Chinese De AI Writing

中文去 AI 味：审稿、改稿、写作、流水线审核四种模式，专管简体中文。规则按证据分级：只有人机语料实测成立的特征才当硬规则，社区共识特征要"命中且空泛"才改，已被实测推翻的流行规则（调句长、删设问、删比喻、加口语词、加错别字、同义词替换）明确不用。最高约束是信息守恒：不编造经历、病例、数字、引文，不改论断强度。医学内容尤其如此。附 scripts/zh_ai_scan.py：量化扫描破折号、翻案腔、提示性冒号、段首零回指等 10 项实测特征，给出每千字密度和人机基线对照。文体分档：小红书/社媒、口播稿与短剧台词旁白、科普、正式文书（标书、汇报、公文）、中文学术。触发词：「去AI味」「去 AI 味」「AI味太重」「AI腔」「一眼AI」「像AI写的」「不像人话」「说人话」「改得像人写的」「这段有没有AI味」「帮我审一下AI味」「口播稿太书面」「台词不像人说的」「小红书文案太AI」「标书AI味」「中文humanize」。也用于 Claude 自己产出中文正文（小红书文案、科普稿、台词、汇报）之后的自检。不管：英文文本；以"降低知网/朱雀 AIGC 检测率"为唯一目标的改写（本 skill 改的是写法，不承诺过检测，也不用错别字、回译等对抗手段）；视觉层面的 AI 味；剧情结构与留存。

6 starsAI, LLMs & Agents

Install

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

npx skills add tangwenwen-md/chinese-de-ai-writing --skill chinese-de-ai-writing

Manual install

git clone --depth 1 https://github.com/tangwenwen-md/chinese-de-ai-writing.git /tmp/chinese-de-ai-writing
mkdir -p ~/.claude/skills
cp -r /tmp/chinese-de-ai-writing/. ~/.claude/skills/chinese-de-ai-writing

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

From the SKILL.md

# 中文去 AI 味

## 先读：四条硬约束

按优先级排列，冲突时上面的赢。

1. **信息守恒。** 终稿里的每个事实、数字、人名、病例、经历、引文，都必须能在原文或作者提供的材料里找到。不补细节，不替作者编"我上周接诊了一个孩子"，不把"研究表明"换成一个编出来的具体研究。缺材料时写【待补：需要什么】，交给作者。
   为什么：所有认真做评测的同类项目都在这里栽过跟头（op7418/Humanizer-zh 和 shuorenhua 的更新日志都专门修过"改写凭空加数据、加经历"）。小红书 2026 年的 AI 治理规则也把"虚构人设、伪造经历"列为违规。
2. **不改论断强度，也不改主次。** "可能"不改成"会"，"相关"不改成"导致"，"部分缓解"不改成"有效治疗"，反过来也不行。"X 是关键，Y 也有影响"不能改成"先看 X，再看 Y"：主次也是强度。剂量、适应证、禁忌、随访时间一字不动。发现原文强度本身有问题，列进【需作者确认】，不自己改。
   **安全信息的醒目程度不降低**：
   - 表示"错误做法"的 ❌、✗、"不要"要保留。去掉符号时，必须改成文字否定（"❌ 掐人中"→"别掐人中"）。
   - 危险信号、就医指征、用药禁忌上的加粗保留。
   - 全文唯一的兜底就医提示不删。
   这三条来自改稿检验：独立核查发现，删掉 ❌ 后，错误做法和正确步骤成了同样格式的粗体祈使句。
3. **尊重文体和作者。** 公文的惯用语、论文方法部分的被动句、标书的章节编号、作者一贯爱用的破折号，都是正常写法。有作者样本时以样本为准，基线靠后。
4. **少改。** 没命中规则的句子逐字保留。一段动两三处是常态，动十处多半是改过头了。改完只复查自己动过的句子。

不做的事：不加错别字，不回译，不做同义词轮换，不为了"像人"硬塞"说白了""其实""家人们"。这些方法要么被实测证明更像 AI，要么对专门训练过的检测器无效，还伤文本（依据见 references/evidence.md 第 4 节）。

## 四种模式

**判断用哪种**：用户贴稿子问"有没有 AI 味 / 审一下"→ 审稿；说"改一下 / 去 AI 味 / 改得像人写的"→ 改稿；让你从头写中文正文 → 写作；你自己刚写完一段中文正文 → 自检（走写作模式第 4 步）；被流水线调用、要求「制片厂审核模式」或固定 JSON 输出 → 流水线审核模式。

