feat: activity video track (pulsing orb) for voice sessions

- ActivityVideoPublisher renders animated orb on 160x120 canvas
- Integrated into both agent.py and voice.py
- Updates confluence-collab submodule

voice.py

@@ -18,6 +18,7 @@ import aiohttp
 import httpx
 from livekit import rtc, api as lkapi
 from livekit.agents import Agent, AgentSession, StopResponse, function_tool, room_io, llm
+from activity_video import ActivityVideoPublisher
 from livekit.plugins import openai as lk_openai, elevenlabs, silero
 from openai import AsyncOpenAI
 
@@ -463,6 +464,8 @@ class VoiceSession:
         self.lk_room = None
         self.session = None
         self._task = None
+        self._activity_video = None
+        self._activity_task = None
         self._http_session = None
         self._caller_key: bytes | None = None
         self._caller_identity: str | None = None
@@ -575,6 +578,10 @@ class VoiceSession:
                         await obj.close()
                 except Exception:
                     pass
+        if self._activity_video:
+            self._activity_video.stop()
+        if self._activity_task and not self._activity_task.done():
+            self._activity_task.cancel()
         if self._task and not self._task.done():
             self._task.cancel()
             try:
@@ -779,6 +786,17 @@ class VoiceSession:
             if remote_identity:
                 logger.info("Linking to remote participant: %s", remote_identity)
 
+            # Publish activity video track (animated waveform bars)
+            try:
+                self._activity_video = ActivityVideoPublisher()
+                video_track = rtc.LocalVideoTrack.create_video_track("activity", self._activity_video.source)
+                pub_opts = rtc.TrackPublishOptions(source=rtc.TrackSource.SOURCE_CAMERA)
+                await self.lk_room.local_participant.publish_track(video_track, pub_opts)
+                self._activity_task = asyncio.create_task(self._activity_video.run())
+                logger.info("Activity video track published")
+            except Exception as e:
+                logger.warning("Failed to publish activity video: %s", e)
+
             # Load memories and user preferences for this caller
             memory_section = ""
             user_timezone = None
@@ -1217,6 +1235,12 @@ class VoiceSession:
             )
             logger.info("Voice pipeline started (voice=%s, linked_to=%s)", voice_id, remote_identity)
 
+            # Wire agent state to activity video animation
+            if self._activity_video:
+                @self.session.on("agent_state_changed")
+                def _on_state_changed(ev):
+                    self._activity_video.set_state(ev.new_state)
+
             try:
                 await asyncio.wait_for(
                     self.session.generate_reply(