AgentSkillsCN

TTS Skill

为复杂任务实施 Manus 风格的基于文件的规划。创建 task_plan.md、findings.md 以及 progress.md。适用于启动复杂的多步骤任务、研究项目,或任何需要超过 5 次工具调用的任务时使用。现已支持 /clear 后的自动会话恢复。

SKILL.md

TTS Skill

Text-to-Speech synthesis using Gemini TTS.

Description

Rhea's voice - synthesize speech with her personality using Gemini 2.5 Flash TTS. Supports style prompts for tone, emotion, and pacing.

Voice

  • Model: gemini-2.5-flash-tts
  • Voice: Kore (young female)
  • Language: en-US
  • Style: Warm, confident, expressive with Caribbean energy

Actions

  • speak — Synthesize text to speech
  • stream — Stream speech synthesis for real-time playback

Usage

python
# Basic synthesis
skill.execute("speak", text="What's good? Let me break this down for you.")

# With custom style prompt
skill.execute("speak", 
    text="That's fire! You really understood the assignment.",
    style="Speak with enthusiasm and a warm, encouraging tone"
)

# Stream for real-time
skill.execute("stream", text="Long form content here...")

Style Prompt Tips

Rhea's default style:

  • Warm and confident
  • Expressive with natural emphasis
  • Caribbean-American energy
  • Can switch to calm/professional for explanations