Technical SEO Audit

Categories

1. Crawlability

• robots.txt: exists, valid, not blocking important resources
• XML sitemap: run claude-seo run sitemap_discovery.py <url> --json; require a valid entry in found, and report stale or unsafe robots.txt declarations separately from working fallback locations
• Noindex tags: intentional vs accidental
• Crawl depth: important pages within 3 clicks of homepage
• JavaScript rendering: check if critical content requires JS execution
• Crawl budget: for large sites (>10k pages), efficiency matters
• Googlebot fetch limits: Googlebot fetches the first 2MB of HTML and first 64MB of a PDF (uncompressed; 15MB is the broader crawler-infra default). Long-standing, not a 2026 change, but inline base64 images, oversized inline CSS/JS, or bloated nav can push critical content/JSON-LD past the cap and out of the index. Keep key content + structured data within the first 2MB.
• Crawl rate auto-adjusts (backs off on 5xx/slow responses); there is no manual crawl-rate control (the legacy Search Console setting was removed Jan 2024). Influence crawling via sitemaps, server responsiveness, and robots controls.
• Google's canonical crawling/robots reference moved to developers.google.com/crawling (migrated 2025-11-20); IP-range files relocated to /crawling/ipranges/ and googlebot.json was renamed common-crawlers.json.

AI Crawler Management

As of 2025-2026, AI companies actively crawl the web to train models and power AI search. Managing these crawlers via robots.txt is a critical technical SEO consideration.

Known AI crawlers:

Key distinctions:

• Blocking Google-Extended prevents Gemini training use but does NOT affect Google Search indexing or AI Overviews (those use Googlebot)
• Blocking GPTBot prevents OpenAI training but does NOT prevent ChatGPT from citing your content via browsing (ChatGPT-User)
• ~3-5% of websites now use AI-specific robots.txt rules

Example, selective AI crawler blocking:

# Allow search indexing, block AI training crawlers
User-agent: GPTBot
Disallow: /

User-agent: Google-Extended
Disallow: /

User-agent: Bytespider
Disallow: /

# Allow all other crawlers (including Googlebot for search)
User-agent: *
Allow: /

Recommendation: Consider your AI visibility strategy before blocking. Being cited by AI systems drives brand awareness and referral traffic. Cross-reference the seo-geo skill for the full AI crawler/fetcher taxonomy.

User-triggered fetchers ignore robots.txt by design. Google now documents Google-Agent (Project Mariner, agentic browsing) plus Google-NotebookLM and Google Messages as user-triggered fetchers that cannot be blocked via robots.txt. Use server-side access controls instead. By contrast, Google-Extended and Google-CloudVertexBot obey robots.txt. Emerging: Web Bot Auth (RFC 9421) lets bots authenticate cryptographically via a Signature-Agent header + key directory at agent.bot.goog (used by Google-Agent); reverse-DNS verification remains the fallback.

2. Indexability

• Canonical tags: self-referencing, no conflicts with noindex
• Duplicate content: near-duplicates, parameter URLs, www vs non-www
• Thin content: pages below minimum word counts per type
• Pagination: rel=next/prev or load-more pattern
• Hreflang: correct for multi-language/multi-region sites
• Index bloat: unnecessary pages consuming crawl budget

3. Security

• HTTPS: enforced, valid SSL certificate, no mixed content
• Security headers:
  • Content-Security-Policy (CSP)
  • Strict-Transport-Security (HSTS)
  • X-Frame-Options
  • X-Content-Type-Options
  • Referrer-Policy
• HSTS preload: check preload list inclusion for high-security sites
• Back-button hijacking (spam-policy violation, malicious practices): flag pages that defeat the Back button via history.pushState/replaceState (including scripts injected by third-party ad/library platforms). Added to Google's spam policies 2026-04-13; enforcement live since 2026-06-15 (manual actions + automated demotions): treat as Critical.

