article-page-generator by kostja94/marketing-skills
npx skills add https://github.com/kostja94/marketing-skills --skill article-page-generator指导单篇文章页面的结构、SEO 和用户体验——包括布局、元数据、结构化数据和技术要求。关于文章正文内容(引言、正文、结论、写作),请参阅 article-content。区别于 blog-page-generator,后者涵盖博客索引/列表页。
调用时机:在首次使用时,如果有助于理解,可以用 1-2 句话开头说明此技能涵盖的内容及其重要性,然后提供主要输出。在后续使用或用户要求跳过时,直接进入主要输出。
输出工作流程:始终按顺序输出:0. 研究阶段(关键词、搜索意图、竞争对手)→ 1. 意图分析 → 2. 内容分析 → 3. 建议。不要跳过任何步骤。当通过网页搜索执行研究阶段时,显示搜索结果和发现。
文章分析和创建基于四项输入。在输出建议前,收集或推断它们:
| 输入 | 目的 | 来源 |
|---|---|---|
| 产品 | 产品关联、功能、用例、行动号召放置位置 | project-context(第 1-4 节,9-11 节);文章内容;网页搜索 |
| 关键词 | 目标关键词、主要/次要放置位置 | project-context 第 6 节;keyword-research;文章 |
广告位招租
在这里展示您的产品或服务
触达数万 AI 开发者,精准高效
| 信息型、商业型、交易型、导航型;驱动结构、行动号召、SEO 深度 |
| project-context 第 6 节(目标意图);文章导向;内容类型 |
| 竞争对手文章 | 可借鉴的结构、内容差距、长度目标、关键词机会 | 用户提供的 URL;project-context 第 11 节;网页搜索 |
当任何输入缺失时:主动询问或搜索。对于文章分析:默认执行研究阶段(关键词搜索、搜索意图、竞争对手文章)——参见研究阶段部分。对于产品/关键词/意图,从文章推断或提示用户添加 project-context。
1. 产品 / 公司上下文
利用现有上下文提供定制化分析:
| 来源 | 用于 |
|---|---|
| project-context.md | 关键词(第 6 节)、竞争对手(第 7 节)、内容策略(第 11 节)、产品关联 |
| 文章内容 | 提取产品名称、功能、URL;推断目标关键词和受众 |
| 网页搜索 | 当分析已知品牌时:搜索“[产品] 功能”、“[产品] vs 竞争对手”、公司定位——用于验证产品关联、建议缺失的功能/用例,并改进竞争对手差距分析 |
如果不存在 project-context,则从文章推断,并可选择搜索公司/产品信息以丰富建议。
为文章分析进行轻量级研究。当分析或审核文章时,执行搜索并在第 0 节输出结果。仅当用户明确要求跳过时才跳过(例如“跳过搜索”)。
输出格式:参见下方输出格式第 0 节。
/blog/how-to-optimize-seo)首先检查项目上下文: 如果存在 .claude/project-context.md 或 .cursor/project-context.md,请阅读它以了解主题、受众、关键词和第 11 节(内容/博客/文章策略)。
识别:
产品关联内容:文章应与产品关联(解决的问题、功能、用例)。避免完全通用、与产品无关的内容。在结论中或上下文合适时,自然地链接到产品/功能页面。
根据导向选择结构、SEO 深度和结构化数据。完整的文章导向(融资/公关、产品更新、指南、新闻、常青内容)、SEO 驱动与非 SEO 驱动、常青与时效性,请参见 content-marketing。
意图分析输出:导向、主要目标、SEO 与非 SEO、常青与时效性——参见输出格式第 1 节。
| 部分 | 目的 |
|---|---|
| Hero/Header | 标题(H1)、作者、单一日期(参见 schema-markup 日期显示以提升点击率)、阅读时间(字数 ÷ 200;向上取整)、特色图片、分享按钮 |
| TL;DR 或关键要点 | 内容参见 article-content;置于引言之后;支持 GEO/AI 引用 |
| 引言 | 参见 article-content 了解钩子、长度、关键词放置 |
| 正文 | 参见 article-content 了解 QAE、段落长度、可扫描性 |
| 结论 | 参见 article-content 了解总结、行动号召、产品关联 |
| 相关文章 | 3-6 个上下文链接;文末推荐 |
| 作者简介 | E-E-A-T;资历、照片、作者页链接——参见 eeat-signals |
参见 image-optimization(文章 / 博客 hero)。结构化数据、Open Graph、Twitter Cards 使用同一张图片;最小宽度 1200px,绝对 URL。参见 open-graph、twitter-cards。
参见 article-content 了解 TL;DR、关键要点、QAE 模式、答案优先;完整的 GEO 策略参见 generative-engine-optimization。
| 元素 | 指南 |
|---|---|
| 标题 | 55 个字符;主要关键词靠近开头;使用强力词 |
| 元描述 | 150–160 个字符;包含行动号召;主要关键词 |
| H1 | 每页一个;与标题匹配;自然地包含主要关键词 |
参见 article-content 了解可读性、深度、原创性、按类型的字数。E-E-A-T:作者简介、引用、更新日志、专家引述——参见 eeat-signals。
使用 url-slug-generator 创建 slug。关键规则:
