Agent Skill
Human Writing skill
human-writingKKKKhazix/human-writinghuman-writing/SKILL.md
通用中文创作与改稿 Skill。用于知乎回答、论坛长帖、公众号文章、博客、评论、人物故事、历史叙事、新闻与行业解读、科普、教程、评测、个人叙事、小说、故事、对白、口播和演讲稿。默认写成一个见过事、查过材料、愿意把来龙去脉讲清楚的人在说话,重点保留中文互联网长回答与长帖的活人感和自然中文韵律,避免空泛的机构腔、喊口号式演说腔、营销腔和模型腔。非虚构长文先检查材料够不够,材料不足时研究、追问或缩短,绝不用重复解释灌字数。现实内容额外核验事实、引语、数据与用户亲历,虚构内容可以创造人物、场景、对白、心理与情节。成稿正文严禁冒号、破折号、“不是……而是……”及同类翻案句,并清除商业黑话和模型惯用黑话。不创建作者画像、个人规则库或个人写作 Skill。
always on
257 tokname + description
once used
3,765 tokSKILL.md body
bundled files
91 scripts
repo stars
4.2kGitHub
Install the Human Writing skill
Skills CLI (Claude Code, Codex, Cursor and others)
npx skills add https://github.com/KKKKhazix/human-writing --skill human-writing
Claude Code, by hand (all projects)
git clone --depth 1 https://github.com/KKKKhazix/human-writing.git /tmp/human-writing mkdir -p ~/.claude/skills cp -R /tmp/human-writing/human-writing ~/.claude/skills/human-writing
For one project only, copy the folder into .claude/skills/ in that project instead. The agent picks it up when a task matches its description.
What it costs your context
Agent Skills load in three steps. Tokens counted with the o200k tokenizer; your model's count will differ a little.
| When | What loads | Cost |
|---|---|---|
| Every session | Name and description | 257 tok |
| When the agent uses it | The SKILL.md instructions | 3,765 tok |
| Only if opened | 9 bundled files (73 KB) | varies |
Until it is used, human-writing costs about 5.3× less than the median MCP server we measured (1,363 tokens on every turn). Skills vs MCP study
Checks
- Ships scripts the agent can run. Read them before installing.
Checked 2026-10-22 · method
What's inside
The headings of its SKILL.md. Read the full file on GitHub.
- 活人感写作 1.1.0
- 第一关先看作品靠什么站住
- 按任务读取
- 多条规则同时出现时
- 先确定作品对现实作出什么承诺
- 现实写作
- 虚构创作
- 混合创作
- 动笔前先找到说话位置
- 第一稿直接带着人写
- 不要穿论坛服装
- 成稿绝对不能出现
- 交稿