Back to Details

seo-readiness

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

seo-readiness

seo-readiness

File existence is the easy 10%. The failures that cost launches are contradictions: canonical says localhost, sitemap advertises 
/admin
, a 
noindex
left over from staging, robots blocking the pages the sitemap lists. This skill reads the relationships.
文件存在只是最容易解决的 10%。真正会导致上线失败的是矛盾：canonical 指向 localhost，sitemap 公布了 
/admin
，遗留自 staging 的 
noindex
，或者 robots 屏蔽了 sitemap 列出的页面。此 skill 会读取并分析这些关系。

Activate when

何时启用

  • The product has public pages meant to be found, or launch/domain/redesign is near.
  • Titles look like "Vite + React", canonicals or sitemap hosts look wrong, or the site has never been indexed.
  • Not for a private/internal product (confirm it is consistently 
    noindex
    ), not for ranking strategy or keyword research.
  • 产品有需要被搜索引擎发现的公开页面，或即将上线、切换域名或重新设计。
  • 标题看起来像“Vite + React”，canonical 或 sitemap 主机看起来不对，或者网站从未被索引。
  • 不适用于私有/内部产品（确认其始终为 
    noindex
    ），也不适用于排名策略或关键词研究。

Inspect

检查

  1. Intent. Should this site be indexed? Production origin (scheme, host, 
    www
    )? Public vs private routes? Ask only if 
    .readyvibe/context.md
    and config do not say (
    site-reconnaissance
    ).
  2. Run the sweep (paths relative to this skill's folder; add 
    --render
    for client-rendered apps; serve a build locally if needed):
    bash
    node scripts/inspect-metadata.mjs --url <site> --render --private-path /dashboard
node scripts/check-links.mjs      --url <site> --render     # sitemap/canonical target status
    Read rendered HTML: frameworks can override tags after hydration. Compare with source (
    <Head>
    , 
    metadata
    exports, 
    next-seo
    , 
    react-helmet
    , 
    index.html
    ) to find where each value comes from, so fixes land in the right place.
  3. Relationships to reason about:
    Signal Avs. Signal BProblem
    canonical hostproduction originlocalhost/staging/preview host in a canonical (HIGH)
    sitemap 
    <loc>
    host    		canonical hostsearch engines get two truths
    sitemap lists pathpage is 
    noindex
    or private route    		contradictory signals; private exposure (HIGH for admin/dashboard/api)
    sitemap lists pathrobots blocks itunfetchable listing
    sitemap lists pathcanonical points elsewherenon-canonical URL listed
    sitemap lists pathreturns 404/redirect/soft-404dead listing (
    check-links
    )
    robots 
    Sitemap:
    		actual sitemap URL/hostwrong or dev sitemap advertised
    robots 
    Disallow: /
    		production intentwhole site blocked (HIGH)
    robots 
    Disallow: /_next/
    /assets/
    		renderingcrawlers cannot render
    robots lists 
    /admin
    		route reachable?robots is not access control; hands attackers a map
    noindex
    on home/most pages    		launch intentstaging leftover (HIGH)
    x-robots-tag
    header    		meta robotsheader may override; check both
