staruhub/ClaudeSkills / skills/Geek-skills-deck-studio
deck-studio
VerifiedPPT 生产 Agent:理解场景 → 推荐风格 → 先出大纲 → 定义页面 → 双通路交付(PPT 内容稿 / 逐页视觉图 / 信息图组图)。当用户要做汇报、路演、培训课件、提案、咨询报告的演示文稿,或要把大纲/文章做成 slides、deck、PPT、信息图、小红书图文、可视化传播图时使用。支持指定风格或从风格库推荐,支持 image 模型逐页出视觉稿。前身为 notion-infographic(v2)并吸收 ppt-designer 的设计原则。不用于:pptx 文件的纯技术操作(拆分/合并/提取,用宿主 pptx skill)、演讲稿或文章本体的写作(用 wechat-article-writer)、需要先做研究的内容调研(先走 deep-research 再回来)、单张海报或 logo 设计。
examples/polar-night-ai-native/generate.js
View on GitHub// deck-studio 实战:AI Native 主题 · 极夜科技风格
// 教训应用:重锚点(粗黑标题+数据英雄)保留,细体只做对比;母题=节点网格+发光青线(每页≤2处光效)
const fs = require("fs");
const path = require("path");
const T = {
bg:"#0A0E1A", card:"#141B2E", card2:"#1C2333", line:"#232B42", line2:"#2E3852",
txt:"#E6EDF3", dim:"#8B94A7", faint:"#525B70",
cyan:"#00D4FF", violet:"#7B61FF",
sans:'"PingFang SC","Hiragino Sans GB",sans-serif',
mono:'"SF Mono","DIN Alternate",monospace',
};
const CSS = `
*{margin:0;padding:0;box-sizing:border-box;-webkit-font-smoothing:antialiased;}
html,body{width:1280px;height:720px;overflow:hidden;}
body{background:${T.bg};color:${T.txt};font-family:${T.sans};position:relative;
background-image:linear-gradient(${T.card}55 1px,transparent 1px),linear-gradient(90deg,${T.card}55 1px,transparent 1px);
background-size:48px 48px;}
.mono{font-family:${T.mono};font-feature-settings:"tnum";}
.top{position:absolute;left:88px;right:88px;top:50px;display:flex;justify-content:space-between;align-items:center;z-index:5;}
.kick{font-size:11px;letter-spacing:.34em;color:${T.dim};font-weight:600;text-transform:uppercase;display:flex;align-items:center;gap:12px;}
.kick::before{content:"";width:7px;height:7px;background:${T.cyan};display:inline-block;box-shadow:0 0 10px ${T.cyan};}
.brand{font-size:10.5px;letter-spacing:.3em;color:${T.faint};}
.foot{position:absolute;left:88px;right:88px;bottom:42px;display:flex;justify-content:space-between;align-items:center;z-index:5;}
.src{font-size:10px;letter-spacing:.14em;color:${T.faint};text-transform:uppercase;}
.pg{font-size:11px;color:${T.dim};letter-spacing:.14em;}
.stage{position:absolute;left:88px;right:88px;top:118px;bottom:100px;z-index:2;display:flex;flex-direction:column;}
/* 重锚点字阶(v4教训:标题回归 800,细体只做对比层) */
.h{font-weight:800;letter-spacing:.005em;line-height:1.22;}
.thin{font-weight:200;}
.lead{font-weight:300;color:${T.dim};line-height:1.7;}
.body{font-weight:400;color:#B8C0D0;line-height:1.62;}
.cy{color:${T.cyan};} .vi{color:${T.violet};}
.mark{width:44px;height:4px;background:${T.cyan};box-shadow:0 0 14px ${T.cyan}88;}
.hair{height:1px;background:${T.line};}
.card{background:${T.card};border:1px solid ${T.line};}
.ghost{position:absolute;font-family:${T.mono};font-weight:800;color:#10162A;z-index:0;line-height:.8;letter-spacing:-.04em;}
`;
function P(section, inner, no, opt={}) {
return `<!doctype html><meta charset="utf-8"><style>${CSS}</style><body>
<div class="top"><span class="kick">${section}</span><span class="brand">AI NATIVE · GEEK SKILLS</span></div>
${opt.ghost||""}
