update mobile pairing flow and audio session handling

Improve pairing reliability and UX by letting users choose among scanned hosts with health checks and cleaner row styling while shrinking QR payloads. Handle iOS call-time audio session conflicts more gracefully with user-friendly messaging and lower-noise logs.

Author: R44VC0RP

packages/mobile-voice/app.json

@@ -82,9 +82,7 @@
       }
     },
     "owner": "anomaly-co",
-    "runtimeVersion": {
-      "policy": "appVersion"
-    },
+    "runtimeVersion": "1.0.0",
     "updates": {
       "url": "https://u.expo.dev/50b3dac3-8b5e-4142-b749-65ecf7b2904d"
     }

packages/mobile-voice/notes.md

@@ -1,4 +1,5 @@
 - While the model is loading for the first time, there should be some fun little like onboarding sequence that you can go through that makes sure the model is automated properly.
-- When a permission/session complete notification is sent, if you click on it, the session/server should auto be selected. 
+- When a permission/session complete notification is sent, if you click on it, the session/server should auto be selected.
 - We need some sort of permissions UI in the top half of the generation.
-- Need to figure out a good way to start new sessions.
+- Need to figure out a good way to start new sessions.
+- When an agent returns a generation, we should be able to expand it into a reader mode view.

packages/mobile-voice/src/app/_layout.tsx

@@ -10,6 +10,8 @@ import {
 LogBox.ignoreLogs([
   "RecordingNotificationManager is not implemented on iOS",
   "`transcribeRealtime` is deprecated, use `RealtimeTranscriber` instead",
+  "Parsed error meta:",
+  "Session activation failed",
 ])
 
 configureNotificationBehavior()