/blog/ai-people-search 而非 /blog/ai-search-engine-finding-people-speed-discovery-outreach/blog/2025/01/15/article-title);复制粘贴完整标题参见 schema-markup(日期显示以提升点击率):仅显示一个可见日期;优先使用 dateModified。
参见 schema-markup 了解 Article/BlogPosting/NewsArticle 类型选择、必需属性、JSON-LD 示例和日期显示。使用 Rich Results Test 进行验证。
对文章页面使用 og:type: article(而非 website):
<meta property="og:type" content="article">
<meta property="og:article:published_time" content="2025-01-15T09:00:00Z">
<meta property="og:article:modified_time" content="2025-02-01T14:30:00Z">
<meta property="og:article:author" content="https://example.com/author/jane">
| 元素 | 指南 |
|---|---|
| 数量 | 正文中 3-5 个上下文链接 + 相关文章中 3-6 个 = 每篇文章总计 6-11 个 |
| 第一段 | 1 个链接指向支柱或关键相关内容 |
| 正文 | 2-4 个上下文链接;每个主要部分在相关时放置一个 |
| 相关文章 | 3-6 个文末链接;同一主题集群 |
| 锚文本 | 描述性(例如“2025 年 SEO 清单”、“如何优化元标签”);避免“点击此处”、“了解更多”、“阅读更多” |
| 变化 | 混合精确匹配、部分匹配、品牌锚文本;避免过度优化 |
| 防止孤立 | 每篇文章至少有 ≥1 个来自中心/支柱或导航的内部链接 |
| 元素 | 指南 |
|---|---|
| 数量 | 每篇文章 2-5 个外部链接;引用权威来源 |
| 使用时机 | 统计数据、研究、定义、工具比较、专家引述 |
| 锚文本 | 描述性(例如“Google 搜索质量指南”、“SEO 研究”);链接到来源 |
| 同一 URL | 每页的链接权益只计算一次;无需重复 |
| E-E-A-T | 指向信誉良好来源的外部链接传递信任信号——参见 eeat-signals |
引用格式参见 article-content;E-E-A-T 及何时包含引用参见 eeat-signals。
AI 辅助内容指南参见 article-content;E-E-A-T 参见 eeat-signals。
审核或优化文章时,应用内容审核清单。完整维度参见 article-content。
分析或审核文章时,在意图分析之前输出此部分。包括搜索来源和发现。如果用户要求跳过搜索,注明并仅从文章推断。
| 部分 | 输出 |
|---|---|
| 关键词搜索 | 主要关键词(来自文章或搜索)、次要关键词、关键词机会(来自搜索结果页/竞争对手分析)。如果执行了搜索:使用的查询、观察到的顶部结果。 |
| 搜索意图 | 主要关键词的意图(信息型/商业型/交易型/导航型)、2-3 个次要关键词的意图、文章内容是否匹配意图。如果执行了搜索:观察到的搜索结果页摘要类型。 |
| 竞争对手文章 | 如果搜索了:2-3 个 URL、简要结构(字数、H2 数量)、内容差距、长度目标。如果用户提供了 URL:同上。完整方法参见 competitor-research。如果跳过:“竞争对手分析已跳过。” |
在任何建议之前,输出简要分析:
| 维度 | 输出 |
|---|---|
| 导向 | 融资/公关、产品更新、指南、新闻、常青内容 |
| 主要目标 | 品牌、公关、教育、产品采用、自然流量、… |
| SEO 与非 SEO | SEO 驱动 / 非 SEO 驱动 / 混合 |
| 常青与时效性 | 常青 / 时效性 |
| 影响 | 1-2 句话:例如“SEO 优先级低 → 关注清晰度、可分享性”或“SEO 驱动 → 完整关键词 + GEO 优化” |
应用上述内容分析表。按维度输出简要评估(✅ / ⚠️ / ❌ + 一行说明)。
为每个项目分配优先级:P0(关键)、P1(高)、P2(中)、P3(锦上添花)。输出为带优先级前缀的表格或列表。
| 优先级 | 使用时机 |
|---|---|
| P0 | 阻碍 GEO/SEO;缺少核心元素(TL;DR 或关键要点、关键词在前 100 词、结构化数据) |
| P1 | 对流量、点击率或转化有显著影响(标题长度、分享按钮、行动号召) |
| P2 | 改善用户体验或权威性(相关文章、作者简介、内部链接) |
| P3 | 优化完善(图片优化、可读性微调) |
示例:[P0] 添加 TL;DR 或关键要点 —— GEO,AI 引用
每周安装量
209
仓库
GitHub 星标
239
首次出现
2026 年 3 月 1 日
安全审计
安装于
kimi-cli191
cursor191
gemini-cli190
codex190
github-copilot190
opencode190
Guides structure, SEO, and UX for individual article pages — layout, metadata, schema, technical. For article body content (intro, body, conclusion, writing), see article-content. Distinct from blog-page-generator , which covers the blog index/listing page.
