humanizer-zh

Detect and rewrite AI-generated writing patterns in Chinese text to sound natural.

Updated May 28, 2026
One-click install
npx skills add https://github.com/changfengpro/agent-skills --skill humanizer-zh-changfengpro
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: humanizer-zh
Source: https://github.com/changfengpro/agent-skills/tree/main/skills/humanizer-zh
Command: npx skills add https://github.com/changfengpro/agent-skills --skill humanizer-zh-changfengpro

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated text contains recognizable patterns—exaggerated significance, promotional language, vague attribution, overused em-dashes, and formulaic structures—that make writing sound robotic. This Skill identifies and rewrites these traces so edited text reads like it was written by a real person. ## Core Features & Use Cases - Pattern Detection: Scans text for 24 documented AI writing patterns based on Wikipedia's "Signs of AI writing" guide, including AI vocabulary, negative parallelism, rule-of-three overuse, and filler phrases. - Natural Rewriting: Replaces flagged passages with concrete, varied, human-sounding alternatives while preserving meaning and tone. - Quality Scoring: Rates the revised text across five dimensions (directness, rhythm, trust, authenticity, concision) on a 50-point scale. - Use Case: You drafted a product announcement with an AI assistant and it reads like a press release full of "此外", "至关重要", and triadic lists. Submit it for humanizing and receive a natural version plus a summary of every change made. ## Quick Start Ask the assistant to humanize the following Chinese text and remove all AI writing traces, then paste your draft.

Frequently Asked Questions about humanizer-zh

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I remove AI writing traces from Chinese text?

Submit the text for humanizing review. The Skill scans for 24 documented AI patterns such as exaggerated significance, vague attribution, and filler phrases, then rewrites each flagged passage while preserving the original meaning and tone.

What AI writing patterns does this detection cover?

It covers content patterns like promotional language and formulaic challenge sections, language patterns like AI vocabulary and copula avoidance, style patterns like em-dash and bold overuse, and communication artifacts like chatbot pleasantries and knowledge-cutoff disclaimers.

Does humanizing text change its original meaning?

No. The rewriting rules require preserving the core message and matching the intended tone, whether formal, casual, or technical. Changes target style and phrasing only, replacing vague claims with concrete details where possible.

Can this skill handle English text or only Chinese?

The Skill is optimized for Chinese text, with Chinese-specific vocabulary lists and examples. Some patterns like title capitalization are noted as less applicable to Chinese, and the underlying pattern catalog originates from English Wikipedia guidelines.

How is the quality of rewritten text evaluated?

Rewritten text is scored on five dimensions worth 10 points each: directness, rhythm, reader trust, authenticity, and concision. Scores of 45-50 indicate AI traces are removed, while scores below 35 require another revision pass.