4. URL Structure

• Clean URLs: descriptive, hyphenated, no query parameters for content
• Hierarchy: logical folder structure reflecting site architecture
• Redirects: no chains (max 1 hop), 301 for permanent moves
• URL length: flag >100 characters
• Trailing slashes: consistent usage

5. Mobile Optimization & Page Experience

• Responsive design: viewport meta tag, responsive CSS
• Touch targets: minimum 48x48px with 8px spacing
• Font size: minimum 16px base
• No horizontal scroll
• Mobile-first indexing: Googlebot Smartphone is the primary crawler (rollout completed 2024). A mobile version is not strictly required (Google says "very strongly recommended"), sites that don't work on mobile can still be indexed, but the real risk is content/parity loss, not hard exclusion.
• Mobile/desktop content parity (highest-value mobile check): equivalent primary content, matching robots meta tags, matching titles/descriptions, equivalent structured data, crawlable resources; avoid lazy-loading primary content that requires user interaction.
• Intrusive interstitials / ad density: flag full-page interstitials, standalone consent-redirect pages, persistent blocking dialogs, and excessive/distracting ad density (a named page-experience aspect). Acceptable: small banners, standard CMS/legal dialogs.
• "Read more" deep links: keep key content immediately visible on load (not behind tabs/accordions), don't hijack scroll on load, and preserve URL hash fragments, content hidden behind expandable sections is less likely to qualify.

Page experience is guidance, not a single ranking system. Only Core Web Vitals feeds ranking directly; HTTPS is a confirmed but lightweight signal (affects <~1% of queries). Relevance can still win even when page experience is sub-par, so don't over-weight security headers. Note: the standalone Page Experience report was removed from Search Console (monitor via the Core Web Vitals + HTTPS reports).

6. Core Web Vitals

• LCP (Largest Contentful Paint): target <=2.5s
• INP (Interaction to Next Paint): target <=200ms
  • INP replaced FID on March 12, 2024. FID was removed from Chrome's field-data tools (CrUX API, PageSpeed Insights) on September 9, 2024 (Lighthouse is a lab tool that never reported FID). Do NOT reference FID anywhere.
• CLS (Cumulative Layout Shift): target <=0.1
• Evaluation uses 75th percentile of real user data
• Use PageSpeed Insights API or CrUX data if MCP available

7. Structured Data

• Detection: JSON-LD (preferred), Microdata, RDFa
• Validation against Google's supported types
• See seo-schema skill for full analysis

8. JavaScript Rendering

• Check if content visible in initial HTML vs requires JS
• Identify client-side rendered (CSR) vs server-side rendered (SSR)
• Flag SPA frameworks (React, Vue, Angular) that may cause indexing issues
• Verify dynamic rendering setup if applicable

JavaScript SEO: Canonical & Indexing Guidance (December 2025)

Google updated its JavaScript SEO documentation in December 2025 with critical clarifications:

1. Canonical conflicts: If a canonical tag in raw HTML differs from one injected by JavaScript, Google may use EITHER one. Ensure canonical tags are identical between server-rendered HTML and JS-rendered output.
2. noindex with JavaScript: If raw HTML contains <meta name="robots" content="noindex"> but JavaScript removes it, Google MAY still honor the noindex from raw HTML. Serve correct robots directives in the initial HTML response.
3. Non-200 status codes: Google does NOT render JavaScript on pages returning non-200 HTTP status codes. Any content or meta tags injected via JS on error pages will be invisible to Googlebot.
4. Structured data in JavaScript: Product, Article, and other structured data injected via JS may face delayed processing. For time-sensitive structured data (especially e-commerce Product markup), include it in the initial server-rendered HTML.

Best practice: Serve critical SEO elements (canonical, meta robots, structured data, title, meta description) in the initial server-rendered HTML rather than relying on JavaScript injection.

9. IndexNow Protocol

• Check if site supports IndexNow for Bing, Yandex, Naver
• Supported by search engines other than Google
• Recommend implementation for faster indexing on non-Google engines

Agent-Friendly Pages & Agentic Browsing

AI agents (not just AI summarizers) increasingly read sites through three channels: vision models on screenshots, raw HTML/DOM, and the accessibility tree (the cleanest signal). Audit criteria: semantic HTML (real <button> and <a>, not <div onclick>), label associations, interactive target sizing, layout stability across templates, cursor: pointer correctness, live in references/agent-friendly-pages.md.

Google now ships a Lighthouse Agentic Browsing category (default-on since Lighthouse 13.3.0, Chrome 150+; buckets: agent-centric accessibility, CLS + llms.txt, three WebMCP audits). It reports a fractional pass-ratio (X of N), not a 0-100 score, keep that distinct from this skill's own Agent-UX 0-100 heuristic below. The PSI REST API does not expose it; run via Lighthouse CLI --only-categories=agentic-browsing, DevTools, or the PSI web UI. See references/agent-friendly-pages.md.

