feat(dross): voice Phase 2a — local whisper transcribe + mic (2.12.0)

faster-whisper (small.en, GPU+CPU fallback) on CT 102 → POST
/api/voice/transcribe (multer→whisper client) → mic in the bubble
records (MediaRecorder), uploads, drops the transcript into the input
to review-and-send. Infra scripts in deploy/whisper/. Retention (P2b)
next. NOTE: mic needs a secure context (the https domain), not the LAN IP.

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

public/style.css

@@ -765,6 +765,8 @@ body.drawer-open #scrim { opacity: 1; pointer-events: auto; }
 .dross-btnrow{display:flex;gap:10px}
 .dross-mic{flex:1;height:50px;border-radius:12px;border:1px solid var(--dross-dim);background:var(--dross-soft);color:var(--dross-glow);cursor:pointer;display:flex;align-items:center;justify-content:center;gap:9px;font-family:var(--font-ui);font-size:13px}
 .dross-mic[disabled]{opacity:.5;cursor:not-allowed}
+.dross-mic.rec{background:#3a1010;border-color:var(--accent);color:#fff;animation:dross-rec 1.2s infinite}
+@keyframes dross-rec{0%,100%{box-shadow:0 0 0 0 rgba(255,79,46,.5)}50%{box-shadow:0 0 0 8px rgba(255,79,46,0)}}
 .dross-send{width:64px;height:50px;border-radius:12px;border:1px solid var(--dross-dim);background:linear-gradient(180deg,var(--dross),var(--dross-dim));color:#fff;cursor:pointer;display:grid;place-items:center}
 .dross-collapse{display:flex;align-items:center;justify-content:center;gap:8px;height:34px;cursor:pointer;color:var(--muted);
   font-family:var(--font-ui);font-size:11px;letter-spacing:.12em;text-transform:uppercase;background:#0b0810;border-top:1px solid var(--border)}