    several pagesthe same canonical URL (
    CANONICAL_DUPLICATE_TARGET
    )    		template copied from the home page; tells engines to ignore the rest
    robots blocks a linked public page (
    ROBOTS_BLOCKS_PAGE
    )    		page's title, description, and 
    noindex
    are never read    		crawler cannot see it
    trailing slash / 
    www
    / 
    http
    variants    		canonical + redirectsduplicate URL forms
  4. Titles and descriptions: unique per page, specific to the page's content, front-load the topic, roughly 50-60 / 120-160 characters as a guideline (truncation, not a rule); no framework defaults; no keyword stuffing. Derive copy from the page's visible content and product facts; do not invent claims.
  5. Multi-language: 
    hreflang
    pairs, per-language canonicals (with 
    multilingual-readiness
    ).
  6. JS-only pages: if content appears only after JavaScript and no server rendering or prerender exists, say indexing depends on the crawler executing JS: SOURCE-INDICATED risk, not proven.
  1. 意图。 该网站是否应该被索引？生产环境源站（协议、主机、
    www
    ）是什么？哪些是公开路由，哪些是私有路由？仅当 
    .readyvibe/context.md
    和配置未说明时再询问（
    site-reconnaissance
    ）。
  2. 运行全面检查（路径相对于此 skill 所在文件夹；对于客户端渲染的应用添加 
    --render
    ；如有需要，在本地提供构建产物）：
    bash
    node scripts/inspect-metadata.mjs --url <site> --render --private-path /dashboard
node scripts/check-links.mjs      --url <site> --render     # sitemap/canonical target status
    读取渲染后的 HTML：框架可能会在 hydration 后覆盖标签。将结果与源码（
    <Head>
    、
    metadata
    exports、
    next-seo
    、
    react-helmet
    、
    index.html
    ）对比，以确定每个值的来源，从而将修复落到正确位置。
  3. 需要推理的关系：
    信号 A对比信号 B问题
    canonical 主机生产环境源站canonical 中出现 localhost/staging/preview 主机（HIGH）
    sitemap 
    <loc>
    主机    		canonical 主机搜索引擎收到两套相互矛盾的信息
    sitemap 列出路径页面为 
    noindex
    或私有路由    		信号相互矛盾；私有内容暴露（admin/dashboard/api 为 HIGH）
    sitemap 列出路径robots 阻止该路径条目无法抓取
    sitemap 列出路径canonical 指向其他位置列出的是非 canonical URL
    sitemap 列出路径返回 404/重定向/soft-404失效条目（
    check-links
    ）
    robots 
    Sitemap:
    		实际 sitemap URL/主机通告了错误或开发环境的 sitemap
    robots 
    Disallow: /
    		生产环境意图整个网站被屏蔽（HIGH）
    robots 
    Disallow: /_next/
    /assets/
    		渲染爬虫无法完成渲染
    robots 列出 
    /admin
    		路由是否可访问？robots 不是访问控制机制；会向攻击者提供路径地图
    首页或大多数页面存在 
    noindex
    		上线意图staging 遗留设置（HIGH）
    x-robots-tag
    header    		meta robotsheader 可能覆盖 meta robots；两者都要检查
    多个页面指向同一个 canonical URL（
    CANONICAL_DUPLICATE_TARGET
    ）    		模板从首页复制而来；这会告诉搜索引擎忽略其他页面
    robots 阻止已链接的公开页面（
    ROBOTS_BLOCKS_PAGE
    ）    		页面的标题、描述和 
    noindex
    从未被读取    		爬虫无法看到它
    尾部斜杠 / 
    www
    / 
    http
    变体    		canonical + 重定向重复 URL 形式
  4. **标题和描述：**每页唯一、与页面内容相关、将主题放在前面；大致 50-60 / 120-160 个字符仅作为指导（用于截断，不是硬性规则）；不使用框架默认值；不堆砌关键词。文案应源自页面可见内容和产品事实；不要编造声明。
  5. **多语言：**成对的 
    hreflang
    、各语言 canonical（配合 
    multilingual-readiness
    ）。
  6. **仅含 JS 的页面：**如果内容只在 JavaScript 执行后出现，且不存在服务端渲染或预渲染，则说明索引取决于爬虫是否执行 JS：这是 SOURCE-INDICATED 风险，而非已证实的问题。

