humanize-chinese

What This Skill Does

The humanize-chinese skill is a powerful command-line interface tool designed to bridge the gap between machine-generated prose and natural human expression. At its core, the tool employs a sophisticated detection engine capable of identifying over 20 distinct AI-writing patterns—ranging from rigid three-part structures and mechanical connectors to the overuse of corporate buzzwords. By assigning a weighted 0-100 severity score, it allows users to quantify exactly how 'robotic' their text appears.

Beyond detection, the skill features a comprehensive transformation engine. It can automatically restructure sentences, diversify vocabulary, and apply one of seven specific stylistic transforms (including Zhihu, Xiaohongshu, Weibo, and academic tones) to ensure the final output feels authentic. Using a pure Python, zero-dependency architecture, v2.0.0 provides a lightweight and robust solution for content creators, copywriters, and developers seeking to refine AI-assisted content without external API dependencies.

Installation

To integrate this skill into your environment, use the OpenClaw repository manager:

clawhub install openclaw/skills/skills/0xspeter/humanize-chinese-2-0-0

Ensure your local environment has a standard Python 3.x installation, as the tool relies on native libraries for all text-processing logic.

Use Cases

1. Content Marketing: Rapidly convert AI-drafted articles into engaging social media posts for platforms like Xiaohongshu or WeChat.
2. Professional Editing: Identify and strip out excessive 'AI-isms' like empty corporate jargon from business reports.
3. Academic Refinement: Adjust the tone of research summaries to match specific literary or academic standards while maintaining core data integrity.
4. Personalized Communication: Adapt robotic, generated email responses into more casual, human-centric language for better engagement.

Example Prompts

1. 'Run a detection analysis on blog_post.txt and tell me which sentences are the most robotic so I can rewrite them.'
2. 'Humanize the content in marketing_copy.txt, apply the xiaohongshu style, and save the result as social_final.txt.'
3. 'Check the current AI-generated draft in draft.txt and compare it against a more casual tone, then show me the improvement in a side-by-side report.'

Tips & Limitations

• Weighting: The detection system categorizes patterns into Critical, High, Medium, and Style signals. Prioritize fixing 'Critical' signals (like mechanical connectors) first, as they contribute most significantly to the 0-100 score.
• Aggressive Mode: Use the -a (aggressive) flag for content that sounds overly formal; however, review the results carefully to ensure meaning remains intact.
• Consistency: If working on a long-form document, process it in sections rather than as a single massive file to maintain better context-awareness.
• Limitations: The tool is optimized for Chinese text. While effective, it cannot replicate unique human experiences or factual domain knowledge; it focuses primarily on stylistic flow and linguistic structure.