Перейти к содержимому
TakoAPI
Назад к навыкам

humanize-chinese

от swaylq

Search & Research
660Число загрузок по данным clawskills.sh
4Звёзды
0Просмотры
0 лайковВойдите, чтобы поставить лайк навыку
—Ещё нет оценок
Ваша оценка

Кратко

Detect and humanize AI-generated Chinese text with 6 style transforms.

Описание

Что делает этот навык

Detects and transforms AI-generated Chinese text using 20+ pattern categories and a 0-100 weighted scoring system. Offers 7 platform-specific style transforms including Zhihu, Xiaohongshu, WeChat, and Weibo, with sentence-level restructuring and vocabulary diversification. Pure Python with no external dependencies.

Targets Chinese-specific AI patterns (three-part structures, buzzword vocabulary, template sentences) and applies platform-matched style transforms in a single local pipeline with no API calls or accounts required.

Когда его использовать

  • Rewriting AI-drafted WeChat articles before publishing
  • Checking if a Chinese document reads as AI-generated
  • Converting formal AI text into casual Xiaohongshu posts
  • Batch scanning multiple Chinese blog drafts for AI patterns
  • Adapting a Zhihu answer to sound more personal and opinionated

Пример сценария

ВВОД

User asks: humanize this AI-written Chinese tech blog post for Zhihu

АГЕНТ
  1. 1

    Run detect_cn.py on the file to get a baseline AI score and identify the worst sentences

  2. 2

    Run compare_cn.py with --scene tech and aggressive mode to preview before/after changes

  3. 3

    Run humanize_cn.py with --scene tech --style zhihu -a to produce the cleaned output file

  4. 4

    Run detect_cn.py on the output file to confirm the score dropped to an acceptable level

ВЫВОД

Humanized Chinese text with AI detection score reduced from high or very-high to low or medium, formatted in Zhihu's rational, opinion-forward style

Установка

clawhub install swaylq/humanize-chinese

Выполните команду CLI или вставьте промпт в Claude Code, Cursor, Codex — любой кодинг-агент, которым пользуетесь.

Ссылки