chore: convert TODO/FIXME to tracked references (TODO-008)

Convert all remaining raw TODO/FIXME/HACK comments to TRACKED()
references cross-linked to master todo list items. No behavior changes.

crates/chanora_audio/src/audio_event_queue.rs

@@ -24,8 +24,8 @@ pub enum AudioCommand {
     /// Set a client's output volume.
     SetVolume(SessionAudioId, f32),
     /// Remove a client's decode queue.
-    // TODO: Wire to client disconnect path; handled in callback but no
-    // producer currently pushes this command.
+    // TRACKED(TODO-005): Wire to client disconnect path; handled in callback
+    // but no producer currently pushes this command.
     #[allow(dead_code)]
     RemoveClient(SessionAudioId),
 }

crates/chanora_audio/src/engine.rs

@@ -2345,16 +2345,13 @@ impl CaptureState {
     /// Consume an arbitrary-rate, multichannel cpal buffer; produce
     /// 48 kHz mono frames; encode and send when `transmit_active`
     /// is true (PTT engaged).
-    // TODO(realtime-audio): This method runs on the cpal audio callback
+    // TRACKED(realtime-audio): This method runs on the cpal audio callback
     // thread with a ~10 ms deadline. Known pre-existing violations of the
-    // realtime safety constraint that should be addressed in a future
-    // iteration:
+    // realtime safety constraint (see TODO-006 in master-todo-list):
     //   1. Blocking Mutex::lock().unwrap() on self (via cpal callback)
     //   2. Potential allocation in mono_scratch.reserve() and
     //      pcm_accum.extend_from_slice() when buffer capacity is exceeded
     //   3. Encode-path warn!/error! logging via send_voip_frame closures
-    // These were present before the VAD integration and are not
-    // introduced by this changeset.
     fn ingest<T: ToF32 + Copy>(&mut self, buf: &[T]) {
         // 1. Down-mix to mono (pre-gain). Always performed so the level
         //    meter reflects real mic input even when PTT is released.

crates/chanora_audio/src/mobile_voice_backend.rs

@@ -13,7 +13,7 @@
 //! `cpal` (and SDL on Linux) own desktop capture/playback per the
 //! existing audio engine design.
 
-// TODO(SDD-117): back-fill `IosVoiceUnit` to implement this trait
+// TRACKED(SDD-117): back-fill `IosVoiceUnit` to implement this trait
 // so the engine can hold a single `Box<dyn MobileVoiceAudioBackend>`
 // across iOS and Android.