Skip to content
cccskills
FreePublished on GitHub

ai-article

Write AI tool reviews, setup tutorials, interview preparation articles, and technical explanations, or update existing articles against current source code. The workflow covers research, outline approval, prose checks, and source documentation.

OriginalShow the original description

AI 类文章和项目教程的撰写与重构优化。支持新写和重构优化(结合最新源码重写已有文章)两种输入模式,四种风格:安装教程、产品评测、面试八股、深度拆解。覆盖 AI Coding 工具实测、AI 开发框架应用、大模型测评、Agent/Skills/RAG 技术讲解。

See how to install

Good for

  • Write a setup tutorial for an AI tool
  • Update an article against current source code
  • Prepare technical interview explanations
  • Draft an AI product review

Files included (30)

  • SKILL.md8.5 KB
  • biaoti.md4.9 KB
  • inbox.md1.0 KB
  • knowledge/agent-benchmarks.md17.4 KB
  • knowledge/ai-compiler-gpu.md10.6 KB
  • knowledge/china-ai-company-checklist.md9.7 KB
  • knowledge/deepseek-harness-desktop.md23.3 KB
  • knowledge/employee-series-published.md18.7 KB
  • knowledge/fable-5-1-harness.md17.1 KB
  • knowledge/nvidia-inference-agent.md18.5 KB
  • knowledge/openmuse.md6.4 KB
  • knowledge/paicli-architecture.md37.8 KB
  • knowledge/paicli-bugfix-loop.md17.4 KB
  • knowledge/paismart-rag.md55.0 KB
  • knowledge/xiaomi-mimo.md6.8 KB
  • references/agent-mianshi-xiaomi.md18.7 KB
  • references/anshui-yin-mianshi.md18.3 KB
  • references/claude-code-grep-vs-rag.md11.8 KB
  • references/deep-analysis.md1.9 KB
  • references/deepseek-tui-review.md10.5 KB
  • references/deepseek-v4.md8.4 KB
  • references/human-tone.md2.3 KB
  • references/interview-style.md3.5 KB
  • references/OpenClaw-install.md12.2 KB
  • references/preface-examples.md8.5 KB
  • references/review-style.md2.0 KB
  • scripts/check_body_length.py1.6 KB
  • scripts/check_prose.py21.8 KB
  • scripts/wording_rules.py4.9 KB
  • sucai.md704 B

SKILL.md (original)

Review the instructions the agent will receive before you install.

AI 技术文章写作

硬性规则

  • 正文标题 ## 01、标题,标题只写名称,不加冒号和后缀解释
  • 列表只用于短小并列的项目(工具清单、检查项、对照表),成段的内容用段落展开,不要连续几个章节都用列表
  • 叙述句和对话引导不用冒号,冒号只用于列表引导句、列表项的“术语:解释”和简历字段名;不用破折号
  • 正文叙述不出现项目自定义类名(AgentBudget、McpServerManager),用中文功能描述代替,类名只在代码块里出现
  • 读者是有经验的 Java 后端开发,熟悉 AI Agent,Kafka 消费者、位移、死信队列、CAS、Full GC、令牌桶、熔断、topK 这类常见概念直接用,不加解释。只解释项目自定义的东西和不常见的术语,工具名、角色名、参数名第一次出现时说清是什么、谁在用
  • 零歧义,读者不需要多想一步。指代写成具体对象,主语、宾语写全,不用需要先翻译一遍的比喻,不生造术语
  • 用户纠正过的用词和禁用词收在 scripts/wording_rules.py,由检查脚本执行,新的纠正追加进去,同时补进 references/human-tone.md「词语」表
  • 量化数据和时间估算必须有出处(源码、文档、实测),没有出处就用模糊表达,禁止编造精确数字
  • 技术断言要么有出处,要么标注“我的推断是”。正文直接讲消化后的事实,不复述“帖子里说”“官方介绍”,出处放进交付给用户的来源清单
  • 模型和产品举例用最新的(Fable 5.1、Opus 5.5、Sonnet 5.5、GPT-6 Astra、GPT-6 Luna、DeepSeek V4、GLM-5.3),AI Coding 工具用 Claude Code、Codex、Qoder 举例;热点词汇先查清当下业界的真实所指
  • 不写没有来源的场景细节(时间、天气、神态、对白)
  • 不用翻案腔,也就是先立一个读者并没有的误解,再推翻它给下文抬价。“不是A,而是B”“你以为A,其实B”“看似A,实则B”都是它的变形,修法是把判断从正面直接说出来。文章确实用一手材料走完了从误解到修正的过程时可以写,但不套这些句式。直接引文里的违禁内容改成间接引述
  • 章节和答案按因果往前推进,前一段的结论引出后一段的问题,不用“三层原因”“分三步”这类分类代替叙述(题目本身问的就是分层结构时除外)
  • 不写总结段和升华句,事情讲完就停

