WaterPaper
水论文skill一句话出论文:多角度选题、真实文献采集、大纲写作、降低AIGC率、降低重复率、绘制图表、DOCX交付
Install / Use
npx skills add ThisIsLittleSky/WaterPaperInstalls into whichever agent you are using.
SKILL.md
Installable skill definition
Quality Score
Category
Content & MediaSupported Platforms
Our assessment of WaterPaper
WaterPaper scores 74/100 on our quality scale, 1056th of 1,171 Content & Media skills we index.
Its SKILL.md is 7.7 KB long, well organised into 10 sections with 4 code examples: a thorough specification that gives an agent plenty to work with.
It has 40 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated about 4 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
- Our last check on 2026-10-18 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 90/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
Safety scan
No issues foundOur scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.
AI review by kimi-k2.7-code on 2026-09-24. Automated pattern scan on 2026-09-24. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.
WaterPaper compared with similar skills
All 4 of these similar skills score higher than WaterPaper; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| WaterPaper (this skill)by ThisIsLittleSky | 74 | 40 | 4mo ago | SKILL.md |
| Agent-Reachby Panniantong | 100 | 109.5k | 1d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 79.0k | today | CLAUDE.md |
| Scraplingby D4Vinci | 100 | 98.7k | today | MCP Server |
| crawl4aiby unclecode | 100 | 87.0k | 2d ago | MCP Server |
Frequently asked questions
- How do I install WaterPaper?
- Run
npx skills add ThisIsLittleSky/WaterPaper. The install tabs above show the steps for each supported agent. - Which AI agents does WaterPaper work with?
- It is written for Claude Code, as a SKILL.md file. Other agents that read the same format can often use it too.
- Is WaterPaper safe to use?
- Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 90/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is WaterPaper still maintained?
- The repository was last updated about 4 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
Skill content
View source on GitHub效果展示
<table align="center"> <tr> <td align="center"><img src="images/效果图1.png" alt="效果展示1" width="350"></td> <td align="center"><img src="images/效果图2.png" alt="效果展示2" width="350"></td> <td align="center"><img src="images/效果图3.png" alt="效果展示3" width="350"></td> </tr> </table>快速开始
把本项目的 GitHub 地址或下载到本地的文件夹路径,发给你的 AI 编程助手(Claude Code / Codex / Trae / Cursor 等),然后说:
帮我安装这个 skill:https://github.com/ThisIsLittleSky/WaterPaper.git
AI agent会自动完成配置
装好之后,直接说你的需求:
用这个skill帮我写一篇关于移动通讯技术的期末论文,4000 字,格式模板是模板.docx
帮我给我的论文降低AI率,【论文路径】、【查重报告.html(可选)】
AI 会走完整流程:要模板 → 出选题 → 爬文献 → 出大纲 → 写正文 → 出图 → 交 .docx
为什么你需要这个
期末周,N 篇课程论文压过来。每篇的流程都一样:
想题目 → 搜文献 → 看文献 → 写大纲 → 凑字数 → 调格式 → 降低AI率 → 交付
【水论文.skill】把这些事自动化了:
