feat(ai_client): isolation of current_tier using threading.local() for parallel agent safety

src/multi_agent_conductor.py

@@ -300,7 +300,7 @@ def run_worker_lifecycle(ticket: Ticket, context: WorkerContext, context_files:
    old_comms_cb(entry)
 
  ai_client.comms_log_callback = worker_comms_callback
- ai_client.current_tier = "Tier 3"
+ ai_client.set_current_tier("Tier 3")
  try:
   comms_baseline = len(ai_client.get_comms_log())
   response = ai_client.send(
@@ -313,7 +313,7 @@ def run_worker_lifecycle(ticket: Ticket, context: WorkerContext, context_files:
   )
  finally:
   ai_client.comms_log_callback = old_comms_cb
-  ai_client.current_tier = None
+  ai_client.set_current_tier(None)
  if event_queue:
   # Push via "response" event type — _process_event_queue wraps this
   # as {"action": "handle_ai_response", "payload": ...} for the GUI.