When invoking : On first use , if helpful, open with 1–2 sentences on what this skill covers and why it matters, then provide the main output. On subsequent use or when the user asks to skip, go directly to the main output.
Output workflow : Always output in order: 0. Research Phase (keywords, search intent, competitors) → 1. Intent Analysis → 2. Content Analysis → 3. Recommendations. Do not skip steps. When Research Phase was performed via web search, show the search results and findings.
Article analysis and creation rest on four inputs. Gather or infer them before outputting recommendations:
| Input | Purpose | Source |
|---|---|---|
| Product | Product connection, features, use cases, CTA placement | project-context (Sections 1–4, 9–11); article content; web search |
| Keywords | Target keyword, primary/secondary placement | project-context Section 6; keyword-research; article |
| Article intent | Informational, commercial, transactional, navigational; drives structure, CTA, SEO depth | project-context Section 6 (target intent); article orientation; content type |
| Competitor articles | Structure to adopt, content gaps, length target, keyword opportunities | User-provided URLs; project-context Section 11; web search |
When any input is missing : Proactively ask or search. For article analysis: perform Research Phase (keyword search, search intent, competitor articles) by default — see Research Phase section. For product/keywords/intent, infer from article or prompt user to add project-context.
1. Product / company context
Use available context to give tailored analysis:
| Source | Use for |
|---|---|
| project-context.md | Keywords (Section 6), competitors (Section 7), content strategy (Section 11), product connection |
| Article content | Extract product name, features, URLs; infer target keyword and audience |
| Web search | When analyzing a known brand: search for "[product] features", "[product] vs competitors", company positioning — use to validate product connection, suggest missing features/use cases, and improve competitor gap analysis |
If no project-context exists, infer from the article and optionally search for company/product info to enrich recommendations.
Lightweight research for article analysis. When analyzing or auditing an article, perform searches and output the results in Section 0. Skip only if user explicitly asks to skip (e.g. "skip search").
Output format : See Output Format Section 0 below.
/blog/how-to-optimize-seo)Check for project context first: If .claude/project-context.md or .cursor/project-context.md exists, read it for topics, audience, keywords, and Section 11 (Content/Blog/Article Strategy).
Identify:
Product-linked content : Articles should tie to the product (problem it solves, features, use cases). Avoid purely generic content with no product relevance. Link to product/feature pages naturally in conclusion or when context fits.
Choose structure, SEO depth, and schema based on orientation. See content-marketing for full Article Orientations (Funding/PR, Product update, Guide, News, Evergreen), SEO-driven vs non-SEO-driven, Evergreen vs Timely.
Intent Analysis output : Orientation, primary goal, SEO vs non-SEO, Evergreen vs timely — see Output Format Section 1.
| Section | Purpose |
|---|---|
| Hero/Header | Title (H1), author, single date (see schema-markup Date display for CTR), reading time (word count ÷ 200; round up), featured image, share buttons |
| TL;DR or Key Takeaways | See article-content for content; placed after intro; supports GEO/AI citation |
| Introduction | See article-content for hook, length, keyword placement |
| Body | See article-content for QAE, paragraph length, scannability |
| Conclusion | See article-content for summary, CTA, product connection |
| Related posts | 3–6 contextual links; end-of-article recommendations |
| Author bio | E-E-A-T; credentials, photo, link to author page — see eeat-signals |
See image-optimization (Article / Blog hero). Same image for Schema, Open Graph, Twitter Cards; min 1200px wide, absolute URL. See open-graph , twitter-cards.
See article-content for TL;DR, Key Takeaways, QAE pattern, answer-first; generative-engine-optimization for full GEO strategy.
| Element | Guideline |
|---|---|
| Title | 55 chars; primary keyword near start; power words |
| Meta description | 150–160 chars; CTA; primary keyword |
| H1 | One per page; matches title; primary keyword naturally |
See article-content for readability, depth, originality, word count by type. E-E-A-T : Author bio, citations, changelog, expert quotes — see eeat-signals.
Use url-slug-generator for slug creation. Key rules:
/blog/ai-people-search not /blog/ai-search-engine-finding-people-speed-discovery-outreach/blog/2025/01/15/article-title); copy-pasting full titleSee schema-markup (Date display for CTR): show only one visible date; prefer dateModified.
