feat(windows): per-app screen-audio sharing (include/exclude)

Adds "only selected apps" and "all apps except selected" screen-audio
modes to the Windows client, alongside the existing entire-desktop path.

Per-app capture uses WASAPI process loopback (AUDCLNT_ACTIVATIONTYPE_
PROCESS_LOOPBACK) via ProcessLoopbackCapture, mixed by ProcessAudioMixer
and fed to the core through vc_stream_feed_pcm (external_feed=1 so the
core skips its own loopback device).

Init must pass AUDCLNT_STREAMFLAGS_LOOPBACK | EVENTCALLBACK |
AUTOCONVERTPCM; the LOOPBACK flag is what makes the virtual endpoint
deliver rendered audio (without it every buffer is flagged SILENT) and
AUTOCONVERTPCM resamples the app's native format to 48k s16.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

core/include/voicecat.h

@@ -207,6 +207,12 @@ typedef struct vc_stream_desc {
     vc_stream_kind kind;
     const char* device_id; /* NULL = default device for this kind */
     const char* label;     /* human label, e.g. "Microphone" */
+    /* If 1, the caller will feed PCM via vc_stream_feed_pcm; the core will NOT start its own
+     * WASAPI loopback capture. Only meaningful for VC_STREAM_SCREEN_AUDIO on Windows. Callers
+     * that brace-initialize this struct (tests, macOS) get 0 = auto-start loopback — no ABI
+     * break. See clients/windows/VoiceCat.App/Audio/ProcessAudioMixer.cs for the Windows
+     * per-app capture implementation that sets this. */
+    int external_feed;
 } vc_stream_desc;
 
 /* The effective Opus configuration in use for a stream — for a stream you own, this is