<div class="stage">${inner}</div>
<div class="foot"><span class="src">keqian-method × xuefeng-method</span><span class="pg mono">${String(no).padStart(2,"0")} — 09</span></div>
</body>`;
}
const pages = [];
/* 1 封面:粗黑大标题(锚点回归)+ 节点网络母题 */
pages.push(P("AI Native · Cover", `
<!-- 大尺度几何母题:同心环 + 主节点,与 p7 的巨型水印同语言 -->
<svg data-deck-background-bleed="true" aria-hidden="true" focusable="false"
style="position:absolute;right:-140px;top:30px;width:640px;height:640px;z-index:0;pointer-events:none;"
viewBox="0 0 640 640">
<circle cx="320" cy="320" r="290" fill="none" stroke="${T.line}" stroke-width="1.5"/>
<circle cx="320" cy="320" r="210" fill="none" stroke="${T.line2}" stroke-width="1.5"/>
<circle cx="320" cy="320" r="130" fill="none" stroke="${T.line2}" stroke-width="2"/>
<line x1="320" y1="30" x2="320" y2="610" stroke="${T.line}" stroke-width="1"/>
<line x1="30" y1="320" x2="610" y2="320" stroke="${T.line}" stroke-width="1"/>
<circle cx="320" cy="190" r="10" fill="${T.cyan}" style="filter:drop-shadow(0 0 14px ${T.cyan})"/>
<circle cx="450" cy="320" r="6" fill="${T.violet}"/>
<circle cx="320" cy="450" r="6" fill="${T.line2}"/>
<rect x="312" y="312" width="16" height="16" fill="${T.txt}"/>
</svg>
<div style="flex:1;display:flex;flex-direction:column;justify-content:center;max-width:860px;">
<div class="mark" style="margin-bottom:38px;"></div>
<div class="h" style="font-size:76px;">AI Native:</div>
<div class="h" style="font-size:76px;">当核心决策<span class="cy">交给模型</span></div>
<div class="lead" style="margin-top:34px;font-size:17px;">两条经过实战的方法论路线 · 克谦式 × 雪峰式</div>
</div>
<div class="hair" style="margin-bottom:18px;"></div>
<div style="display:flex;gap:48px;align-items:baseline;">
<span class="mono" style="font-size:14px;font-weight:600;color:${T.txt};">2026.07</span>
<span class="lead" style="font-size:12px;letter-spacing:.08em;">GEEK SKILLS 方法论分享</span>
</div>`, 1));
/* 2 目录 */
pages.push(P("目录 · Agenda", `
<div class="h" style="font-size:42px;margin-bottom:10px;">三个问题</div>
<div class="lead" style="font-size:15px;">判断 → 方法 → 选择</div>
<div style="flex:1;display:flex;flex-direction:column;justify-content:center;">
${[["01","是什么","AI Native 与 +AI 的分界线在哪",1],
["02","怎么做","封闭走克谦式,开放走雪峰式",0],
["03","怎么选","第零步判型,选错方法比没有方法更危险",0]].map(([n,t,d,hot])=>`
<div class="hair"></div>
<div style="display:flex;align-items:center;padding:27px 4px;">
<span class="mono" style="font-size:48px;font-weight:200;width:140px;color:${hot?T.cyan:T.faint};">${n}</span>
<span class="h" style="font-size:25px;width:250px;font-weight:700;">${t}</span>
<span class="lead" style="font-size:14.5px;flex:1;">${d}</span>
</div>`).join("")}
<div class="hair"></div>
</div>`, 2));
/* 3 问题:分界线(左右分屏对比,重锚点) */
pages.push(P("01 是什么 · 分界", `
<div class="h" style="font-size:44px;">分界线只有一条:<span class="cy">用户行为能否穷举</span></div>
<div style="flex:1;display:flex;gap:1px;background:${T.line};margin-top:36px;">
<!-- 三段网格(消灭空腹):① 标题头 ② 图形居中占中段固定高 ③ 正文贴中段 + 结论行 -->