See schema-markup for Article/BlogPosting/NewsArticle type selection, required properties, JSON-LD example, and date display. Validate with Rich Results Test.
Use og:type: article for article pages (not website):
<meta property="og:type" content="article">
<meta property="og:article:published_time" content="2025-01-15T09:00:00Z">
<meta property="og:article:modified_time" content="2025-02-01T14:30:00Z">
<meta property="og:article:author" content="https://example.com/author/jane">
| Element | Guideline |
|---|---|
| Volume | 3–5 contextual links in body + 3–6 in Related posts = 6–11 total per article |
| First paragraph | 1 link to pillar or key related content |
| Body | 2–4 contextual links; one per major section when relevant |
| Related posts | 3–6 end-of-article links; same topic cluster |
| Anchor text | Descriptive (e.g. "SEO checklist for 2025", "how to optimize meta tags"); avoid "click here", "learn more", "read more" |
| Variation | Mix exact-match, partial-match, branded anchors; avoid over-optimization |
| Orphan prevention | Every article has ≥1 internal link from hub/pillar or nav |
| Element | Guideline |
|---|---|
| Volume | 2–5 external links per article; cite authoritative sources |
| When to use | Statistics, research, definitions, tool comparisons, expert quotes |
| Anchor text | Descriptive (e.g. "Google's Search Quality Guidelines", "SEO study"); link to source |
| Same URL | Counts once per page for link equity; no need to repeat |
| E-E-A-T | External links to reputable sources signal trust — see eeat-signals |
See article-content for citation format; eeat-signals for E-E-A-T and when to include.
See article-content for AI-assisted content guidance; eeat-signals for E-E-A-T.
When auditing or optimizing an article, apply the Content Audit Checklist. See article-content for full dimensions.
When analyzing or auditing an article, output this section before Intent Analysis. Include search sources and findings. If user asked to skip search, note that and infer from article only.
| Section | Output |
|---|---|
| Keyword Search | Primary keyword (from article or search), secondary keywords, keyword opportunities (from SERP/competitor analysis). If search was performed: query used, top results observed. |
| Search Intent | Intent for primary keyword (Informational/Commercial/Transactional/Navigational), intent for 2–3 secondary keywords, whether article content matches intent. If search was performed: SERP snippet types observed. |
| Competitor Articles | If searched: 2–3 URLs, brief structure (word count, H2s), content gaps, length target. If user provided URLs: same. See competitor-research for full methodology. If skipped: "Competitor analysis skipped." |
Before any recommendations, output a brief analysis:
| Dimension | Output |
|---|---|
| Orientation | Funding/PR, Product update, Guide, News, Evergreen |
| Primary goal | Brand, PR, education, product adoption, organic traffic, … |
| SEO vs non-SEO | SEO-driven / Non-SEO-driven / Hybrid |
| Evergreen vs timely | Evergreen / Timely |
| Implications | 1–2 sentences: e.g. "Low SEO priority → focus on clarity, shareability" or "SEO-driven → full keyword + GEO optimization" |
Apply the Content Analysis table above. Output a brief assessment per dimension (✅ / ⚠️ / ❌ + one-line note).
Assign priority to each item: P0 (critical), P1 (high), P2 (medium), P3 (nice-to-have). Output as table or list with priority prefix.
| Priority | Use when |
|---|---|
| P0 | Blocks GEO/SEO; missing core element (TL;DR or Key Takeaways, keyword in first 100 words, schema) |
| P1 | Significant impact on traffic, CTR, or conversion (title length, share buttons, CTA) |
| P2 | Improves UX or authority (related posts, author bio, internal links) |
| P3 | Polish (image optimization, readability tweaks) |
Example: [P0] Add TL;DR or Key Takeaways — GEO, AI citation
Weekly Installs
209
Repository
GitHub Stars
239
First Seen
Mar 1, 2026
Security Audits
Gen Agent Trust HubPassSocketPassSnykWarn
Installed on
kimi-cli191
cursor191
gemini-cli190
codex190
github-copilot190
opencode190
README 国际化工具:自动化翻译与多语言文档管理 | readme-i18n
30,900 周安装
前端代码审计工具 - 自动化检测可访问性、性能、响应式设计、主题化与反模式
38,600 周安装
前端动画设计指南:提升用户体验的微交互与动效策略
38,600 周安装
跨平台设计适配指南:移动端、桌面端、平板、打印及邮件适配策略与实施方法
38,800 周安装
前端打磨(Polish)终极指南:提升产品细节与用户体验的系统化检查清单
39,900 周安装
Web应用测试指南:使用Python Playwright自动化测试本地Web应用
39,500 周安装
Azure Cloud Migrate:AWS Lambda到Azure Functions迁移工具 - 微软官方评估与代码迁移
38,700 周安装