docs(plans): close residual #6 (R5 merged); register row 40 (wall-clock-sleep absence-assertion test class)

docs/plans/2026-08-15-arch-review-remediation-execution-log.md

@@ -124,8 +124,18 @@ Deliberately not fixed in this program — each has a stated reason, not an over
    5 places (kept at 16 as the DB read-page bound per finding F2) and documented
    `MaxBatchBytes`. Full 57-project suite green bar one unrelated pre-existing flake (register
    row 38). Two OtOpcUa-side findings recorded as register rows 38–39.
-6. **Fragile `SandboxTests` timing pin.** Pre-existing, unrelated to this remediation's changes;
-   noted so it isn't mistaken for a regression if it flakes later.
+6. ~~**Fragile `SandboxTests` timing pin.**~~ **RESOLVED 2026-08-15 (`ca30d17f`+`9fb52153`, R5).**
+   Root-caused causally, not by loosening timeouts. The pin was a wall-clock
+   `CancellationTokenSource` (100 ms) raced against a fixed workload measured at 298 ms — fails
+   outright on faster hardware, and passes *vacuously* in the opposite direction because a
+   pre-cancelled token throws before the script body is entered. Replaced with a deterministic
+   same-thread cancellation edge plus a negative control; 600 ms of test sleeping removed. The
+   deferred sweep ran with it: 3 further instances of the `c4caebe9`/`cfa6acbf` flake class fixed
+   (`ScriptRunShedTests` — where `CollectReport` DRAINS the shed counter, so the fix accumulates
+   across polls; `SyncCallEmissionEndToEndTests`; `DebugStreamBridgeActorTests`), each with
+   fail-before-fix by injected delay + a suppression control; 2 ambiguous candidates cleared with
+   product-code evidence. Test-side only (`git diff src/` empty). A different timing-pin class
+   (wall-clock sleeps asserting absences) was found and registered, not fixed (register row 40).
 7. ~~**Target-scale load test (deferred-work register #25).**~~ **RESOLVED 2026-08-15.** The
    follow-on validation ran: 10 sites x 500 instances x 75 tags = **375,000 live tag
    subscriptions**, 37,518 updates/s achieved vs 37,500 nominal, **45,021,375 updates** over a