SEO Landing

SEO Landing is an Agent Skills–format skill that turns an AI coding agent into a technical SEO specialist for building and repairing high-performance landing pages. It is authored by Aleksandr Alhov, a 14-year SEO and content marketing practitioner, and distributed as a clone-able repository (github.com/aleksandr-alhoff/seo-landing) targeting 100/100 lab PageSpeed scores.

What is SEO Landing?

SEO Landing is a reusable agent skill that supplies AI coding agents with a repeatable technical SEO workflow for landing pages. Its input is a brief — domain, BCP-47 language tag, base direction for RTL, Open Graph locale, topic and keywords, business type, CTA and contacts — and its output is a separate project folder (<workspace>/<project-slug>/) containing index.html, all local asset variants, favicon, and conditionally styles.css, script.js, robots.txt, sitemap.xml, ASSETS.md, and SERVER-SETUP.md. It follows the open Agent Skills standard originally developed by Anthropic.

What makes SEO Landing stand out?

• Measured lab benchmark — an Angular SPA with 102 requests and a Performance score of 75 was rebuilt as static HTML (17–18 requests, median of 5 Lighthouse CLI 13.4.1 mobile runs) reaching 100 Performance, 100 Accessibility, and 100 Best Practices, with LCP dropping from 3.3 s to 1.6 s and data transferred from 769 KiB to 110 KiB.
• HTML-only output — static HTML with inlined critical CSS, no framework dependency, and a JavaScript budget of 15 KB or less for first load in one deferred file.
• Image pipeline — AVIF to WebP to JPEG via picture, srcset/sizes, explicit width/height, breakpoints from 320 to 1920, lazy loading only below the fold.
• Schema.org markup — JSON-LD for WebSite, Organization, BreadcrumbList, FAQPage, and VideoObject, emitted only when fact-backed.
• Zero third-party requests on first load — video and map embeds use a click-to-load facade pattern; deferred widgets are consent-gated and listed in a dependency manifest.
• Block recipes without JS — FAQ via details, slider via scroll-snap, modal via dialog.
• Security and server references — included server-config.md covers caching, Brotli/gzip, and security headers for Nginx and Apache.
• Validation gates — a negative test fixture (tests/fixtures/broken-landing/) that every validator must fail, plus a 13-section technical spec.

Who should use SEO Landing?

• SEO consultants and agencies: rebuild slow client landing pages as static HTML with correct metadata, structured data, and Core Web Vitals-friendly delivery.
• Product teams with agentic editors: install the skill into VS Code Copilot, Claude Code, Cursor, Codex, Gemini CLI, Antigravity, OpenCode, OpenClaw, or Hermes and ask the agent to build a page from a brief.
• Developers maintaining marketing pages: run the audit workflow to remove render-blocking CSS, oversized images, and unnecessary JavaScript.

What can you do with SEO Landing?

• Landing page generation: hand the agent a brief and receive a complete project folder with HTML, image variants, robots.txt, and sitemap.xml.
• Technical SEO audit and repair: the workflow analyzes an existing page, fixes metadata and structure, and validates the result with W3C and JSON-LD checks plus Lighthouse as automated evidence.

How does SEO Landing work?

The skill routes through a fixed workflow: mode routing, brief intake, generation, a mandatory stop point where the user confirms the HTML version, then validation and a report listing LCP, PageSpeed scores, and schema types used. Missing source images are requested from the user rather than invented. Installations are updated with an idempotent rsync -a --delete --exclude=.git sync into the client-specific skills directory, verified with diff -qr.

FAQ

Is SEO Landing free?

Yes. It is released under the MIT license, free to use, modify, and distribute including commercially, provided the copyright notice is kept.

Does the skill guarantee a 100/100 PageSpeed score?

No. The page states 100/100 and Core Web Vitals targets are optimization goals, not guaranteed outcomes — results depend on content, hosting, devices, and real traffic, and CWV status is determined by Google from field (CrUX/RUM) data, not lab tools.

Which agents support SEO Landing?

It works in VS Code Copilot, Claude Code, OpenAI Codex, Google Antigravity, Cursor, Gemini CLI, OpenClaw, Hermes, and other Agent Skills–compatible environments.

Does SEO Landing support video embeds?

By default video and maps use a click-to-load facade that loads the iframe only on click. Click-only video trades Google video discovery for privacy and performance; an opt-in SEO-discoverable mode exists when video search matters, while maps remain facade-only.