| 环节 | 手工方式 | 大模型客户端 | 【水论文.skill】 | |------|----------|-------------|------------| | 选题 | 绞尽脑汁想 1 个 | AI 生成选题,但无真实文献支撑 | 一句话出 5 角度 × 2 选题 = 10 个 | | 文献 | 翻 CNKI、万方,复制粘贴 | ❌ 编造参考文献,标题作者看似真实 | 多源爬虫自动抓取,去重去假 | | 格式 | 对着模板一行行调 | 需手动描述格式,无法精确复刻 | 上传 .docx 模板 → 自动提取 → 严格复刻 | | 图表 | 用 Excel 画、截图、插入 | 无法生成或生成简陋 | HTML 渲染科研级 SVG 图表 → 自动插 | | 降AI | 硬着头皮改套话、调句式 | 无降AI能力 | D0-D7 七维约束 + PaperPass 五模式扫描 + 9 大改写技法 | | 降重 | 手动改写重复段落 | 无降重能力 | 深度语义改写 + 表述角度转换 + 引用归并 |
大模型客户端写论文的致命问题是编造参考文献——标题、作者、期刊看起来像真的,实际上不存在。【水论文.skill】所有参考文献均来自 CrossRef / Semantic Scholar / 百度学术的实际抓取,每条标注来源和可核验状态,不确定的条目直接丢弃。
你只需要两件事:说需求 和 做选择。剩下的交给脚本。
文献数据源
| 数据源 | 接入方式 | 中文覆盖 | 可靠性 | 速率 | |--------|---------|----------|--------|------| | CrossRef | REST API(免费) | 一般 | ✅ 高 | 50 req/s | | Semantic Scholar | REST API(免费) | 一般 | ✅ 高 | 受 API Key 控制 | | 百度学术 | Web 抓取 | ✅ 好 | ⚠️ 中(需核验) | — | | CNKI | Web 抓取 | ✅ 最好 | ⚠️ 低(反爬严格) | — |
中文文献建议以百度学术为主、CrossRef/Semantic Scholar 补英文。CNKI 反爬严格,如果抓取失败属正常现象,不影响整体流程。
论文结构(课程论文)
| 论文字数 | 中文文献 | 英文文献 | 章节结构 | |----------|----------|----------|----------| | 3000-4000 | 4-6 | 1-2 | 引言 + 2 章正文 + 结论 | | 4000-6000 | 5-7 | 2-3 | 引言 + 2-3 章正文 + 结论 | | 6000-8000 | 6-8 | 2-4 | 引言 + 3 章正文 + 结论 |
支持学科类型:经管、人文社科、理工、案例分析。
图表类型
- 柱状图/条形图(类别对比)
- 折线图(趋势变化)
- 流程图/框架图(理论框架、研究流程)
- 数据表格(多维度对比)
- 饼图/环形图(占比展示)
所有图表为 HTML/SVG 渲染、2x DPR 高清输出,学术配色风格。
常见问题
<details> <summary><b>Q: 学校模板里奇怪的格式能复刻吗?</b></summary>能。analyze_template.py 会逐段分析 .docx 文件,提取每种元素的字体、字号、加粗、对齐、行距、段前段后、首行缩进、是否分页,生成结构化 JSON。generate_paper_docx.py 按这个 JSON 严格复刻。
如果你的模板格式超出脚本分析范围(比如特殊的页眉页脚、水印),可以手动编辑 style_profile.json 补充。
【水论文.skill】内置了两道防线,在 PaperPass 上实战验证过:
第一道:降AI处理(D0-D7 七维约束 + PaperPass 五模式扫码)
- 整合了 thesis-optimizer 项目的 30+ AI 模式分类学和三级词汇黑名单
- D0 最小干预原则:句内微调,不搞大段 AI 式重写
- D1 句长分布:主动制造长短句交替,打破 AI 的钟形句长分布
- D2 段落结构:5 种模板随机切换,相邻段落结构不重复
- D3 信息密度:核心段高密度 + 过渡段低密度,形成"高-低-高"交替
- D4 连接词控制:红色高风险词("此外""至关重要""综上所述")直接删除,黄色中风险词控制密度
- D5 术语语境:heavy 档启用术语变体
- D6 逻辑人性化:打破 AI 式"问题→方法→结论"直线型逻辑,保留探索的曲折性
- D7 并列必死:知网/PaperPass 对并列结构极其敏感——"第一/第二/第三""首先/其次/最后"直接粉碎,改用碎片化断句和空行分段打断并列
- 写作完成后自动运行
humanize_check.py验证(含并列检测),不通过不交付
第二道:降重处理(深度语义改写)
- 标准定义段:重新组织语序,避免教科书式表述
- 文献综述段:分类归纳 + 多源归并引用,不做文献流水账
- 方法描述段:增加"为什么选择此方法"的动机说明
- 结论总结段:用具体发现替换泛泛总结
- 全程术语保护(专业缩写、数学模型、引用编号禁止修改)
拿到成稿后建议通读一遍,加入你自己的观点和课程相关的具体内容。水论文是工具,不是代写。
</details>工作流
<details> <summary><b>点击展开:完整 10 步工作流</b></summary> <br/>flowchart TD
YOU["<b>你</b><br/>帮我写一篇XX课期末论文"]:::user
subgraph R1 [ ]
direction LR
S1["<b>1. 格式提取</b><br/>上传模板,自动对比默认格式"]:::step
S2["<b>2. 选题生成</b><br/>AI 生成 5 角度 × 2 选题"]:::step
S3["<b>3. ⚡ 文献采集</b><br/>CrossRef + SS + 百度学术 + CNKI"]:::step
S1 --> S2 --> S3
end
subgraph R2 [ ]
direction RL
S4["<b>4. 文献核验</b><br/>标注可信度,✅高 / ⚠️中"]:::step
S5["<b>5. 大纲 + 字数预算</b><br/>基于真实文献分配引用"]:::step
S6["<b>6. 逐章写作</b><br/>按预算控字,句句标引"]:::step
S4 --> S5 --> S6
end
subgraph R3 [ ]
direction LR
S7["<b>7. 图表渲染</b><br/>HTML/SVG 科研图表 → PNG"]:::step
S8["<b>8. 降AI检查</b><br/>D0-D7 七维 + PaperPass 五模式"]:::step
