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 qingshanliuci/cnki-aigc---skill --skill cnki-aigc-skillgit clone --depth 1 https://github.com/qingshanliuci/cnki-aigc---skillWrote 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/qingshanliuci/cnki-aigc---skill/cnki-aigc-skill)<a href="https://agentmods.dev/skills/qingshanliuci/cnki-aigc---skill/cnki-aigc-skill"><img src="https://agentmods.dev/badge/skills/qingshanliuci/cnki-aigc---skill/cnki-aigc-skill/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/qingshanliuci/cnki-aigc---skill/cnki-aigc-skill"><img src="https://agentmods.dev/badge/skills/qingshanliuci/cnki-aigc---skill/cnki-aigc-skill.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 | $0.00126 | $0.02438 |
| Opus 5.5 | $0.00050 | $0.00975 |
| Sonnet 5.5 | $0.00025 | $0.00488 |
| Haiku 4.5 | $0.00013 | $0.00244 |
Grade A, and why
cnki-aigc-rewrite 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 4d 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 — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.
When to use
用户做以下任一动作时触发:
- 提到"知网AIGC检测""AI率""红色显著""降AIGC""降AI率"
- 上传AIGC检测报告(PDF/截图)
- 贴出一段被标为红色或疑似的论文文本
- 询问"怎么改这段不像AI写的"
仅适用于中文学术论文(学位论文、期刊投稿、学术报告)。不适用于:
- 摘要/Abstract(结构化文本天然像AI,但通常不计入AI字符数)
- 文献引用密集段("X等(2020)发现……"难以脱模板)
- 研究方法固定描述句("采用双向固定效应模型"是规范表达)
- 非学术文本(自媒体、营销文案、小说)
核心认知:检测器抓的是什么
知网AIGC检测器不是抓"AI写的",而是抓5种语言模式。理解这5种模式比记规则更重要:
模式1:句法节奏的可预测性 — LLM输出句长集中在20-35字(正态分布),人类写作是齐夫分布(少数特别长、少数特别短、多数中等)。改写要刻意打破均匀。
模式2:信息密度的均匀性 — LLM每段密度都差不多,人类写作有"高密度论证段"和"低密度铺陈段"。检测器跟踪段级特征值方差。
模式3:术语的句法位置固定 — LLM倾向把术语放在主语位置("逆向技术溢出转化为创新能力"),人类更常放在话题或宾语位置("企业把外部技术转化为创新能力")。不动术语本身,但挪动它在句子里的角色。
模式4:连接词功能重叠 — LLM爱用"因此""从而""进而""由此",功能高度重叠都在做因果推进。人类的连接词承担多种功能:补充(也就是说)、转述(换句话说)、强调(——)、对照(同向)。
模式5:模板段功能全等性 — 三/四项工整排比、抽象名词链、段尾元话语收束、模板化报告句、平行铺陈+末尾收束。这5个具体特征是上面5种模式的表层呈现。
工作流:根据输入长度自动切换
路线A:交互式精修(输入<800字)
用户贴一段红色显著文本进来,按以下步骤处理:
步骤1:识别高危特征 扫描这段文本,列出它命中了下面哪几个具体特征(不必全列,列最关键2-3条):
- 三/四项工整排比("从X看……从Y看……从Z看……" / "对A……对B……对C……" / "一方面……另一方面……" / "A以a为主;B以b为主;C以c为主")
- 抽象名词链("传导机制""调控变量""动态监测体系""未被纳入检验的路径仍是传导重要环节")
- 段尾元话语("这一结论""上述发现表明""为后续……提供了实证依据")
- 模板化报告句("X的研究表明""在统计推断意义上获得支持""与预期方向一致""下降至/上升至")
- 平行铺陈+末尾收束
步骤2:执行改写动作 按下列动作改写。保留所有数字、变量名、文献编号[N]、年份、作者姓名、专业术语:
-
拆三/四项排比:每一项换不同句式起头(描述句/判断句/动宾句/转折句各一)
- ❌ "从产业基础来看……从人力资本来看……从投资动机来看……"
- ✅ "中部处于工业化中期;高校与高职供给的工程技术劳动力够用;OFDI动机上又以战略资产寻求型为主"
-
抽象名词链 → 动词短语:
- "传导机制较为契合" → "跟……同向"
- "在统计推断意义上获得支持" → "被样本较为充分地支持"
- "动态监测体系" → "看看是不是还走在一起"
-
删段尾元话语:让段落自然结束。
-
替换LLM典型语套:
- "研究表明" → "看到的是"/"读出的是"
- "进一步揭示" → "再往下看是"
- "下降至" → "跳到"
- "一致" → "对得上"/"同向"
- "释放/呈现" → 用具体动作替换
-
打乱句长方差:长短句穿插,用"——""换句话说""问题是""也就是说"让节奏不均匀。不要写成短句串+连接词,那是另一种AI腔。
-
挪动术语句法角色:原文术语在主语位置的,改写时换到宾语或话题位置。
步骤3:自检5项
输出前必须过一遍:
- 数字、变量名、引用[N]是不是一字未动?
- 段落里还有没有任何"X的Y的Z"三层名词嵌套?
- 有没有两组以上"A,B,C"并列结构?
- 段尾有没有"这一……""上述……"元话语收束?
- 全段是否倒向口语化(找不到任何学术连接词)?保留至少2-3个学术连接词("鉴于""因此""二者"),否则会从AI腔倒向博客腔。
任一项不通过,重写。
Done when: 输出包含改写版 + 改了哪些高危特征(1-3条最关键的)+ 未动的部分(数字/术语/引用列表)。
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 4d ago First seen · 147 lines · 126 tokens per session scan A 99e7cc39439e
cnki-aigc-rewrite is a skill published in the GitHub repository qingshanliuci/cnki-aigc---skill (51 stars, last pushed 5mo ago), licensed MIT. It adds 126 tokens to every session and 2,438 once invoked, about $0.0005 per session on Opus 5.5. 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 2026-10-21.
Other skills, from other repositories
mem-reading-report
A reading-report and literature-review assistant for master's students, especially ZJU MEM students. It works from several papers plus report and student details to produce or review reports in the required format.
thesis-deai-review
A review guide for Chinese degree theses, especially evidence-based engineering-management theses. It checks for formulaic wording, weak evidence, and writing that may appear machine-generated.
thesis-orchestrator
A coordinator for writing a Zhejiang University MEM thesis, where MEM means Master of Engineering Management. It identifies the current stage, plans the overall route, directs work to one of 21 related skills, and tracks progress and risks.
aigc-reduce-playbook
A Chinese-language academic writing review workflow for reports that flag text as possibly generated by AI. AIGC means artificial intelligence-generated content, and the workflow applies to theses and journal papers.
mem-midterm-report
A Chinese-language workflow for preparing a ZJU MEM midterm report. ZJU is Zhejiang University, and MEM is a Master of Engineering Management degree; the workflow covers the report, registration form, presentation outline, and submission guidance.
thesis-defense
A preparation guide for ZJU MEM thesis defenses, covering midterm reviews, preliminary defenses, blind review, and the final defense. It organizes review checks, presentation planning, likely questions, video-based question analysis, and compact preparation notes.