| name | seo-technical-audit |
| description | Technical SEO sweep — crawlability, indexability, schema markup, sitemaps, OG / Twitter Card metadata, hreflang, structured data. Used by `web-architect` (primary, technical) + `content-strategist` (content-SEO). |
Skill: seo-technical-audit
Purpose: Technical SEO sweep — crawlability, indexability, structured data, social-share metadata, internationalization. Used by web-architect (technical) + content-strategist (content-side).
When to use
- Pre-launch verification
- Search ranking regression
- Migration / re-platform readiness
- Annual / quarterly review cycle
- New section / language launch
The 7 areas
1. Crawlability
robots.txt present at root, valid syntaxrobots.txt doesn't accidentally block / or critical pathssitemap.xml referenced from robots.txt- Sitemap accurate — every URL returns 200, no 404s in sitemap
sitemap.xml includes <lastmod> for cache-aware crawlers- No infinite-redirect loops; no 30x chains > 1 hop
- No
<meta name="robots" content="noindex"> on key pages (the hook catches this for HTML files)
2. Indexability
- Canonical URLs declared on every page (or canonicalized at the server / CDN level via 301)
- Trailing-slash decision made + enforced (one or the other, consistently)
- Case-sensitivity: lowercase URLs enforced; mixed-case redirect-or-404
- No duplicate content across URLs (UTM-only variants, faceted-nav variants)
- Pagination uses
rel="next" / rel="prev" (deprecated by Google but still respected by some crawlers; consistent UX matters)
- Hash-based URLs only for in-page anchors; never for primary navigation
3. URL structure + IA
- Short, readable URLs (
/blog/topic > /post.php?id=42)
- URLs reflect IA (a page at
/blog/foo is a child of /blog)
- Avoid stop-words / dates in primary URLs unless intentional
- Locale prefix consistent (
/en/, /de/, etc.) if multi-language
4. Structured data (Schema.org)
- Schema.org JSON-LD in
<head> (not Microdata or RDFa — JSON-LD is the preferred format)
- Validates against schema.org and Google Rich Results Test
- Match the page's actual content (no spam-typing)
- Common types used appropriately:
Organization (homepage, about)WebSite (homepage with sitelinks search box)BreadcrumbList (every page below root)Article / BlogPosting (blog content)Product (e-commerce)FAQPage (FAQ content; sparingly — Google has narrowed eligibility)LocalBusiness (locations)Person (author bios)
5. Social share (OG + Twitter Card)
<meta property="og:title">, og:description, og:image, og:url, og:type<meta name="twitter:card" content="summary_large_image">, twitter:title, twitter:description, twitter:image- OG image: 1200×630 minimum, < 5 MB, branded consistently
- Twitter image: same as OG image is fine if it fits
- Validate via Twitter Card Validator + Facebook Sharing Debugger
6. Internationalization
<html lang="..."> on every page (locale code, e.g., en-US)hreflang tags on multilingual pages, including self-referencehreflang is bidirectional (page A → page B; page B → page A)x-default declared- Language switcher in nav (not just IP-based redirect)
7. Performance × SEO
- Core Web Vitals pass (LCP / CLS / INP) — see
../core-web-vitals-tuning/SKILL.md
- HTTPS everywhere
- Mobile-friendly (Google's mobile-friendliness test passes)
- Page-experience signals on (CWV is the load-bearing one)
Output
| Area | Score (✅ / ⚠️ / 🔴) | Findings | Severity (P0 / P1 / P2) |
|---|
Plus:
- Top 5 fixes ranked by leverage
- Index-coverage report (Google Search Console pull: indexed / excluded / errors)
- Recommendation: clean / targeted-fixes / comprehensive-remediation
Severity guide
- P0 — issue blocks indexability of critical pages (robots disallow, noindex, broken canonical)
- P1 — issue degrades ranking signal (missing structured data on monetizable pages, broken hreflang)
- P2 — improvable; not currently hurting ranking
Tools
- Google Search Console — index coverage, performance reports, sitemaps, manual actions
- Google Rich Results Test — structured-data validation
- Schema.org validator — schema.org-side validation
- Twitter Card Validator + Facebook Sharing Debugger — social share
- Screaming Frog / Sitebulb — crawler simulation
- Ahrefs / Semrush / Moz — for backlink + keyword data (out of scope for technical audit; content-side)
Anti-patterns the audit catches
robots.txt accidentally Disallow: / (kills indexing)- Canonical URL pointing to itself + a different page (loop)
- Trailing-slash inconsistency (Google sees
/about and /about/ as separate pages)
- OG image missing → bare URL on social share
lang attribute missing or wrong on <html>hreflang declared one-way (A → B without B → A)- FAQ schema used on pages without actual FAQ structure
- Structured data that doesn't match visible content (spam pattern; risks manual action)
- Multiple H1s per page (technically allowed in HTML5; still confusing for some crawlers)
- Duplicate title tags across pages
See also
