Chinese Tech Doc Style skill
A skill for Chinese technical writing: rewrites docs, interface strings and API descriptions, keeps the facts intact and checks output with a linter
- 1,518 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill changes only the text you give it and runs nothing; the linter reads files and writes nothing
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/chinese-tech-doc-styleDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill changes only the text you give it and runs nothing; the linter reads files and writes nothing
This is third-party code. Review the repository files before installing.
What it does
Chinese Tech Doc Style is a skill for writing, rewriting and proofreading Chinese technical text: documentation home pages, API descriptions, installation guides, troubleshooting, interface strings and buttons. Its rules aim at restrained, accurate, readable text without a sales tone, and the recipes strip empty phrases, repetition and internet jargon such as 赋能, 抓手, 闭环 and 打通. The skill puts facts above style. It does not add dates, figures, deadlines or conditions that were not there, and it does not remove limits, warnings or degrees of certainty. Code, JSON keys, URLs, API paths and database fields stay untouched. Among the rules: corner quotes 「」, spacing between Chinese text, Latin letters and digits, one paragraph per source line and careful rendering of English statuses such as Success, Invalid and Bad Request. For instructions, diagnostics and operations docs the repository has a controlled technical Chinese module. It borrows from ASD-STE100 the ideas of one term per concept, one action per step and warnings before the action. Next to it sits `scripts/lint_copy_rules.py`: a dependency-free check with error, warning and style levels.
Who it is for. For technical writers, docs engineers and product teams shipping Chinese-language documentation and interfaces.
Good fit when
- A documentation home page or landing page has come out promotional and vague and needs a rewrite
- Parameter descriptions, error-code tables and changelogs in Chinese docs need a consistent form
- The interface has literally translated statuses such as Success or Bad Request and needs natural wording
- A step-by-step install or troubleshooting guide mixes up who acts and under what conditions, and must be restructured with controlled-language principles
- You want an automated check of Chinese Markdown in CI: the linter catches typical errors and flags doubtful spots as warnings
Not a fit when
- The text is not in Chinese: the rules, terms and spacing are built for Chinese mixed with Latin text and digits
- You need creative or promotional copy: the skill strips pathos and holds a technical tone
- You need ASD-STE100 certification: the author states that the module is not a Chinese version of the standard and that lint results do not confirm compliance
- The wording comes from a law, a contract or an original error message: the skill keeps such fragments as they are and lists its suggestions separately
Example request
Rewrite our Chinese docs home page and error-code list with tech-doc-style-chinese, keeping all facts and limits unchangedLimitations
The skill covers only the Chinese language and Chinese technical text. It does not replace an editor: its priority rules put facts, limits and safety above style, and when information is missing it keeps the original meaning or marks the spot as "待确认", meaning it needs confirmation.
The controlled Chinese module borrows the ASD-STE100 approach, but the author states plainly that the project is not a Chinese version of the standard, promises no conformance with it, and that a lint pass does not prove compliance. STE rules for English, such as the controlled vocabulary, tenses and sentence length, do not transfer to Chinese directly. The linter splits findings into error, warning and style. Only the first are treated as certain mistakes and by default produce a non-zero exit code; the rest need a human decision. Project terms and tone go into a separate overrides file, with a template in `references/project-overrides-example.md`; its examples do not act as project rules.
How to disable. Delete the `tech-doc-style-chinese` folder from `~/.claude/skills/` or the project's `.claude/skills/`, then restart the agent. For a Codex install via `npx skills`, remove it from the Codex skills directory in the same way.
FAQ
How do I install Chinese Tech Doc Style in Claude Code?
Run `npx skills add https://github.com/Fenng/tech-doc-style-chinese -a claude-code`. With `-g` the skill lands in `~/.claude/skills/`; without it, in the project's `.claude/skills/`. Restart Claude Code afterwards, and the skill triggers from the description in SKILL.md.
What does the skill leave alone in a text?
Code literals, JSON keys, URLs, API paths, database fields, commands and configuration options. This is written in SKILL.md and in the README. The spacing and replacement rules apply to visible text only.
Which quotes and spacing does it apply?