前言

前言是一篇独立的短文,包含钩子(热点、反差或痛点),用一句话说清这篇文章做了什么,作者作为同行者带读者往前走。情感方向是“在稳定的状态下变得更好”,不制造焦虑、不功利化。正例和反面模式见 references/preface-examples.md。

截图占位符

每个二、三、四级标题下至少 1 个,长章节 2 个,紧贴它可视化的内容,格式如下。

【截图:<名称>;风格:<风格>;截图目标:<证明什么>;关键词:<关键词1>、<关键词2>、<关键词3>】

风格只在 6 种里选,whiteboard(架构图)、skill-card(技能卡片)、data-board(数据对比)、three-layer(层级关系)、swimlane(泳道流程)、checklist-card(注意事项)。

简历包装

文章涉及实战项目时加一段简历包装,写项目名称、项目简介、技术栈、核心职责(5 条,做了什么 + 解决了什么问题 + 有出处的数据,不出现自定义类名)。

工作流程

1. 确定输入模式

  • 新写:从 Skill 目录下的 sucai.md 出发
  • 重构优化:用户指定已有文章,结合最新源码重写。保留真实截图和准确的代码,校正和最新实现对不上的内容,删掉废话;章节顺序按读者的问题链重组,不默认沿用原文目录。JobClaw 教程以重构优化为主

2. 调查

项目相关内容先调查再写,不凭通用知识猜。安装教程看官方文档,产品评测看 README、issue、PR,面试八股和深度拆解要读源码、用代码证明观点。

  • 先查 Skill 目录下的 knowledge/ 缓存。缓存文件开头写调研日期、仓库路径和 commit,用 git log --since="<调研日期>" 判断要不要增量补充,补充后同步改开头的日期和 commit;没有缓存就全量调研,写成新的缓存文件
  • 读源码交给 Sub-agent,主对话只收结论
  • 源码实现不够好时,答案直接按面试官期望的方案写,正文不承认项目现状的不足。大纲里标注哪些基于源码、哪些是理想方案;交付时把差距写进项目仓库的升级计划(派聪明是 PaiSmart/docs/interview-upgrade-plan.md,其他项目放在 docs/ 下同名文件),每条写清现状(源码位置)和目标,用户会据此迭代源码
  • 公开数据从原始来源获取,访问不到的注明“截至 YYYY-MM-DD”

3. 确定风格和参考文章

自己决定,写进大纲让用户把关。选题不属于任何一种时按素材的问题链组织。

  • 安装教程类:参考 references/OpenClaw-install.md
  • 产品评测类:规范见 references/review-style.md,参考 references/deepseek-tui-review.md、references/deepseek-v4.md
  • 面试八股类:规范见 references/interview-style.md,参考 references/agent-mianshi-xiaomi.md、references/anshui-yin-mianshi.md
  • 深度拆解类:规范见 references/deep-analysis.md,参考 references/claude-code-grep-vs-rag.md

参考文章学判断力和节奏,不照搬句式。

4. 出大纲,用户确认后再写

大纲包含风格和参考文章(各一句理由)、章节结构和每章的推进线(用一句话串起来,不分条)、源码参考策略、完整前言样本(默认约 800 字,素材冲击力强时可以缩到两三百字)。

5. 撰写

扫一眼 Skill 目录下的 inbox.md,看有没有能用的素材。字数不设下限,讲清楚就停,不为凑字数加内容。

头部模板如下。

---
title: # 写完后回填
shortTitle: # 写完后回填
description: # 50-120 字 SEO 描述,含 2-3 个核心关键词
keywords: # 5 个搜索关键词
tag:
  - Agent  # 面试对话类用:- 面试
category:
  - AI
author: 沉默王二
date: # YYYY-MM-DD
---