<div style="flex:1;background:${T.card};border-top:2px solid ${T.faint};padding:36px 40px;display:flex;flex-direction:column;">
<div style="margin-bottom:8px;"><span class="mono" style="font-size:13px;letter-spacing:.2em;color:${T.dim};">TYPE A</span>
<div class="h" style="font-size:29px;margin-top:12px;">+AI<span class="thin" style="color:${T.dim};font-size:21px;">(场景依赖型)</span></div></div>
<div style="flex:1;display:flex;align-items:center;justify-content:center;">
<div style="display:grid;grid-template-columns:repeat(8,26px);gap:9px;">
${Array.from({length:24},()=>`<div style="width:26px;height:26px;background:${T.line2};"></div>`).join("")}
</div>
</div>
<div class="body" style="font-size:14.5px;margin-bottom:16px;">行为可枚举——有限的格栅。AI 辅助执行<span style="color:${T.txt};font-weight:600;">确定性流程</span>:记账、OCR 归档、CRUD 后台。</div>
<div class="mono" style="font-size:13px;color:${T.dim};border-top:1px solid ${T.line};padding-top:16px;">→ 克谦式:单 Agent · SDD · 质量门禁</div>
</div>
<div style="flex:1;background:${T.card2};border-top:2px solid ${T.cyan};padding:36px 40px;display:flex;flex-direction:column;">
<div style="margin-bottom:8px;"><span class="mono" style="font-size:13px;letter-spacing:.2em;color:${T.cyan};">TYPE B</span>
<div class="h" style="font-size:29px;margin-top:12px;">AI Native<span class="thin" style="color:${T.dim};font-size:21px;">(强模型依赖型)</span></div></div>
<div style="flex:1;display:flex;align-items:center;justify-content:center;">
<svg width="280" height="130" viewBox="0 0 280 130">
<circle cx="40" cy="65" r="9" fill="${T.cyan}" style="filter:drop-shadow(0 0 10px ${T.cyan})"/>
${[[120,18],[160,50],[135,92],[200,30],[230,72],[195,112],[260,16],[268,58],[250,104]].map(([x,y])=>
`<line x1="40" y1="65" x2="${x}" y2="${y}" stroke="${T.line2}" stroke-width="1"/><circle cx="${x}" cy="${y}" r="4" fill="${T.line2}"/>`).join("")}
</svg>
</div>
<div class="body" style="font-size:14.5px;margin-bottom:16px;">行为开放——发散的节点。AI 驱动<span class="cy" style="font-weight:600;">核心决策</span>:AI 日历、对话助手、智能推荐。</div>
<div class="mono" style="font-size:13px;color:${T.cyan};border-top:1px solid ${T.line2};padding-top:16px;">→ 雪峰式:多专精 Agent · 快速校准 · 行为审计</div>
</div>
</div>`, 3));
/* 4 第零步判型表 */
pages.push(P("01 是什么 · 判型", `
<div class="h" style="font-size:42px;">第零步:先判型,再选法</div>
<div class="lead" style="font-size:15px;margin-top:10px;">选错方法论,比没有方法论更危险</div>
<div style="flex:1;display:flex;flex-direction:column;justify-content:center;">
<table style="width:100%;border-collapse:collapse;">
<tr>${["产品类型","关键判断","推荐方法"].map((h,i)=>`<th style="text-align:left;font-size:12px;letter-spacing:.16em;color:${i===2?T.cyan:T.dim};font-weight:600;text-transform:uppercase;padding-bottom:16px;border-bottom:1px solid ${T.line2};">${h}</th>`).join("")}</tr>
${[["+AI 场景依赖","能列出所有合法输入输出组合","克谦式(单 Agent + SDD)",0],
["AI Native 强模型依赖","用户的下一步操作无法预测","雪峰式(多专精 Agent)",1],
["混合型","核心流程确定,部分环节开放","两者结合,按模块选用",0]].map(([a,b,c,hot])=>`
<tr><td class="h" style="font-size:19px;font-weight:700;padding:24px 24px 24px 0;border-bottom:1px solid ${T.line};">${a}</td>
