Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add tance-mang/chinese-webnovel-skills --skill deslopgit clone --depth 1 https://github.com/tance-mang/chinese-webnovel-skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tance-mang/chinese-webnovel-skills/deslop)<a href="https://agentmods.dev/skills/tance-mang/chinese-webnovel-skills/deslop"><img src="https://agentmods.dev/badge/skills/tance-mang/chinese-webnovel-skills/deslop/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tance-mang/chinese-webnovel-skills/deslop"><img src="https://agentmods.dev/badge/skills/tance-mang/chinese-webnovel-skills/deslop.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00164 | $0.01776 |
| Opus 5 | $0.00082 | $0.00888 |
| Sonnet 5 | $0.00033 | $0.00355 |
| Haiku 4.5 | $0.00016 | $0.00178 |
Grade A, and why
deslop scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 65 lines — stays where its author put it; the contents beside it link to each section on GitHub.
去 AI 味（deslop）
AI 写的网文常有通病：太工整、爱排比、空洞抒情、翻译腔、用词高级但没网感。本技能把它改成"真人网文作者"的手感。
何时用
- 正文读起来"像 AI / 像作文 / 有翻译腔"
- AI 生成的稿子要落地成网文
- 文字太"端着"，没有网文的爽利感
AI 味的典型症状 → 改法
| 症状 | 表现 | 改法 |
|---|---|---|
| 无缘无故的修辞 ⭐ | 为修辞而修辞的比喻/叠喻，对剧情情绪没帮助 | 每个比喻问"它让读者更懂更爽了吗"，没用就删；爽文优先具体动作不优先比喻 |
| 跨度大/不搭的引用 ⭐ | 朴实文里突然引古诗典故/学术概念/不合身份的梗，语域不一致 | 先定语域全文统一；问"这个叙述者会这么说吗""和全文一个频道吗"，不搭就删换 |
| 排比癖 | 三句排比抒情、"不是…而是…" | 删掉两句，留最有力的一句，或改成动作 |
| 喊情绪（不演） | "我很心疼""他感到无比愤怒" | 用神态/动作/生理反应/潜台词演出来，让读者自己感受（详见 references/show-emotion.md，专项改用 emotion 技能） |
| 翻译腔 | "这是一个如此美丽的……" "令人难以置信的" | 改口语："美得不像话""离谱" |
| 过度工整/句子太顺 | 每句长度差不多、整齐划一、太顺；上下段字数行数雷同 | 句长不均(短长交错+偶尔50字长句)、加省略号/破折号/换行做停顿；段落打破对称（详见 references/sentence-rhythm.md） |
| 重复修辞/句式/词 | 同一个比喻、句式、词反复出现 | 通篇查重，删或换 |
| 同义反复 | "他非常生气，他怒了"——一个意思说两遍 | 留最准的一个，最好换成动作 |
| 极端词滥用 | 非常/极其/极大/无比/十分 堆砌 | 删九成，程度靠具体画面给 |
| 莫名陈述句/强行升华 | 突然插"生活总是充满挑战"这类空话 | 一律删 |
| 句号当顿号 | "苹果。香蕉。橘子。" | 并列用顿号“、”，逗号停顿，句号留给完整句 |
| 装饰符号癖 | 正文/标题撒 ✦✨◆●★ 当点缀或分隔 | 全删，分隔用空行或"——"，标题用纯文字 |
| 省略号用错 | 用 ... 或 。。。 |
中文省略号用“……”(六点一个标点) |
| 破折号滥用 | —— 反复出现 | 极少用，能用句逗就别用；一章顶多一次 |
| 方括号入正文 | 正文用【】 | 正文不用【】(仅系统流弹窗例外) |
| 直角引号/日式引号 | 对话用「」或『』 | 简体网文不用「」『』，一律改弯双引号""，嵌套用弯单引号'' |
| 引号用错 | 平白用单引号''当主引号、或用英文 "" '' | 主引号用弯双引号""，单引号''只用于嵌套 |
| 滥用高级词 | 堆四字成语和书面语 | 换网文常用的直白表达 |
| 解释过度 | 把人物情绪/动机全解释一遍 | 留白，让动作和对话自己说话 |
| AI 转折词 | "然而""与此同时""不仅如此""值得一提的是" | 删掉或换成口语连接 |
| 端着的对话 | 角色说话像书面语 | 改成符合人设的口语，带语气词 |
工作流
- 先读用户的文字，标出 AI 味最重的句子
- 逐处按上表改写，给出"原句 → 改后"对照
- 整体调节奏：加碎句、单句成段、对话提速
- 保留剧情和爽点不变，只改"手感"
完整深度清单（含无缘无故修辞、跨度大引用、语域一致、show-don't-tell）见
references/anti-ai-checklist.md。
网文文字手感
- 短：忌长难句，一句话一个意思
- 快：对话和动作推进，少叙述
- 狠：情绪用具体细节砸出来，不铺排形容词
- 活：适当口语、网感、语气词，但别过火
- 一致：先定语域(朴实/华丽/古风/现代)，全文统一——朴实老成的文就别乱引典故和怪东西，风格一致 > 单句出彩
注意
- 去 AI 味 ≠ 加网络烂梗，过度玩梗也会让稿子廉价
- 保持作者原有风格，只去"机器感"不强行改风格
- 改完可对照原文确认剧情/人设没改变
配合
- 写正文时就少出 AI 味 → 用
expand（已内置避免 AI 腔的要求）。 - 改之前想先量化定位"AI 味多重、扣分在哪几句" → 用
aidetect（AI 味指数 + 八项指标 + 定位清单，见references/ai-detector.md），按它的优先级下刀更快。 - 本技能治表面（修辞/标点/符号/重复/极端词）。若表面改干净了还"像 AI/像说明书"，那是深层缺人味——用
human技能加失控感（非理性行为/情绪不完整/不闭环/私货），见references/human-texture.md。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 65 lines · 164 tokens per session scan A bcc5a7d1166c
deslop is a skill published in the GitHub repository tance-mang/chinese-webnovel-skills (58 stars, last pushed 3mo ago), licensed MIT. It adds 164 tokens to every session and 1,776 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
fec-drawio-studio
An editable workflow for making technical diagrams in draw.io (also called diagrams.net), with the original .drawio file saved alongside exported images or documents. It covers architecture, database, UML, sequence, flow, machine-learning, and code-structure diagrams.
fec-nextjs-project-standard
A development guide for Next.js 14 and newer projects using App Router, Next.js’s file-based routing system. It covers server-rendered pages, layouts, data loading, middleware, metadata, and server actions.
fec-component-testing
A guide for writing and reviewing front-end unit, component, and lightweight integration tests. These tests check small pieces of code and UI behavior without running a full browser journey.
fec-monorepo-project-standard
A guide for organising a monorepo, which is one code repository containing several applications and shared packages. It covers workspace dependencies, folder structure, build tasks, testing, caching, and package releases.
fec-nuxt-project-standard
A guide for building Nuxt 3 applications with Vue 3 and TypeScript. Nuxt is a framework for Vue that can render pages on the server, generate static pages, or run as a browser-only app, using file-based routes and built-in data-fetching patterns.
fec-route-protection
A guide for protecting front-end routes, meaning the pages and URLs of a web application, based on whether someone is signed in and what permissions they have. It covers login checks, role-based access, expired sessions, and redirects.