AI Agent Hub
Back to skills
Multi-Register Chinese AI Writing Humanizer icon

Multi-Register Chinese AI Writing Humanizer

Content Creation Updated 2026.08.30

Paste the following prompt into your AI chat to install this skill:

Please follow the guide at https://skillhub.cn/install/skillhub.md to install @user_8d277d62/humanizer-zh-skill into your AI assistant.

About this skill

Problem

Chinese AI text often fails less because of errors and more because of machine-like phrasing: translationese, formulaic parallelism, hollow corporate nouns, over-hedging, and injected opinion in news or official writing where it should not appear. Editors can sense the issue, but locating it quickly is harder.

How It Works

humanizer-zh-skill first classifies the text as blog, news, official, or formal essay, then applies the matching register rules. In rewriting mode, it reads the full text, lists core facts, diagnoses AI patterns, rewrites problem passages, and checks the quote, fact, and stance red lines. In review mode, it only reports problem locations, severity, and direction without changing the source. Its checklist covers overuse of , weak verbs plus abstract nouns, three-part stacking, and cliched filler.

Boundaries

It is useful for edit-level de-AI-ification of Chinese text, not for inventing content or replacing fact checking. If the register is mixed and unclear, it asks the user before rewriting. Internal scoring and pattern IDs are not meant to be shown to readers.

Use Cases

  • Before publishing a blog post, rewrite an LLM-drafted product essay to remove template phrasing while preserving key metrics and conclusions.
  • While drafting a press release, verify quotes, dates, and qualifiers remain intact and remove promotional filler.
  • Before submitting an internal or government report, clean up cliches while keeping the formal register from becoming colloquial.
  • Before writing a column, review whether unsourced judgments were added and break up three-part parallelism and hollow endings.

Best For

  • Content operators who manage blogs or newsletters and want LLM drafts turned into opinionated but fact-safe Chinese essays.
  • News editors who need pre-publication checks on quotes, facts, and promotional tone without rewriting every paragraph.
  • Policy or administrative writers who need AI-drafted reports to stay formal and standard, not colloquial.
  • Columnists who want to keep arguments intact while removing three-part parallelism, quotable lines, and hollow optimistic endings.