Audit command

# Render with Playwright + capture accessibility tree, then score
claude-seo run agent_ux_check.py https://example.com --json

The scanner outputs an Agent-UX score (0-100) plus itemized issues:

• HTML findings: real buttons / anchors, <div onclick> widgets, semantic landmarks, inputs without <label for>, inputs without ARIA labels
• Accessibility tree findings: total nodes, interactive nodes, unnamed interactive elements, role="generic" ratio

The accessibility-tree snapshot uses Playwright's page.accessibility.snapshot(interesting_only=False). To capture the tree without scoring, use claude-seo run render_page.py <url> --a11y-tree --json.

Surface findings as opportunities, not failures; don't gate audits on a sub-100 Agent-UX score. WebMCP origin-trial/sign-up status needs verification, and absence of WebMCP support is still an opportunity, not a defect.

Output

Technical Score: XX/100

Category Breakdown

Critical Issues (fix immediately)

High Priority (fix within 1 week)

Medium Priority (fix within 1 month)

Low Priority (backlog)

DataForSEO Integration (Optional)

If DataForSEO MCP tools are available, use on_page_instant_pages for real page analysis (status codes, page timing, broken links, on-page checks), on_page_lighthouse for Lighthouse audits (performance, accessibility, SEO scores), and domain_analytics_technologies_domain_technologies for technology stack detection.

Google API Integration (Optional)

If Google API credentials are configured, use claude-seo run pagespeed_check.py <url> --json for real PSI + CrUX field data (replaces lab-only CWV estimates), claude-seo run crux_history.py <url> --json for 25-week CWV trends, and claude-seo run gsc_inspect.py <url> --json for real indexation status per URL.

Error Handling

---

Supporting file: LICENSE.txt

MIT License - see repository root LICENSE file for complete terms.

Copyright (c) 2026 AgriciDaniel
https://github.com/AgriciDaniel/claude-seo

---

Supporting file: references/agent-friendly-pages.md

Agent-friendly pages — audit reference (June 2026)

The next wave of AI search is not summarization — it's agents acting on the user's behalf (search, compare, buy, book). Google's AI optimization guide and the linked web.dev article describe three channels through which agents interpret your site:

1. Screenshots + vision model — interprets visual hierarchy, button prominence, layout. Slow and token-expensive.
2. Raw HTML / DOM — nesting, IDs, classes, data attributes.
3. The accessibility tree — the browser-native semantic distillation (roles, names, states). The cleanest signal of the three.

Modern agents combine all three. Optimizing for the accessibility tree is the single highest-leverage move; if your accessibility tree is broken, no amount of visual polish saves you.

Primary sources:

• Google AI optimization guide: https://developers.google.com/search/docs/fundamentals/ai-optimization-guide
• web.dev (referenced from the guide above): article on building agent-friendly websites

Audit checklist

1. Use real interactive elements

If you cannot use a real interactive tag, supply ARIA: role="button", role="link", tabindex="0", plus key handlers for Enter and Space.

Why it matters: the accessibility tree exposes real interactive elements with their roles. Custom div widgets often appear in the tree with no role at all — agents skip them.

2. Label associations

Every form input must have an associated label:

<label for="email">Email</label>
<input id="email" type="email" name="email">

Or use aria-label / aria-labelledby where a visible label isn't possible. Agents that read the accessibility tree get the field purpose directly from the associated label — without it, the input is a void.

3. Interactive target size

Visual-analysis pipelines filter out interactive elements smaller than ~8 square pixels of unobscured area. Tap-target accessibility minimums (24×24 WCAG AA, 44×44 Apple HIG) are stricter and pass the agent gate by default.

Audit: any clickable element below 24×24px is a candidate for agent invisibility, in addition to the WCAG failure.

4. Don't cover interactive nodes with transparent overlays

Vision models discard covered nodes when computing what's "interactive at this position". Common offenders:

• Full-card click handlers that overlay every child link.
• Transparent cookie-consent layers persisting beyond consent.
• Modal portals with pointer-events: auto left on after dismiss.
• "Ghost" tracking pixels with position: absolute; inset: 0.

5. Layout stability

If "Add to cart" lives in different positions on /category/shoes vs /category/bags, screenshot-based agents have to relearn the page each visit. Keep functionally identical actions in the same screen quadrant across templates.

Cross-reference: this overlaps with CLS (Cumulative Layout Shift) in Core Web Vitals, but the agent-UX concern is broader — it covers page-to-page stability, not just within-page shift.

