Claude SEO

Open-source SEO analysis plugin for Claude Code by AgriciDaniel. Runs 25 sub-skills and up to 15 specialist agents in parallel across technical SEO, content quality (E-E-A-T), Schema.org markup, AI-search optimization (GEO/AEO), local SEO, e-commerce, and international SEO. Every audit produces a prioritized action plan where each recommendation carries a first-principle observation, dependency relationships, an explicit falsifiability check, and a leading indicator.

🧩 What it does

• Full site audit (/seo audit <url>) — parallel sub-agent fan-out, prioritized action plan in 10-15 minutes
• 27 commands across orchestrator + sub-skills: page, technical, content, schema, geo, sitemap, images, plan, programmatic, competitor-pages, local, maps, hreflang, google, backlinks, cluster, sxo, drift, ecommerce, flow, plus optional firecrawl, dataforseo, image-gen extensions
• AI-search first — aligned with Google’s AI Optimization Guide (May 2026); scores passage citability (134-167 word self-contained answer blocks), question-based heading hierarchy, attribution density, entity presence across Wikipedia/Reddit/YouTube/LinkedIn
• Headless rendering everywhere (v2 Phase A) — shared Playwright Chromium with trafilatura + htmldate; SPA-aware fetching auto-detected on Next.js/React/Vue/Nuxt/Astro
• 4-tier Google API integration — PageSpeed/CrUX (key only) → GSC/Indexing (OAuth) → GA4 → Keyword Planner
• SQLite drift monitoring — baseline/compare/history per URL
• PDF reports via WeasyPrint + matplotlib (A4, ~32 pages)

🔗 Links

• Public OSS: https://github.com/AgriciDaniel/claude-seo (MIT)
• Community private mirror (early access): AI-Marketing-Hub/claude-seo — requires AI Marketing Hub Pro membership
• Codex port: https://github.com/AgriciDaniel/codex-seo
• Demo: https://www.youtube.com/watch?v=COMnNlUakQk

🚀 Install

Plugin install (Claude Code 1.0.33+):

/plugin marketplace add AgriciDaniel/claude-seo
/plugin install claude-seo@agricidaniel-claude-seo

Or manual clone + install.sh / install.ps1. Avoids irm | iex deliberately to keep the install script reviewable.

🎨 Why it matters for me

• Direct fit for Spec-driven SEO and GEO — the “code-stack only top 20% controls” thesis becomes concretely actionable when 18 agents enforce it
• E-commerce SEO module is relevant for Qamera AI product-page audits
• Local SEO + maps intelligence layer would slot into PLSoft client work
• Parallel sub-agent pattern is the same architecture used in Awesome Claude Code / Karpathy Skills — concrete reference implementation of the Agent Skills standard

☘️ Methodology

10 principles across PERCEIVE (observe internal/external + listen) → ANALYZE (think + connect lateral/system) → VALIDATE (feel + accept) → ACT (create + grow). Every recommendation carries four fields: first-principle observation, dependency, falsifiability check, leading indicator. Anti-pattern catalog: llms.txt is not a citation lever (primary-source evidence in repo), content chunking not required, AI-specific keyword rewriting unnecessary (synonym understanding is sufficient).

📒 Notable design choices

• MIT, your files, no data leaves machine — vs commercial SEO tools that upload to vendor
• Falsifiability per finding — Google’s own QRG (Sept 2025) heuristics encoded as gates
• Active schema coverage — Organization, LocalBusiness, Article, Product, Review, BreadcrumbList, etc.; explicitly excludes deprecated types (HowTo retired Sept 2023, FAQ restricted to gov/health, SpecialAnnouncement retired July 2025)
• INP-not-FID — FID removed from Chrome tools Sept 2024; only LCP/INP/CLS measured

📖 Further reading

• Spec-driven SEO and GEO
• Agent Skills
• Awesome Claude Code
• Karpathy Skills
• Claude Code
• Claude Code Best Practice