6. 自检

  • python3 <Skill 目录>/scripts/check_body_length.py <稿件路径>,字数写进自检报告,不因为字数多少增删内容
  • python3 <Skill 目录>/scripts/check_prose.py <稿件路径>,“需要修改”的必须改,“需要人工判断”的逐条确认
  • 当成第一次接触这个话题的读者通读一遍,哪里要多想一步才懂、哪个结论超出了材料能支撑的范围,都改掉

给用户一份简短的自检报告。

7. 落盘和交付

  • toBeBetterJavaer 新写:保存到 docs/src/sidebar/itwanger/ai/,文件名用主题关键词
  • JobClaw 新写:放进 docs/ 对应章节目录,拿不准就问
  • 重构优化:覆盖原文件,保留文件名和序号前缀

交付时附上项目升级计划的更新情况、截图来源清单(有链接给链接,需要自己操作的标注“需自行操作”)、事实来源清单(关键数字和技术细节逐条对应源码位置或原始链接),以及 5 个候选标题(素材里有现成标题直接用;没有就参考 biaoti.md 里打开率高的标题)。

作者与项目

  • 作者:沉默王二(二哥),GitHub https://github.com/itwanger
  • 网站:javabetter.cn、paicoding.com(技术派社区)
  • 实战项目源码都在 /Users/itwanger/Documents/GitHub/ 下,涉及项目细节时读源码,不要编造:
    • 技术派(paicoding):前后端分离的技术社区系统,即 paicoding.com
    • PaiCLI(paicli):对标 Claude Code 的 Java Agent 命令行工具
    • PaiAgent(PaiAgent-one):LangGraph4j + Spring AI 的工作流编排平台
    • 派聪明(PaiSmart):基于 ES 混合检索的 RAG 知识库
    • PaiFlow(PaiFlow):可视化 AI Agent 工作流编排平台,类 Dify/Coze/n8n
    • PmHub(pmhub):基于 SpringCloud 和 LLM 的智能项目管理系统
    • 求职派(JobClaw):OpenClaw 风格的多 Agent 求职系统,教程在仓库的 docs/ 下

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

browser-act

FreeSummary

Automate browser tasks such as reading JavaScript-rendered pages, filling forms, clicking through workflows, and taking screenshots. Maintain login sessions, manage separate browsers, and capture network responses, with approval checks for sensitive actions.

  • Extracting content from pages that use JavaScript
  • Automating form entry and page navigation
  • Checking page appearance with screenshots
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

Turn narration scripts into animated explainer videos using Remotion and 火山 TTS, a text-to-speech service. Prepare scripts, synchronize visuals with narration, preview each chapter, correct pronunciations, and export a verified MP4 with voice audio.

  • Turn a narration script into an animated explainer video
  • Synchronize subtitles and visuals with narration
  • Preview and refine a video chapter by chapter
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

Create a coordinated set of video covers in 3:4, 16:9, and 4:3 formats from AI or Java scripts. Choose illustrations featuring a consistent presenter character or text-only layouts with white titles and yellow keywords.

  • Creating matching vertical and horizontal video covers
  • Making text-only covers with white and yellow lettering
  • Keeping the same presenter character across covers
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

video-script

FreeSummary

Write or refine Chinese spoken video scripts about AI technology, interview questions, and current topics. New scripts follow research and outline approval, while revisions provide specific wording changes and reasons for the edits.

  • Writing a spoken script explaining AI technology
  • Turning technical interview questions into video scripts
  • Preparing a script about a current AI topic
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

video-sync

FreeSummary

Match videos from 王二讲Agent to their scripts, then add Bilibili links and Douyin video IDs to the designated documentation files. Identify missing entries while preserving existing links and IDs.

  • Add newly published Bilibili videos to the documentation
  • Register Douyin video IDs for embedded players
  • Check which scripts still lack video references
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

zhihu-search

FreeSummary

Search Zhihu, a Chinese knowledge-sharing platform, for answers, articles, people, and topics. Filter by content type and creation date, sort results, and collect structured records containing excerpts, authors, votes, and links across multiple pages.

  • Finding Zhihu answers and articles on a topic
  • Comparing highly upvoted answers
  • Collecting recently created content
itwanger/toBeBetterJavaer17.8KUpdated Oct 18, 2026

See all skills by itwanger

Report a problem with this skill