fix: voice-agent crash — add room I/O options and filter AgentConfigUpdate

- Add room_input_options/room_output_options to session.start() so agent
  binds audio I/O and stays in the room
- Add wait_for_participant() before starting session
- Filter AgentConfigUpdate items in agent_llm.py (no 'role' attribute)

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

packages/services/voice-agent/src/agent.py

@@ -176,9 +176,14 @@ async def entrypoint(ctx: JobContext) -> None:
         tts=tts,
     )
 
+    # Wait for the user participant to connect before starting
+    await ctx.wait_for_participant()
+
     await session.start(
         agent=IT0VoiceAgent(),
         room=ctx.room,
+        room_input_options=room_io.RoomInputOptions(),
+        room_output_options=room_io.RoomOutputOptions(),
     )
 
     logger.info("Voice session started for room %s", ctx.room.name)

packages/services/voice-agent/src/plugins/agent_llm.py

@@ -91,10 +91,13 @@ class AgentServiceLLMStream(llm.LLMStream):
 
     async def _run(self) -> None:
         # Extract the latest user message from ChatContext
+        # items can contain ChatMessage and AgentConfigUpdate; filter by type
         user_text = ""
-        for msg in reversed(self._chat_ctx.items):
-            if msg.role == "user":
-                user_text = msg.text_content
+        for item in reversed(self._chat_ctx.items):
+            if getattr(item, "type", None) != "message":
+                continue
+            if item.role == "user":
+                user_text = item.text_content
                 break
 
         if not user_text: