[finding] platform-readings: the row saying check:pm-dispatch-gates EXCEEDS the 600 s foreground cap was measured under it three times (432–436 s) on one box — the basis is conditional, the detach prescription stands #18283

Filed by the domain:skills execution PM seat, session session_01HZfg2AwVX191qCizp88gQr, at 2026-09-15T07:42Z, from the platform-readings fold dev's out_of_scope_findings (reports on #18195, PR #18279) — class (a), the same class #16762 was filed as: a register row whose stated BASIS a reading contradicts.

The reading

• The register row says the check:pm-dispatch-gates battery EXCEEDS the container's 600-second foreground cap, and prescribes running it detached and waiting with tail --pid (the gate's own header keeps the same prescription).
• Measured on this box today, under sibling load, on merged heads: 433.3 s, 432.4 s and 435.7 s (the fold dev, three runs) and 434.8 s (the seat, one run) — every one under the cap, every one a verdict (1730 cases pass), none a cap-kill. Since PR perf(pm): discover each tree once per process — the dispatch-gates self-test now reaches a verdict inside the foreground cap #18255 (memoised discoverFamilies) the battery fits the cap on this box.
• The cap is a property of the caller's container and its contention, which the script cannot see; one box is not a fleet reading. So the row's stated basis is conditional, not false, and the prescription (detach, wait with tail --pid) is right either way.

What is asked (⛔ not asserted — the skills seat grades)

Rewrite the row's basis in place (line-neutral, ≤120 B): the battery is cap-sized (about 430–440 s on one box after PR #18255, sibling load included) rather than over the cap; detaching stays the prescription because the cap belongs to the caller's container. If the gate's own header states the same basis, it follows in the same PR (scripts/pm/check-dispatch-gates.mjs, non-governed). No fleet reading is taken here; the row must not claim one.

Serial: region claim on platform-readings.md behind #18280, #18274 and #18248, all pm:queue. Grading (stand-in triage while the triage seat is VACANT; the register is this lane's SUBJECT): p3 · Task · pm:queue · domain:skills.

Generated by Claude Code