AGENT SKILL
voice
Pick the reading voice — list, audition, set
- License 许可
- MIT
- 最近更新
- 2026-09-16
- Stars
- ★ 14
- Forks
- 10
01 前言 Front Matter
- description
- Pick the reading voice — list, audition, set
- argument-hint
- [ "voice id to set directly" ]
03 SKILL.md 正文
If the user supplied an argument, treat it as a voice id: run step 3 with it.
- List what's available:
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/speak.py" --list-voices
Show the result as a short readable list (label and id). If it's long, show the first ~20 and say how many more there are; offer to filter by name.
- When the user wants to hear one, audition it WITHOUT saving:
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/speak.py" --detach --voice VOICE_ID --text "This is how I'd sound reading your replies."
Repeat for as many voices as they like.
- When they choose, save it and confirm in the new voice:
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/speak.py" --set-voice VOICE_ID
python3 "${CLAUDE_PLUGIN_ROOT}/scripts/speak.py" --detach --text "Voice saved. This is me from now on."
Voices belong to the current provider (see /read-aloud:speak-status). To
change provider, edit the config file it names — or /read-aloud:voice-setup
installs the free Kokoro voices.
04 同仓库更多 Skills
voice-setup
Install the free Kokoro neural voice (one-time ~340MB download; big upgrade on Linux)
★ 14
speak
Read Claude's last reply aloud
★ 14
speak-stop
Stop reading aloud
★ 14
speak-status
Show read-aloud configuration (provider, voice, auto-read)
★ 14
speak-auto
Toggle reading every reply aloud automatically
★ 14