Corner quotes 「」 for Chinese text. In visible text a space is set between Chinese characters, Latin letters and digits. Each Markdown paragraph stays on one line with no hard wrap, and `unwrap_md_paragraphs.py` expands paragraphs that were already broken.
How do I check a text automatically?
Run `scripts/lint_copy_rules.py`. It needs no dependencies, protects code blocks, inline code, URLs and paths, and sorts findings into error, warning and style. The repository has a `skill-lint.yml` CI workflow and tests in the tests folder.
Is this a Chinese version of ASD-STE100?
No, the author rules that out. The controlled Chinese module borrows STE ideas: one term per concept, one action per step, warnings before the action. It does not reproduce the standard and promises no conformance.
Discussion
No questions or reviews yet.
Related
SkillsA skill that cleans machine tone from text, Chinese-first: it drops filler and cliches while keeping the facts and the author's voice
A skill that rewrites dense, ambiguous English into ASD-STE100 Simplified Technical English so another agent cannot misread it
A skill that makes an agent write docs and replies in plain English on ASD-STE100 rules: short sentences, one term per meaning, condition before command
A set of LaTeX templates for Chinese research documents paired with Claude Code and Codex skills that fill and verify them
Simplified Technical Russian
Упрощенный технический русский (УТР)
A skill that makes the agent write code comments, docs and error messages by the Simplified Technical Russian rules of GOST R 58049-2017, with a linter
Simplified Technical English Skill
simplified-technical-english
A skill that makes a model write in Simplified Technical English (ASD-STE100), a controlled English for technical documentation. Unofficial
Comparison with similar tools
Entries that share the main search query, and how they differ.
| Entry | Official | Repository stars | Main difference |
|---|---|---|---|
| Chinese Tech Doc Style (this entry) | no | 1,518 | |
| Human Writing | no | 4,134 | writes and edits Chinese in a lively conversational voice, while this one holds the restrained tone of technical documentation |
| The Elements of Agent Style | no | 720 | general English writing rules for AI agents, without this skill's Chinese typography and linter |
README in short
The description, in Chinese, explains the goal: restrained, accurate, readable technical text. It lists the use cases and what not to touch, then the key rules, the STE approach and the repository layout with references, scripts and tests. Install routes for Codex and Claude Code, project overrides and the linter with a CI example follow.
README badge
Original SKILL.md
--- name: tech-doc-style-chinese description: 在撰写、改写、校对或审阅中文技术文档、产品文案、界面文案、Markdown 文档、API 说明、操作手册、故障排查或运维文档时使用。采用克制、准确、可扫读的中文技术写作风格；保留原文事实、限制和机器可读内容；按任务读取术语排版、API 状态文案、项目覆盖或受控中文技术写作参考。 --- # 中文技术文档与产品文案规范 <!-- 作者：Fenng（GitHub：@Fenng） --> ## 适用范围 将本 Skill 用于中文技术内容的撰写、改写、校对和审阅，包括： - 文档首页、产品介绍、解决方案页和更新日志 - API 文档、参数说明、错误码和常见问题 - 操作手册、故障排查、运维 Runbook 和安全说明 - 界面文案、按钮、导航、状态和提示信息 不要改写代码字面量、JSON 键名、URL、API 路径、数据库字段名、命令、配置项或其他机器可读标识符。 ## 规则优先级 发生冲突时，按以下顺序处理： 1. 保留事实、逻辑、限制、安全信息和法律含义。 2. 服从用户明确要求和目标项目约定。 3. 保持技术术语和机器可读内容准确。 4. 改善结构、语义、语气和可扫读性。 5. 最后处理标点、留白、大小写等排版细节。 不得为了句子更短、语气更确定或格式更统一而牺牲更高优先级的信息。 ## 事实保真 - 不新增原文或可靠上下文没有提供的日期、数字、时限、SLA、能力、条件、因果关系或结论。 - 不删除前置条件、适用范围、例外、风险、安全警告、兼容性说明或失败处理。 - 不把可能、计划、建议、通常等不确定表达改成确定事实。 - 信息不足时保留原意，或明确标记「待确认」；不要自行补全。 - 改写引用、法规、合同、错误原文或用户提供的固定文案时，优先保持原文并将建议单独列出。