feat: migrate on-device Qwen3 ASR to CoreML for background Flow dictation

Replace MLX GPU inference with CoreML bundles so transcription continues
while the host app is backgrounded. Adds model download and warm-up,
vendored Qwen3Speech, and updates onboarding, settings, and copy for the
~1.6 GB CoreML package (iOS 18+).

OSGKeyboardShared/Services/FlowContinuousCapture.swift

@@ -34,7 +34,7 @@ private final class FlowCaptureStreamRelay: @unchecked Sendable {
     }
 
     func yield(_ snapshot: AudioBufferSnapshot) {
-        lock.withLock { continuation?.yield(snapshot) }
+        _ = lock.withLock { continuation?.yield(snapshot) }
     }
 
     func finish() {
@@ -250,6 +250,22 @@ public final class FlowContinuousCapture {
         )
     }
 
+    /// Re-activate capture after returning from background without
+    /// reinstalling the tap (iOS may deactivate the audio session).
+    public func reassertIfRunning() {
+        guard isRunning else { return }
+        let session = AVAudioSession.sharedInstance()
+        try? session.setCategory(
+            .playAndRecord,
+            mode: .measurement,
+            options: [.defaultToSpeaker, .allowBluetoothHFP, .mixWithOthers]
+        )
+        try? session.setActive(true, options: .notifyOthersOnDeactivation)
+        if !audioEngine.isRunning {
+            try? audioEngine.start()
+        }
+    }
+
     /// Begin forwarding downsampled buffers to ASR for one utterance.
     public func beginUtterance() -> AsyncStream<AudioBufferSnapshot> {
         let (stream, continuation) = AsyncStream<AudioBufferSnapshot>.makeStream()