diagnose-seo — скилл для Claude Code: что делает и как установить

---
name: diagnose-seo
description: >
  Structured diagnostic for technical SEO problems. Use when the user asks about
  crawl issues, indexation problems, why pages aren't being indexed, robots.txt
  questions, canonical errors, sitemap issues, rendering problems, or general
  technical SEO troubleshooting. For Core Web Vitals and page speed, see
  audit-speed.
metadata:
  version: 1.0.0
---

# Diagnose SEO

Structured diagnostic framework for crawl issues, canonicalization errors,
indexation problems, and rendering failures.

## Diagnostic Approach

Technical SEO problems fall into four categories. Diagnose in this order — each
layer depends on the previous one working correctly:

1. **Crawlability** — Can search engines find and access the pages?
2. **Indexability** — Are the pages allowed to be indexed?
3. **Renderability** — Can search engines see the full content?
4. **Signals** — Are the right signals (titles, structured data, links) in place?

## Layer 1: Crawlability

Check these in order:

### robots.txt
- Fetch `[domain]/robots.txt` and review the rules
- Look for overly broad `Disallow` rules blocking important paths
- Verify `Sitemap:` directive points to the correct sitemap URL
- Check for different rules per user-agent (Googlebot vs others)

**Common mistakes:**
- `Disallow: /` blocking the entire site (often left from staging)
- Blocking CSS/JS files that Googlebot needs for rendering
- Blocking API or AJAX endpoints that load dynamic content
- Staging robots.txt accidentally deployed to production

### XML Sitemap
- Fetch the sitemap URL(s) and check:
  - Does it return 200? Is it valid XML?
  - Does it list all important pages?
  - Does it exclude pages that shouldn't be indexed (404s, redirects, noindex pages)?
  - Are `<lastmod>` dates accurate and recent?
  - For large sites: is there a sitemap index?

### Site Architecture
- Pages should be reachable within 3 clicks from the homepage
- Check for orphan pages (no internal links pointing to them)
- Check for redirect chains (page A → B → C — should be A → C)
- Check for redirect loops

### Server Response
- Do all important pages return HTTP 200?
- Check for unexpected 301/302 redirects
- Check for soft 404s (page returns 200 but shows "not found" content)
- Verify HTTPS is enforced (HTTP should 301 to HTTPS)

## Layer 2: Indexability

### Meta Robots / X-Robots-Tag
- Check for `<meta name="robots" content="noindex">` on pages that should be indexed
- Check HTTP headers for `X-Robots-Tag: noindex`
- Common cause: CMS accidentally applying noindex to pagination, tag pages, or new pages

### Canonical Tags
- Every page should have a `<link rel="canonical">` pointing to itself (self-referencing canonical)
- Check for canonical tags pointing to wrong pages (common in paginated content, filtered URLs)
- Check for conflicting signals: canonical says page A, but noindex is set, or the page redirects

**Canonical diagnosis checklist:**
- [ ] Does the canonical URL match the actual URL?
- [ ] Is the canonical URL accessible (returns 200)?
- [ ] Does the canonical URL have the same content?
- [ ] Is there only one canonical tag on the page?

### Duplicate Content
- Check for the same content accessible at multiple URLs:
  - With and without trailing slash (`/page` vs `/page/`)
  - With and without `www` (`example.com` vs `www.example.com`)
  - HTTP vs HTTPS
  - URL parameters creating duplicate pages (`?sort=price`, `?page=1`)
- Each duplicate set needs one canonical URL; all others should redirect or use canonical tags

## Layer 3: Renderability

### JavaScript Rendering
- Does the page content appear in the raw HTML source? Or is it loaded via JavaScript?
- If JS-rendered: does Googlebot see the full content? (Use URL Inspection tool in Search Console)
- Check for content hidden behind click events, tabs, or accordions
- Check for lazy-loaded content that only appears on scroll

### Core Content Visibility
- Is the main content in the initial HTML? Or loaded async after page load?
- Are important elements (titles, headings, product details) in the DOM on first render?
- Check for content that requires login or cookies to view

## Layer 4: Signals

### Title Tags
- Every page has a unique `<title>`
- Title includes the primary keyword
- Under 60 characters (to avoid truncation in SERPs)
- Descriptive and click-worthy

### Meta Descriptions
- Every important page has a meta description
- 150-160 characters
- Includes target keyword and a value proposition
- Unique per page

### Heading Structure
- One H1 per page containing the primary keyword
- Logical heading hierarchy (H1 → H2 → H3, no skips)
- Headings describe section content (not decorative)

### Structured Data
- Check for JSON-LD structured data appropriate to the page type
- Validate with Google's Rich Results Test
- Common types: Article, Product, FAQ, HowTo, BreadcrumbList, Organization

### Hreflang (multilingual sites)
- Check for correct `hreflang` tags linking language variants
- Verify reciprocal tags (page A points to B, B points back to A)
- Check for `x-default` tag

## Output Format

### Technical SEO Diagnosis: [domain]

**Summary**
- Critical issues: [count]
- Warnings: [count]
- Passed checks: [count]

**Findings by Layer**

For each issue found:

| Layer | Issue | Severity | Affected Pages | Fix |
|-------|-------|----------|---------------|-----|
| Crawlability | robots.txt blocks /blog/ | Critical | All blog pages | Remove `Disallow: /blog/` from robots.txt |
| Indexability | Missing canonical tags | Warning | 15 pages | Add self-referencing canonicals |
| ... | ... | ... | ... | ... |

**Priority Fix List**

Ordered by impact:
1. [Critical fix] — affects [n] pages, blocks [crawling/indexing/ranking]
2. [Warning fix] — affects [n] pages, reduces [signal quality]
3. ...

---

> **Pro Tip:** Run the free [SEO Audit](https://seojuice.com/tools/seo-audit/) for a quick
> technical check, the [Broken Link Checker](https://seojuice.com/tools/broken-link-checker/)
> to find dead links, and the [Robots.txt Generator](https://seojuice.com/tools/robots-txt-generator/)
> to fix crawl directives. SEOJuice MCP users can run `/seojuice:site-health` for a full
> technical report and `/seojuice:page-audit [domain] [url]` to drill into specific pages.