feat(stt-log): log the WHOLE turn pipeline to the transcript channel

The transcript channel only showed successful transcripts, so dropped utterances
(the 47/50 misses) were invisible. Now every captured utterance is mirrored with
its outcome and per-stage timing:
- too-short blip (<300ms), VAD "음성 아님(VAD 차단)", "인식 실패", "답변 없음", or "ok"
- transcript + reply (or "(무응답)")
- ⏱️ stt/llm seconds

The bridge meta now carries note + stt_sec + think_sec; voice.ts fires onTurn for
every turn (not only non-empty transcripts) and for the too-short drop; userbot
formats the diagnostic line.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

bot/src/voice.ts

@@ -69,8 +69,17 @@ export class VoiceSession {
   /** Pending reply clips. Played one at a time so concurrent speakers don't
    *  cut each other's replies off. */
   private playQueue: Buffer[] = [];
-  /** Optional callback to surface transcripts/replies to a text channel. */
-  onTurn?: (info: { user: string; transcript: string; reply: string }) => void;
+  /** Optional callback to surface EVERY heard utterance (and its outcome) to a
+   *  text channel — including ones dropped before/at STT — so misses are
+   *  diagnosable. `note` says why (e.g. "음성 아님(VAD 차단)", "너무 짧음", "ok"). */
+  onTurn?: (info: {
+    user: string;
+    transcript: string;
+    reply: string;
+    note?: string;
+    sttSec?: number;
+    thinkSec?: number;
+  }) => void;
   /** Live screen-share state, sent with each turn so the brain routes search
    *  (Chrome while broadcasting, Gemini when off). */
   getBroadcasting?: () => boolean;
@@ -123,8 +132,12 @@ export class VoiceSession {
 
     if (!chunks.length) return;
     const mono = stereoToMono(Buffer.concat(chunks));
-    // Ignore blips shorter than ~300ms (likely noise / key clicks).
-    if (mono.length < DISCORD_RATE * 0.3 * 2) return;
+    // Ignore blips shorter than ~300ms (likely noise / key clicks) — but still
+    // report them so the transcript channel shows every captured utterance.
+    if (mono.length < DISCORD_RATE * 0.3 * 2) {
+      this.onTurn?.({ user: userId, transcript: "", reply: "", note: "너무 짧음(<300ms)" });
+      return;
+    }
     const wav = pcm16MonoToWav(mono, DISCORD_RATE);
 
     try {
@@ -133,9 +146,16 @@ export class VoiceSession {
       // so the first sentence starts playing while the rest are still spoken.
       await converseStream(wav, this.getBroadcasting?.(), {
         onMeta: async (meta) => {
-          if (meta.transcript) {
-            this.onTurn?.({ user: userId, transcript: meta.transcript, reply: meta.reply });
-          }
+          // Report EVERY turn (even empty/VAD-dropped) so the transcript channel
+          // explains why a turn did or didn't answer.
+          this.onTurn?.({
+            user: userId,
+            transcript: meta.transcript,
+            reply: meta.reply,
+            note: meta.note,
+            sttSec: meta.stt_sec,
+            thinkSec: meta.think_sec,
+          });
           // Apply any broadcast directive the brain requested (e.g. user said
           // "방송 켜줘 / 꺼줘") before the reply audio plays. The meta line
           // always precedes the audio clips, so awaiting here preserves order.