You need to generate voice from text, clone voices, or produce podcast audio, and you want to work with ElevenLabs tooling directly in Claude. This handles TTS generation, speech-to-speech transformation, voice cloning from 30-second samples, and dialogue mixing with proper loudness targeting. The shibboleths table is genuinely useful: it calls out -16 LUFS for podcasts, 5-8kHz de-essing before compression, and the difference between eleven_multilingual_v2 for quality versus eleven_flash_v2_5 for speed. It won't handle spatial audio, music production, or sound effects work (those go to sound-engineer), but for voice-specific tasks like audiobook production or building conversational AI agents with proper vocal processing, it covers the workflow end to end.
npx skills add https://github.com/erichowens/some_claude_skills --skill voice-audio-engineer