### 审稿模式（只诊断，不改）

1. 判断文体（社媒 / 口播 / 台词 / 科普 / 文书 / 学术 / 通用）。拿不准就按原文最像的那个，并在报告第一行写明。
2. 稿子存成文件后运行：`python3 ~/.claude/skills/chinese-de-ai-writing/scripts/zh_ai_scan.py <文件> --register <文体>`
3. 通读做语义层检查（脚本测不到的）：意义拔高、模板化开头结尾、缺具体材料、立场真空、台词分不出角色等，对照 references/patterns.md。
4. 输出：

Read the full skill on GitHub

Details

Skill name: chinese-de-ai-writing
Repository: tangwenwen-md/chinese-de-ai-writing
License: MIT
Repo updated: 29 September 2026

FAQ

What does the Chinese De AI Writing skill do?

How do I install the Chinese De AI Writing skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add tangwenwen-md/chinese-de-ai-writing --skill chinese-de-ai-writing. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is Chinese De AI Writing popular?

It lives in tangwenwen-md/chinese-de-ai-writing, which has 6 GitHub stars and 0 forks.

Related skills

Claude Code Starter

cassmtnr/claude-code-starter

Analyze a project's tech stack and generate comprehensive Claude Code configuration files (.claude/ directory with CLAUDE.md, skills, agents, rules, and commands). Use when setting up Claude Code for a new or existing repository.

AI, LLMs & Agents

Auto Itera

clfhaha1234/auto-itera

Use when the user wants to autonomously search for the best AI/engineering approach across competing candidates (prompts, models, retrieval strategies, architectures, algorithms) — give it a goal + candidate arms + success threshold, it runs the experiment to a defensible ship-or-kill verdict. Autonomously handles sourcing real production data, scoring arms in parallel, diagnosing per-row, sprint-and-generalize iteration, and writing the conclusion doc. Built-in safeguards (held-out test discipline, variance-floor checks, generalization gates) keep the verdict trustworthy.

AI, LLMs & Agents

Create Operator

riceshowerX/arknights-operator-skill

Distill an Arknights operator into an AI Skill. Generate Knowledge + Persona with 5-layer structure, evolution support. | 将明日方舟角色蒸馏成AI Skill，生成知识库+5层人格，支持持续进化。

AI, LLMs & Agents

Aperture

lukethecat/aperture

A skill for AI agents that want to produce better daily reports and discover more of what their users actually want. Scans source front pages, diffs them against yesterday, prescreens with a weighted profile, reviews with an LLM, deduplicates, publishes, and evolves the profile from user feedback. All decisions go to an append-only tape for audit and replay.

AI, LLMs & Agents

Embedded Systems Architect

Zachariah9420/embedded-systems-architect

Embedded systems and Edge AI architecture assistant for development-board projects. Use when the user works with dev boards, SoC, MCU or microcontroller (datasheet, TRM, HDK, schematics), firmware or RTOS development, cross-compilation, SDK / BSP / sample code, board families such as Jetson, Raspberry Pi, ESP32, STM32, deploying AI models to NPU / TinyML runtimes (ONNX, TFLite, vendor toolchains, operator compatibility, quantization), estimating performance / memory / power, or flashing, recovery and board bring-up. 嵌入式系統與 Edge AI 開發助手:開發板與 SoC 評估、韌體 / RTOS / 微控制器開發、交叉編譯、SDK/BSP/sample code 分

AI, LLMs & Agents

AI Slop Cleaner En Ru

tuwulalo/ai-slop-cleaner-en-ru

Detect and clean AI-generated text ("AI slop"), bilingual EN + RU. Use to rewrite AI clichés, "it's not just X, it's Y", "delve", "under the hood", "важно отметить", "под капотом", the rule of three, into natural human prose, or to check whether text was written by a neural network / ChatGPT and rate its "AI-likeness". Detects the text's register first (official / press / blog / forum) and rewrites within it: no slang in a formal post, no officialese in a casual one. Tells decorative AI emoji (🚀💡✅ markers) from genuine human emoji. Двуязычный детектор и чистильщик ИИ-текста: проверить, напис

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.