feat(cli): text-to-speech v1 — select-model

[claude]

Adds the select-model recipe for text-to-speech v1 (cli).

Closes part of #176

[github-actions]

✅ CLI Recipe Tests

── recipes/cli/text-to-speech/v1/select-model ──
PASS
Generating speech with aura-2-arcas-en...
Audio saved to /tmp/deepgram_model_output.mp3 (17,424 bytes)
{
✓ PASSED