Unblock + serially execute 4 PR-cursor-triage fix child tasks (post-stability + Richie a/b/c/d)

Carry-forward from 2026-06-13 PR-cursor-triage parent (b621e74f). Blocked children: - b68e29ec — PR #9584 maskClientId extract (VALID-LOW) - cb05f961 — PR #9996 event-bus ^10.25.0 bump (VALID) - 39fa17b3 — PR #10565 onConnectionWorkerFailed shutdown guard (VALID HIGH) - 24e20961 — PR #10566 tx pass + CustomerStatusEnum rename (VALID, 2 fixes) All currently blocked on (a) host_reboot_storm_2026-06-13 resolution and (b) Richie a/b/c/d decision from 2026-06-13 04:14 UTC DM re: stale PR backlog. Once both gates clear: 1. Verify host stability with bin/host-stability-check.sh (once it ships per priority 2 task). 2. Execute SERIALLY, not in parallel — host storm taught me parallel fan-out into instability is the failure mode. One PR at a time, wait for CI green + DeepSource clean before the next spawn. 3. Start with #10565 (HIGH severity, single-file surgical fix), then #10566, then #9996 (just package bump), then #9584 (trivial refactor). 4. Each fix uses the spawn-commit-trailer.md literal block. Acceptance: 4 small PRs land CI-green over the course of one heartbeat cycle without re-poison-pilling. Reference: LEARNINGS 2026-06-14.