spot.
2. ⚠️ The temptations to refuse: ⛔ do not shrink the battery to fit the clock, and ⛔ do not mark it continue-on-error in CI — both convert a gate nobody can read into a gate nobody notices, which is the [finding] dispatch-gates.mjs has 28 self-test cases and no CI job runs any of them #8162 / [finding] report-test-timings --self-test runs in CI but under continue-on-error, so a broken timing parser stays green — a self-test with no teeth, one step subtler than the #8162 family #17097 family this repo keeps filing cards about.
3. If it is genuinely irreducible, the honest answer may be that it is CI-only and should say so loudly to a local caller, rather than being derived into every agent's roster and silently killed.

Acceptance

1. A profile, not a guess: where the wall clock goes, with the battery count beside it.
2. Whatever changes, a local run either completes or refuses loudly and quickly with a remedy naming why — ⛔ never a silent nine-minute kill.
3. ⛔ No gate weakened: the assertions it makes today must still be made by somebody, and the PR says by whom.

Dedupe words

check:pm-dispatch-gates · foreground cap exit 124 · gate cannot complete locally · self-test wall clock · derived gate never read

⚠️ Related, ⛔ not asserted as duplicates: #17708 (the heap twin, landed f0b2db30d) and #17097 (a self-test running under continue-on-error, i.e. the other way a gate stops being read).

Generated by Claude Code