Evidence that counts

有效证据

Label each claim OBSERVED, SOURCE-INDICATED, DECLARED, INFERRED, UNKNOWN, or REVIEW REQUIRED. UNKNOWN is never a pass and never a failure.
  • Rendered tag from the running site: OBSERVED. Tag in source only: SOURCE-INDICATED.
  • Local checks cannot show what the production CDN redirects, rewrites, or serves. Verify the live URL when one exists; otherwise mark host/redirect behavior UNKNOWN.
  • Never state a page "is indexed", "will rank", or "Google will show X"; that needs Search Console (
    search-console-readiness
    ).
将每项结论标记为 OBSERVED、SOURCE-INDICATED、DECLARED、INFERRED、UNKNOWN 或 REVIEW REQUIRED。UNKNOWN 既不算通过，也不算失败。
  • 运行中网站渲染出的标签：OBSERVED。仅存在于源码中的标签：SOURCE-INDICATED。
  • 本地检查无法显示生产 CDN 会如何重定向、重写或提供内容。存在线上 URL 时应进行验证；否则将主机/重定向行为标记为 UNKNOWN。
  • 绝不能声称页面“已被索引”“会获得排名”或“Google 会显示 X”；这些结论需要 Search Console（
    search-console-readiness
    ）。

May change

可以修改的内容

Set/repair titles, descriptions, canonicals (only once the production origin is known: DECLARED or asked), robots.txt, sitemap.xml or its generator (
next-sitemap
, 
app/sitemap.ts
, 
@astrojs/sitemap
), 
noindex
on genuinely private routes, removal of private routes from the sitemap, redirects for duplicate URL forms. Preserve intentional owner copy; improve only starter, missing, duplicate, or placeholder values.
设置或修复标题、描述、canonical（仅在生产环境源站已知后：DECLARED 或已询问）、robots.txt、sitemap.xml 或其生成器（
next-sitemap
、
app/sitemap.ts
、
@astrojs/sitemap
）、对确实私有的路由设置 
noindex
、从 sitemap 中移除私有路由，以及为重复 URL 形式添加重定向。保留所有者有意编写的文案；仅改进初始模板、缺失、重复或占位值。

Must not claim

不得声称的内容

"SEO-optimized", "will be indexed", "will rank", "Google-friendly" or a numeric score. Do not claim a sitemap is correct because it exists; correctness is the agreement list above.
“已完成 SEO 优化”“会被索引”“会获得排名”“对 Google 友好”，也不要给出数字评分。不要仅因为 sitemap 存在就声称其正确；正确性取决于上述各项是否一致。

Verify

验证

Re-run 
inspect-metadata
and 
check-links
. Pass = no HIGH; every sitemap URL is public, canonical, 200, not noindex, not blocked; every canonical target resolves and is on the production host; no localhost/staging strings remain; robots and sitemap agree. Then spot-check one page in the rendered DOM.
重新运行 
inspect-metadata
和 
check-links
。通过标准 = 无 HIGH；每个 sitemap URL 都是公开页面，采用 canonical，返回 200，未设置 
noindex
且未被阻止；每个 canonical 目标均可访问并位于生产主机上；不残留 localhost/staging 字符串；robots 与 sitemap 相互一致。然后在渲染后的 DOM 中抽查一个页面。

Escalate

升级处理

Unknown production origin (ask before writing absolute URLs); a private route that is reachable (hand to 
web-security
); site migrations with redirects; sites with several domains or heavy JS rendering where search visibility matters (deep technical SEO is beyond this pass).
生产环境源站未知（写入绝对 URL 前先询问）；私有路由可被访问（移交给 
web-security
）；带重定向的网站迁移；拥有多个域名或大量 JS 渲染且搜索可见性重要的网站（深度技术 SEO 超出本次检查范围）。

No change is valid when

以下情况不应进行修改

The site is intentionally private and consistently noindex; a one-page site has no sitemap by design; an apparent finding is explained by context (
noindex
on 
/dashboard
). Do not generate a sitemap or rewrite good copy to hit a length target.
网站有意设为私有且始终使用 
noindex
；单页网站按设计没有 sitemap；表面上的发现已由上下文解释（例如 
/dashboard
上的 
noindex
）。不要为了达到长度目标而生成 sitemap 或重写优质文案。