6. cursor: pointer as a legitimate signal

Vision models read cursor: pointer (set by default on <a> / <button>) as a hint that an element is actionable. Do not override it to cursor: default on truly interactive elements just for visual minimalism.

Inverse: do not apply cursor: pointer to non-interactive elements — that makes agents click things that do nothing.

7. Stable, meaningful selectors

Agents that fall back to DOM parsing rely on:

• Real semantic tags (<nav>, <main>, <article>, <section>, <aside>)
• Stable id attributes on top-level layout containers
• data-* attributes that describe purpose, not implementation

Avoid auto-generated class names like __sc_a4b7d9e2 as the only handle on a critical interactive element — agents can target them but cannot tell what they mean.

Lighthouse Agentic Browsing category (shipped)

Google now ships a dedicated Agentic Browsing Lighthouse category — created in Lighthouse 13.2.0 (2026-05-01) and on by default since 13.3.0 (2026-05-07) in DevTools and the CLI (Chrome 150+). Category id: agentic-browsing. Unlike other categories it reports a fractional pass-ratio (X of N checks passed), not a 0–100 weighted score — so do not compute or report a weighted agentic "score". It groups deterministic audits into three buckets:

• Agent-centric accessibility — reuses three a11y audits: Names and labels (every interactive element has a programmatic name), Tree integrity (valid roles + parent-child relationships), Visibility (interactive content not hidden from the accessibility tree). These map directly onto checklist items 1–2 above and the agent_ux_check.py heuristics.
• Stability & discoverability — CLS (visual stability for element positioning; see checklist item 5) + an llms.txt presence-at-domain-root check (13.4.0 relaxed it to allow leading whitespace).
• WebMCP integration — three audits, see below.

Run paths: DevTools Lighthouse panel (Chrome 150+, default-on, no toggle); CLI npx lighthouse@latest <url> --only-categories=agentic-browsing (Node.js 22.19+, per-audit JSON for CI); PageSpeed Insights web UI and API. Lighthouse 13.4.1 enabled the category through the PSI API after 13.4.0 had disabled it. CrUX never provides these lab audits.

WebMCP (proposed standard; status needs recheck)

WebMCP lets a site declare structured tools/actions for agents rather than agents inferring intent from the DOM. It is no longer "years away": a flag-gated Early Preview (Chrome 146 Canary) opened 2026-02-10, and Chrome 149 origin-trial/sign-up status needs verification from a Google-owned source (local dev flag: chrome://flags/#enable-webmcp-testing). It remains a proposed standard from the Web Machine Learning community group (github.com/webmachinelearning/webmcp), not W3C-finalized.

Lighthouse 13.2+ ships three WebMCP audits under the Agentic Browsing category (require Chrome 150+ and origin-trial registration). Shipped audit ids in the 13.2.0 release notes:

• webmcp-form-coverage: informational; flags <form> elements lacking toolname / tooldescription.
• webmcp-registered-tools: informational; lists tools registered via the declarative (HTML) and imperative (JS) APIs.
• webmcp-schema-validity: validates WebMCP form schemas.

(Code parsing Lighthouse JSON should expect the release-note ids above.)

Audit posture: WebMCP is an active opportunity, not a finding to hard-fail — absence is not a defect. Surface it as an opportunity for sites that want first-class agent actions, and note origin-trial enrollment is required for the audits to fire.

Quick-audit one-liner

For a fast smoke check, capture the accessibility tree via Lighthouse or Chrome DevTools and look for:

• Any interactive element with role="generic" → broken semantics.
• Any input without an accessible name → missing label.
• Any <div> with onclick and no role / tabindex → custom widget that agents won't see.

claude-seo run render_page.py <URL> --mode auto --a11y-tree --json loads the page headlessly and captures Chromium's full accessibility tree through CDP. Use claude-seo run agent_ux_check.py <URL> --json for the bounded Agent-UX heuristic and its explicit complete, partial, or unavailable score status.

Last verified

2026-08-25. Google has published the Lighthouse Agentic Browsing framework, Lighthouse 13.4.1 exposes it through the PSI API, and the local renderer now captures the Chromium accessibility tree through CDP. Update when:

• WebMCP graduates from origin trial to a stable/shipped API (or W3C status changes).
• Google adds or renames audits in the agentic-browsing category.
• web.dev publishes a follow-up article with revised criteria.