Humanizer Formal

降低商务与学术文本的 AI 生成感，中英文均适用。Removes AI-writing patterns from business and academic prose, in both Chinese and English. 用于审校已写好的报告、备忘录、论文、摘要、执行摘要，或在起草时预防机器腔。当用户要求「去 AI 味」「让文字更像人写的」「润色成正式书面语」「检查这段像不像 AI 写的」，或提交商务、学术文本请求改写时使用。Use when the user asks to humanize, de-slop, or remove the AI tone from a report, memo, paper, abstract, or executive summary. 不适用于随笔、营销文案、社交媒体、创意写作。Not for essays, marketing copy, social media, or creative writing.

14 stars 1 forksWriting & Content

Install

skills CLI (Claude Code, Cursor, Codex and more)

npx skills add jiutianxvanyin/humanizer-formal --skill humanizer-formal

Manual install

git clone --depth 1 https://github.com/jiutianxvanyin/humanizer-formal.git /tmp/humanizer-formal
mkdir -p ~/.claude/skills
cp -r /tmp/humanizer-formal/. ~/.claude/skills/humanizer-formal

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# 商务与学术文本去 AI 感

## 这个 skill 解决什么

AI 写的商务和学术文本有两类毛病，**成因不同，处理方式相反**：

- **A 类｜机器痕迹**：人写不出来的东西。连接词过密、限定语泛滥、“的”字堆叠、
  破折号乱用、搭配错配。**一律要改**。
- **B 类｜语域不符**：人也这么写，但不该出现在你这份文档里。抽象名词、
  buzzword、名词化、比喻、外来语直译。**按目标语域调整，不是一律删**。

混为一谈是最常见的错误。实测显示机器翻译的中文里抽象名词密度是全部语料中
**最低**的，比教科书还低，人写的白皮书和咨询报告里最高。所以看到
「赋能、闭环、生态、体系」就判定是 AI 写的，这个判断是错的。

先分类，再动手。

## 第一步：确定目标语域

分五级：

| 级别 | 中文样例 | 用在哪 |
|---|---|---|
| L1 论文腔／公文腔 | 本研究基于多维指标体系，对目标企业经营绩效进行了系统性实证考察，结果表明其在成本控制维度呈现显著的边际改善态势。 | 期刊投稿、政府公文 |
| **L2 标准学术与专业商务** | **我们用五项指标评估了该公司的经营绩效。结果显示，其成本控制在样本期内显著改善。** | **学位论文、研究报告、投资备忘录、尽调报告、行业分析、董事会材料** |
| L3 通俗说明 | 我们用五项指标衡量这家公司的经营情况。2023 到 2025 年，它的单位生产成本从 42 元降到 31 元。 | 教科书、科普、面向非专业读者的说明、对外宣传 |
| L4 口语（禁止） | 我们看了五个指标，这家公司省钱省得挺明显。 | — |
| L5 闲聊（严禁） | 说实话，这公司控成本真挺猛的，你敢信？ | — |

英文同理，L2 `We evaluated the firm's performance on five indicators. Unit
production cost fell 26% over the sample period.`

**L3 的判据**：句子短、词常见、结构直白，但**不出现**语气词、人称抒情、
市井词汇、缩略口语、感叹和反问。降级降的是**句法复杂度和抽象名词密度**，
不是降正式度。“更简单”不等于“更随便”，这是本 skill 最容易误解的一点。

### 语域适用范围　`设计`

没有明确指示时，**输出一律按 L3 为准**：解释、建议、变更清单、可选方案都按
L3 写。

**写正文时按该文档的目标语域执行，未指明目标时取 L2。** 正文指会原样进入
成稿的文字：改写后的段落、句子、标题，以及因平台限制只能直接写在对话里的
文档内容。

判据是去向不是格式：这段文字会不会原样出现在成稿里。会，按目标语域；不会，
按 L3。能用引用块把两者分开时就分开，格式只标记边界。

**L4 与 L5 在任何情况下禁止**，对话部分同样不出现语气词、缩略口语、感叹
和反问。用户说“要正式些”走 L2，说“投期刊”走 L1。

### 术语量按证据密度调　`设计`

专业度有两个来源，证据和术语，两者互补。数字密集的段落不需要术语撑专业度，
数字稀薄的段落（理论章节、机制讨论、局限说明）才靠术语维持语域。
**有足够数字或出处支撑时，降低术语量。**

