Task name in DB differs from slug `ticket-pickup`. Let me find the actual failed timeout run around 08:35.

Last timeout = 08:35:50. MC-4683 fix deployed ~08:57. Check runs of that task since fix.

No runs since 08:35:50 — 28min gap. Either task disabled or dispatch dead. Let me check task file + whether pickup still happening.

Clear now. ticket-pickup.md intentionally disabled 08:24 (MC-4683/MC-4639 dispatch hardening — legacy duplicate dispatcher, control-room cr/ workers own dispatch now). Last timeout 08:35:50 = tail before disable took effect. Watchdog ticket = artifact of already-resolved outage. Verify dispatch still alive (this session itself = proof, but confirm control-room healthy):

Dispatch alive: this session (mc-MC-4703) + cr-MC-4547 control-room worker both spawned by dispatcher. No timeouts since 08:35:50. Task now disabled. Full picture confirmed: - **Root cause**: timeout inversion (pool claim 600s > task cap 180s) → SIGKILL mid-claim → pool slot corruption → recurring `ticket-pickup` timeouts. Compounded by legacy duplicate dispatcher. - **Already fixed**: MC-4683 (F1-F4 + Q3 timeout coordination) deployed ~08:57 + ticket-pickup.md disabled 08:24 (control-room cr/ workers own dispatch now). - **Verified**: zero timeouts since 08:35:50 (28min clean); task `enabled: false`; dispatch pipeline functional. Watchdog ticket = tail artifact of the 08:35:50 run (last before disable propagated). No code change needed — fix already landed. Work class: operational/watchdog. Report done.