S9["<b>9. 降重处理</b><br/>语义改写 + 表述角度转换"]:::step
S7 --> S8 --> S9
end
S10["<b>10. .docx 成稿</b><br/>严格按模板格式,标题命名"]:::step
YOU --> S1
S3 --> S4
S6 --> S7
S9 --> S10
classDef user fill:#fff3cd,stroke:#ffc107,color:#856404,stroke-width:2px
classDef step fill:#e3f2fd,stroke:#2196f3,color:#0d47a1,stroke-width:2px
style R1 fill:transparent,stroke:transparent
style R2 fill:transparent,stroke:transparent
style R3 fill:transparent,stroke:transparent
</details>
项目结构
<details> <summary><b>点击展开:项目目录结构</b></summary> <br/>WaterPaper/
├── SKILL.md # 技能主定义
├── requirements.txt # Python 依赖
│
├── prompts/ # AI 提示词模板
│ ├── format_extractor.md # 格式提取(.docx 模板 & 文字描述)
│ ├── topic_selector.md # 5 角度 × 2 选题生成
│ ├── outline_builder.md # 大纲 + 字数预算 + 文献分配
│ ├── chapter_writer.md # 逐章写作 + 引用规则 + 降重约束
│ ├── chart_designer.md # 科研 HTML 图表模板
│ ├── humanize_constraints.md # D0-D7 降AI写作约束(含 PaperPass 专项表)
│ ├── humanize_pass.md # 独立降AI改写流程(36 键 + 9 技法 + 五模式扫描)
│ ├── detection_pass.md # 降AI检测流程(含 PaperPass 反馈迭代改写)
│ └── plagiarism_pass.md # 独立降重改写流程
│
├── tools/ # Python 工具脚本
│ ├── analyze_template.py # DOCX 模板格式分析器
│ ├── literature_scraper.py # 多源文献爬虫(4 数据源)
│ ├── render_html_chart.py # HTML → PNG 渲染(Playwright)
│ ├── count_words.py # 中英混合字数统计
│ ├── generate_paper_docx.py # Markdown → DOCX(python-docx)
│ └── humanize_check.py # 降AI效果验证(三级词汇报告 + 密度/并列检测)
│
└── references/ # 参考规范
├── course_paper_structure.md # 4 类学科论文结构模板
├── default_format.md # GB/T 7714 默认格式规范
├── detection_principles.md # 各平台AI检测原理分析
├── humanize_platforms.md # 各平台降AI策略参考(含知网 v3.0)
├── humanize_matrix_template.md # humanize_matrix.md 模板
├── paperpass_patterns.md # PaperPass 五大致命模式 + 破解实例 + 降分数据
├── rewrite_methods.md # 9 大降AI改写技法速查
├── ai_pattern_taxonomy.md # 30+ AI 模式分类学(S01-S10 含 PaperPass 专项)
├── ai_vocabulary_blacklist.md # 三级词汇黑名单
└── term_whitelist.md # 术语保护白名单
</details>
License
MIT
<p align="center"> <sub>水论文不是真的水。把体力活交给脚本,把时间留给真正重要的东西。</sub> </p> <p align="center"> <a href="https://github.com/ThisIsLittleSky/WaterPaper/issues">Issues</a> · <a href="https://github.com/ThisIsLittleSky/WaterPaper/pulls">Pull Requests</a> </p> <p align="center"> <sub>欢迎提交 Issue 和 Pull Request!</sub> </p>
Related Skills
Agent-Reach
109.5kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
headroom
79.0kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Scrapling
98.7k🕷️ The BEST Web Scraping framework that handles everything from a single request to a full-scale crawl! Don't be shy, join here: https://discord.gg/EMgGbDceNQ and follow here for daily tips and tricks: https://x.com/Scrapling_dev
crawl4ai
87.0kOfficial Crawl4AI Cloud: a hosted web scraping, extraction, and search API for AI agents. Free API key, 1,000 credits a month, no card. Fully open source: this repo is the Apache-2.0 crawler that powers it, and you can self-host it.
Languages
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