<td class="body" style="font-size:15px;padding:24px 24px 24px 0;border-bottom:1px solid ${T.line};">${b}</td>
<td class="body" style="font-size:15px;padding:24px 0;border-bottom:1px solid ${T.line};${hot?`color:${T.cyan};font-weight:600;`:""}">${c}</td></tr>`).join("")}
</table>
</div>`, 4));
/* 5 两条路线对比(双卡,重锚点) */
pages.push(P("02 怎么做 · 路线", `
<div class="h" style="font-size:42px;">两条路线,各管一半世界</div>
<div style="flex:1;display:flex;gap:24px;margin-top:34px;">
${[["克谦式","确定性交付","单 Agent 极致论","文档驱动开发 SDD","质量门禁闭环","概率乘:环节越少越可靠",T.violet,"KEQIAN"],
["雪峰式","开放性生存","多养专精虾:每只只干一种活","唯快不破:快速上线快速校准","行为审计替代质量门禁","穷举是死循环",T.cyan,"XUEFENG"]].map(([name,tag,p1,p2,p3,p4,color,en])=>`
<div class="card" style="flex:1;padding:36px;display:flex;flex-direction:column;border-top:2px solid ${color};">
<div style="display:flex;justify-content:space-between;align-items:baseline;margin-bottom:6px;">
<span class="h" style="font-size:28px;">${name}</span>
<span class="mono" style="font-size:11px;letter-spacing:.2em;color:${T.faint};">${en}</span></div>
<div class="mono" style="font-size:13px;color:${color};margin-bottom:24px;">${tag}</div>
${[p1,p2,p3,p4].map((p,pi)=>`<div style="display:flex;gap:14px;padding:12px 0;border-top:1px solid ${T.line};align-items:baseline;">
<span class="mono" style="color:${color};font-weight:600;font-size:13px;min-width:24px;">0${pi+1}</span><span class="body" style="font-size:14.5px;">${p}</span></div>`).join("")}
</div>`).join("")}
</div>`, 5));
/* 6 雪峰三原则(三卡片,一个点亮) */
pages.push(P("02 怎么做 · 心法", `
<div class="h" style="font-size:42px;">开放世界的三条生存法则</div>
<div class="lead" style="font-size:15px;margin-top:10px;">来自雪峰式方法论 · AI 日历等产品的实战沉淀</div>
<div style="flex:1;display:grid;grid-template-columns:1fr 1fr 1fr;gap:24px;margin-top:32px;">
${[["穷举是死循环","试图列出所有用户行为,是在和无限对赌。接受不可穷举,设计兜底。",0],
["多养专精虾","每只虾只干一种活:路由、执行、校验、表达分开,出错能定位,升级不牵连。",1],
["唯快不破","校准到 95% 再上线 = 永远不上线。快速上线,用行为审计追漂移,边跑边修。",0]].map(([t,d,hot],i)=>`
<div class="card" style="padding:34px;display:flex;flex-direction:column;${hot?`border-color:${T.cyan};box-shadow:0 0 24px ${T.cyan}22;`:""}">
<span class="mono" style="font-size:40px;font-weight:200;color:${hot?T.cyan:T.faint};">0${i+1}</span>
<div class="h" style="font-size:22px;margin:18px 0 14px;">${t}</div>
<div class="body" style="font-size:14px;flex:1;">${d}</div>
</div>`).join("")}
</div>`, 6));
/* 7 数据页:概率乘(数据英雄,占屏大) */
pages.push(P("02 怎么做 · 铁律", `
<div class="ghost mono" style="right:-20px;bottom:-50px;font-size:280px;">×</div>
<div class="h" style="font-size:42px;">克谦铁律:概率乘</div>
<div style="flex:1;display:flex;align-items:center;gap:56px;">
<div>
<div style="display:flex;align-items:baseline;">
<span class="mono" style="font-size:120px;font-weight:800;line-height:.9;">0.99</span
><span class="mono" style="font-size:48px;font-weight:300;color:${T.dim};position:relative;top:-44px;margin-left:4px;">51</span>
<span class="mono" style="font-size:56px;font-weight:200;color:${T.dim};margin:0 24px;">=</span>
