60db TTS provider integration

[Nitish-Kumar-Bote]

Adds 60db (TTS_PROVIDER=60db) as a TTS provider in packages/core, implemented alongside ElevenLabs
with the same factory pattern (createTTS()), per-agent voice maps, Prometheus metrics, and cost tracking

• synthesize() uses the HTTP endpoint POST https://api.60db.ai/tts-synthesize and decodes the base64 MP3
  envelope; synthesizeStream() drives the full 60db TTS WebSocket protocol (create_context → send_text → flush_context → audio_chunk* → flush_completed → close_context) and yields concatenatable LINEAR16 PCM chunks
  (24 kHz mono)
• checkHealth() probes GET /voices; estimateCost() = $0.00002/char; default voices: agent 0 → 60db system
  default, agent 1 → docs example voice (overridable via voiceMap)
• Wires '60db' into the Zod VoiceConfigSchema, VoiceConfig type, CLI (SIXTYDB_API_KEY env mapping),
  .env.example, and docs/providers.md
• Adds ws + @types/ws dependencies to packages/core