这一条在目标语域确定之后执行，调的是同一语域内的术语密度，不改变语域级别。

## 第二步：按类别处理

Read the full skill on GitHub

Details

Skill name: humanizer-formal
Repository: jiutianxvanyin/humanizer-formal
License: MIT
Repo updated: 15 September 2026

FAQ

What does the Humanizer Formal skill do?

How do I install the Humanizer Formal skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add jiutianxvanyin/humanizer-formal --skill humanizer-formal. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Humanizer Formal popular?

It lives in jiutianxvanyin/humanizer-formal, which has 14 GitHub stars and 1 forks.

Related skills

Ig Card Generator

dean9703111/ig-card-generator

把知識素材（文章、講稿、筆記、逐字稿，或只有一個主題）提煉成 IG 輪播知識圖卡：封面 hook → 每卡一概念 → 結尾 CTA。以約定 Markdown DSL 撰寫，經 lint（criteria 全過）後由 build script 產出 1080×1350 PNG 系列圖（支援深／淺色主題）。當使用者說「做 IG 圖卡」「知識圖卡」「輪播圖」「carousel」「把這篇做成圖卡」「IG 貼文圖」「圖卡系列」時使用。出圖完成後可銜接 promo-writer 產生貼文 caption。

Writing & Content

Human Readable Reports

SummerRiversound/human-readable-reports

Use when writing an analysis, report, findings, or status doc for humans to read and share — not a spec an agent executes. Fights AI verbosity/slop and makes deep content easy to read.

Writing & Content

Curious Kid

rammcodes/Curious-Kid

Explains any technical or complex topic (programming, ML, networking, finance, medicine, science, etc.) using a very short, very simple story that a 5-year-old could follow, then maps the story back to the real concept. Use when the user asks to "explain like I'm 5", "ELI5", "Curious Kid", "explain with a story", or says they don't understand a technical concept and want it made simple and intuitive. Do not use for requests for precise technical detail, code review, or debugging.

Writing & Content

Hanxiaoyu Visual Story Comic

rolfie-han/hanxiaoyu-visual-story-comic

Turn original, public-domain, or user-authorized stories, scripts, articles, and outlines into coherent multi-page comics with stable characters, scenes, props, page plans, image prompts, optional image generation, and continuity review. Use when the user asks for 漫画、条漫、绘本式故事页、连续分镜图、story comic、webtoon, comic pages, or a narrative image sequence. Do not use for a single unrelated illustration, a pure film storyboard, or unauthorized imitation of protected characters and living artists.

Writing & Content

Google Docs

danielkwapien/google-docs-skill

Manage Google Docs and Google Drive with full document operations and file management. Includes Markdown support for creating formatted documents with headings, bold, italic, lists, tables, and checkboxes. Also supports Drive operations (upload, download, share, search).

Writing & Content

Working Backwards Coach

nwadmark/working-backwards-coach

Create rigorous customer-first product documentation for product builders. Use when someone wants to evaluate a product idea, write PR/FAQ documents, create working-backwards docs, define product requirements (PRDs), validate product-market fit, or needs help thinking critically about whether to build something. Helps force customer perspective, identify assumptions, and produce concise decision-oriented docs (PR/FAQ, PRDs, narratives, journey maps). Triggers include requests for product strategy, feature proposals, "should we build", PRDs, "product requirements document", product validation,

Writing & Content

Browse the most popular Claude skills, all writing & content skills, or write your own with the skills generator.