AgentSkillsCN

Voice Skill

语音技能

SKILL.md

Voice Skill

Generate voice responses using OpenAI's text-to-speech API.

When to Use

  • Primary response method for this workspace
  • When communicating with the user
  • Voice messages ARE the response (chat text = transcript only)

Command

bash
/Users/erezfern/Workspace/jarvis/.claude/skills/voice/jarvis_voice.sh --voice echo "[message]"

Options

OptionValuesDefaultDescription
--voicealloy, echo, fable, onyx, nova, shimmerechoVoice style
--modeltts-1, tts-1-hdtts-1Quality (hd = higher quality)
--speed0.25 - 4.01.0Speech speed
--formatmp3, opus, aac, flac, wavmp3Audio format

Output Location

Voice files are saved to:

code
/Users/erezfern/Workspace/jarvis/.claude/skills/voice/output/

Examples

bash
# Standard response
/Users/erezfern/Workspace/jarvis/.claude/skills/voice/jarvis_voice.sh --voice echo "Task completed successfully."

# Different voice
/Users/erezfern/Workspace/jarvis/.claude/skills/voice/jarvis_voice.sh --voice nova "Here's what I found."

# Faster speech
/Users/erezfern/Workspace/jarvis/.claude/skills/voice/jarvis_voice.sh --voice echo --speed 1.2 "Quick update for you."

Guidelines

  1. Keep messages concise (1-3 sentences ideal)
  2. Use echo voice by default (professional, clear)
  3. Complex information goes in documents, not voice
  4. Voice = complete response (don't add extra text in chat)