<span class="mono" style="font-size:120px;font-weight:800;line-height:.9;color:${T.cyan};text-shadow:0 0 30px ${T.cyan}55;">0.59</span>
</div>
<div class="lead" style="font-size:16px;margin-top:30px;max-width:900px;line-height:1.85;">
每个环节 99% 可靠,51 个环节相乘后只剩 59 分——不及格。<br>
<span style="color:${T.txt};font-weight:600;">所以:链条越短越好,每个环节都要过门禁,能用确定性代码就不用模型。</span></div>
</div>
</div>
<div style="display:flex;gap:64px;">
${[["≤5","关键链路环节数目标"],["100%","每环节过质量门禁"],["0","能写 if 就不调模型的例外"]].map(([n,c])=>`
<div style="flex:1;"><div class="hair" style="margin-bottom:14px;"></div>
<span class="mono h" style="font-size:40px;">${n}</span>
<div class="body" style="font-size:13px;margin-top:6px;color:${T.dim};">${c}</div></div>`).join("")}
</div>`, 7));
/* 8 行动三步 */
pages.push(P("03 怎么选 · 行动", `
<div class="h" style="font-size:42px;">落地三步</div>
<div style="flex:1;display:flex;flex-direction:column;justify-content:center;">
${[["判型","用第零步表格给产品定性:+AI / AI Native / 混合","本周就能做"],
["选法","封闭模块走克谦式流水线,开放模块交给雪峰式专精 Agent","架构评审一次定"],
["上线校准","AI Native 部分不等 95%,上线 + 行为审计 + 快速迭代","每周一个校准循环"]].map(([t,d,tag],i)=>`
<div class="hair"></div>
<div style="display:flex;align-items:center;padding:24px 4px;">
<span class="mono" style="width:110px;font-size:17px;font-weight:600;color:${T.cyan};">STEP ${i+1}</span>
<span class="h" style="width:200px;font-size:22px;font-weight:700;">${t}</span>
<span class="body" style="flex:1;font-size:14.5px;">${d}</span>
<span class="mono" style="font-size:12px;color:${T.faint};">${tag}</span></div>`).join("")}
<div class="hair"></div>
</div>`, 8));
/* 9 封底 */
pages.push(P("AI Native", `
<div class="ghost mono" style="left:-20px;bottom:-80px;font-size:330px;">→</div>
<div style="flex:1;display:flex;flex-direction:column;justify-content:center;">
<div class="mark" style="margin-bottom:36px;"></div>
<div class="h" style="font-size:64px;">穷举是死循环,</div>
<div class="h" style="font-size:64px;"><span class="cy">唯快不破</span>。</div>
<div class="lead" style="font-size:16px;margin-top:36px;">完整方法论:keqian-method · xuefeng-method<br>
<span class="mono" style="font-size:13px;">github.com/staruhub/ClaudeSkills</span></div>
</div>`, 9));
const outputDir = path.join(__dirname, "html");
fs.mkdirSync(outputDir,{recursive:true});
pages.forEach((h,i)=>fs.writeFileSync(path.join(outputDir, `p${i+1}.html`),h));
console.log(pages.length+" 页 AI Native deck 生成");
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| canvas-designanthropics/skillsCreate beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| docxanthropics/skillsUse this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pdfanthropics/skillsUse this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pptxanthropics/skillsUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates (.potx), layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx or .potx filename, regardless of what they plan to do with the content afterward. If a .pptx or .potx file needs to be opened, created, or touched, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |