菜单
Use when auditing or implementing technical SEO — semantic HTML, meta tags, structured data, canonical URLs, sitemap, robots.txt, page speed as a ranking factor, and accessible content. Trigger when the user mentions SEO, search engine, meta tags, structured data, sitemap, or crawlability.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use when defining or auditing a product's written voice — the consistent personality across all UI copy. Covers voice vs tone, voice attribute definition, do/don't tables, and how to keep AI-generated copy on-brand. Trigger when the user mentions brand voice, tone of voice, writing style guide, copy guidelines, or product personality.
Use when auditing or reducing JavaScript and CSS bundle sizes, setting performance budgets, choosing lighter alternatives to heavy libraries, configuring tree shaking, code splitting, and lazy loading. Trigger when the user mentions bundle size, JavaScript size, tree shaking, code splitting, lazy loading, or performance budget.
Use when designing or auditing for color vision deficiency — ensuring that color is never the only visual signal, checking charts and status indicators, picking colorblind-safe palettes, and testing against deuteranopia, protanopia, and tritanopia simulations. Trigger when the user mentions color blindness, color vision, CVD, deuteranopia, protanopia, accessibility color, or color-only signal.
Use when planning content hierarchy, information density, content-first design, microcopy systems, or editorial patterns. Covers the content-before-design principle, content inventory, progressive disclosure, scannability, and content governance. Trigger when the user mentions content strategy, content hierarchy, information architecture, microcopy, or editorial design.
Use when designing or auditing cursor behavior — matching cursor shape to affordance, custom cursors, drag cursors, resize cursors, and the hover+cursor pairing rule. Trigger when the user mentions cursor, pointer, mouse cursor, drag cursor, or resize handle.
Use when designing, auditing, or implementing dark mode. Covers lightness inversion (not hue shift), elevation via lightness not shadow, surface stacking, contrast rechecking, and the prefers-color-scheme media query. Trigger when the user mentions dark mode, dark theme, night mode, theme toggle, or color scheme.
| name | canon-seo |
| description | Use when auditing or implementing technical SEO — semantic HTML, meta tags, structured data, canonical URLs, sitemap, robots.txt, page speed as a ranking factor, and accessible content. Trigger when the user mentions SEO, search engine, meta tags, structured data, sitemap, or crawlability. |
Technical SEO is about making content findable and understandable by search engines. Not tricks — structure.
Search engines understand semantic HTML natively. Every rule in canon-accessibility that improves screen reader experience also improves SEO.
<h1> once per page, matching the page topic.<nav>, <main>, <article>, <aside> help crawlers understand page layout.<a href> for links (not <div onClick>).<img alt> describes the image for image search.<head>
<title>Page Title — Site Name</title>
<meta name="description" content="Concise summary under 155 characters.">
<link rel="canonical" href="https://example.com/page">
<meta name="robots" content="index, follow">
</head>
<title>: 50–60 characters. Unique per page. Primary keyword near the front.<meta description>: 120–155 characters. Compelling summary. Unique per page.<link rel="canonical">: prevents duplicate content penalties.<meta robots>: index, follow is default. noindex for admin pages.<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "Page Title",
"author": { "@type": "Person", "name": "Author" },
"datePublished": "2026-04-18"
}
</script>
Use for: articles, products, FAQs, breadcrumbs, organizations, events. Google uses structured data for rich results (stars, prices, FAQ dropdowns).
Core Web Vitals are a Google ranking signal. See canon-performance for targets:
robots.txt: allow crawlers access to important pages, block admin/API routes.sitemap.xml: list all public pages with <lastmod> dates.| Anti-pattern | Why it fails |
|---|---|
Same <title> on every page | Search engines can't distinguish pages |
No <meta description> | Search engine generates one (often badly) |
| Keyword stuffing | Penalized by modern search engines |
| Client-rendered content without SSR/SSG | Crawlers may not execute JS |
| No canonical URL | Duplicate content across www/non-www, http/https |
| Blocking CSS/JS in robots.txt | Crawlers can't render the page |
<title> per page, 50–60 characters<meta description> per page, 120–155 characters<link rel="canonical"> on every page<h1> per page matching the topicalt textsitemap.xml submitted to search consolesrobots.txt allows crawling of public pagescanon-performance for Core Web Vitalscanon-accessibility for semantic HTML