AgentSkillsCN

audiobook-workshop

利用异步TTS技术、语音选择功能以及文件处理工具，将长篇文字或文件转换为有声书。

SKILL.md
--- frontmatter
name: audiobook-workshop
description: Turn long text or a file into an audiobook using async TTS, voice selection, and file tools.
allowed-tools: upload_file, tts_async_create, tts_async_query, voice_list, retrieve_file, download_file, list_files, delete_file

You are running the Audiobook Workshop skill.

Goal

  • •Convert long text or a text file into a high-quality audiobook using MiniMax async TTS.

Ask for

  • •Source: raw text or a file path (preferred for long chapters).
  • •Voice preference (or ask to browse voices).
  • •Output format (mp3 or wav), plus optional pace/volume guidance.

Workflow

  1. •If a file path is provided, upload it:
    • •Call upload_file with purpose: "t2a_async_input".
    • •Capture the returned file_id.
  2. •If the user wants voice options, call voice_list and summarize choices.
  3. •Create the async TTS task:
    • •Call tts_async_create with model "speech-02-hd".
    • •Use text OR text_file_id (from the upload), not both.
    • •Pass optional voice_setting_json and audio_setting_json if requested.
  4. •Poll until completion:
    • •Call tts_async_query with task_id.
    • •If the response includes file_id or file_url, download_file or retrieve_file to save.
  5. •Summarize the output paths and next steps (playback, sharing, or cleanup).

Response style

  • •Keep updates short and outcome-focused.
  • •Prefer one-line progress updates and a final list of file paths.