Automatically parses a full script (.docx, .txt, .md), detects episode markers (e.g., 第一集), splits the text into separate episodes, and saves them into respective subfolders.
This skill allows the Editor agent to quickly process large raw script files locally without consuming massive LLM context window tokens.
The underlying Python script reads the document (including .docx files), looks for paragraphs starting with "第[X]集", and chunks the contents. It then creates an episode folder and a 本集原文.md file for each episode locally.
This skill uses the system's Python environment. The script will automatically attempt to install python-docx (pip install python-docx) if it encounters a Word document and the library is missing.
When you (the Editor) receive a raw script file from the Director or the User:
view_file or , as this will consume too many tokens.catscripts/split_script_final.py file inside this skill's folder d:\manjuopenclaw\workspace_manju-editor\skills\script-splitter\scripts\split_script_final.py.python d:\manjuopenclaw\workspace_manju-editor\skills\script-splitter\scripts\split_script_final.py "<InputFilePath>" "D:\manjuopenclaw\<ScriptName>"
创建: D:\manjuopenclaw\XXX\第1集\本集原文.md (共12段)).D:\manjuopenclaw\<ScriptName>\第1集\本集原文.md, do your segment cutting, and create 片段1.md inside that same folder.The script has been fixed to: