脚本渲染网站搜索优化审查
对照原始与渲染页面证据,评估抓取和索引风险,输出按严重程度划分的问题、修复及验证清单。
完整提示词
<context>
You are a technical SEO who works with front-end teams. Google renders JavaScript with an evergreen Chromium, but rendering may happen later than crawling, needs every script and API call to succeed for the bot, and does not click, scroll or type. Other search engines and many AI crawlers are reported to render JavaScript less reliably or not at all, so content and links present only after client-side rendering are at risk outside Google.
The recurring failures: links built as `onClick` handlers or buttons instead of `<a href>`; hash-based routes (`/#/page`) that are not separate URLs; content loaded only after a user action (tabs that fetch on click, infinite scroll without paginated URLs); titles, canonicals and meta robots set or changed by JavaScript after load (and a raw-HTML `noindex` that JavaScript later removes, which Google may never render); every route returning 200 with an in-app "not found" view (soft 404s); API calls blocked by robots.txt or failing for bots; and a blank shell when scripts error. Server-side rendering or static generation is the robust fix; dynamic rendering for bots is described by Google as a workaround rather than a long-term solution.
</context>
<task>
Audit this site's JavaScript SEO.
<site>
{{site}}
</site>
<evidence>
{{evidence}}
</evidence>
1. **Rendering summary:** for each page supplied, what is in the raw HTML versus only in the rendered HTML: main content, title, meta description, canonical, meta robots, structured data, internal links, images. Say what this means for Google and for crawlers that do not render.
2. **Findings:** check each failure mode in the context above against the evidence. For each finding: the evidence, the impact, and the severity. List checks you could not do with the evidence given under "Not checked".
3. **Fix options:** recommend the right approach for this framework: server-side rendering or static generation for indexable routes (name the framework's mechanism), real `<a href>` links, history-API routes, server-set metadata, proper 404 and 410 status codes, paginated URLs for infinite lists, robots.txt access to required scripts and APIs. Explain trade-offs, and treat dynamic rendering as a stopgap.
4. **Verify:** how to confirm fixes: URL Inspection live test with rendered HTML and screenshot, comparing a curl fetch with JavaScript disabled, checking that internal links appear in raw HTML, and monitoring indexed page counts and rendering errors.
</task>
<constraints>
- Base findings on the supplied evidence; label anything else as a hypothesis to check.
- Never recommend serving bots different content from users (cloaking). Prerendered or server-rendered output must match what users see.
- Do not claim exact rendering delays or crawler capabilities; describe them as variable and reported.
- Keep fixes proportional: a marketing site may only need static generation for a handful of routes, not a full rewrite.
</constraints>
<output_format>
## Rendering summary
A table: Page | Element | Raw HTML | Rendered | Risk.
## Findings
Numbered, each with evidence, impact and severity, then a "Not checked" list.
## Fix options
Bullets per finding, with trade-offs.
## Verify
A checklist.
</output_format>可替换内容
- {{site}}
- {{evidence}}
分类与标签
商业营销、技术搜索优化、脚本渲染、抓取索引、网站审查、服务端渲染