Claude SEO - SEO Audit Skill for Claude Code
SEO analysis skill for Claude Code. 25 sub-skills (21 core + 1 orchestrator + 1 framework integration + 2 extension mirrors) and 18 sub-agents covering technical SEO, on-page analysis, content quality (E-E-A-T), schema markup, image optimization, sitemap architecture, AI search optimization (GEO), local SEO, maps intelligence, semantic topic clustering, search experience optimization (SXO), SEO drift monitoring, e-commerce SEO, international SEO, FLOW framework integration, Google SEO APIs, PDF report generation, and strategic planning.
Fork of AgriciDaniel/claude-seo — this fork removes community branding from AI context files and integrates additional methodology depth from seo-for-ai and seo-optimization skill sets. See What's different in this fork below.
Table of Contents
What's different in this fork
Branding cleanup
All community/personal branding removed from SKILL.md files, agent context files, and scripts. Credits remain in CONTRIBUTORS.md and CHANGELOG.md where they belong.
GEO methodology depth (seo-geo)
- Weighted Citation Share (WCS) — north-star KPI formula with intent/engine/quality weights
- Citation Survival Rate (CSR) — 4-stage pipeline (Track → Measure → Analyze → Optimize)
- Baseline guard — hard rule: no content changes without a recorded WCS baseline
- Platform rubrics — 100-point scoring for Google AIO, ChatGPT, Perplexity, Gemini, Bing Copilot
- Entity modeling — Entity Authority Score 0-100,
sameAs priority list, schema patterns
- E-E-A-T scoring rubric — 100-point rubric with YMYL × schema interaction
- Brand authority signals — AI crawler UA list (14 bots), correlation data, sentiment audit
- 3 helper scripts:
geo_checker.py, llms_txt_generator.py, prompt_cluster_generator.py
Active link building (seo-backlinks)
- 6 campaign types — guest post, resource page, broken link, HARO, digital PR, directories
- Outreach templates — 5 email templates with personalization, A/B testing, response handling
- Directory submissions — Tier 1 (DA 80+), Tier 2 (DA 50-79), 8 niche verticals
Programmatic SEO depth (seo-programmatic)
- Best practices — content uniqueness strategy (>80% target, HARD STOP <30%), pre-launch checklist
- Templates & URLs — full Nunjucks syntax,
PSEOURLGenerator class with dedup and length cap
- Scale architecture —
BatchProcessor (1000-row chunks), WorkerPool (4-8 threads), CheckpointManager, per-batch quality validation with 10%-failure abort
Semantic SEO (seo-content)
- Entity optimization — types, salience, relationships,
about/mentions schema markup
- Topical authority hub-spoke — coverage checklist, authority signals, SERP-based clustering
- Semantic/LSI variations — integration strategy with density guidelines
- NLP-friendly structure — Q&A format, semantic HTML, logical content flow
- 5-step audit workflow with before/after example
Installation
Manual Install (Unix/macOS/Linux)
git clone --depth 1 https://github.com/lugondev/claude-seo.git
bash claude-seo/install.sh
One-liner (curl, review then run)
curl -fsSL https://raw.githubusercontent.com/lugondev/claude-seo/main/install.sh > install.sh
cat install.sh # review
bash install.sh # run when satisfied
rm install.sh
Windows (PowerShell)
git clone --depth 1 https://github.com/lugondev/claude-seo.git
powershell -ExecutionPolicy Bypass -File claude-seo\install.ps1
Why git clone instead of irm | iex? Claude Code's own security guardrails flag irm ... | iex as a supply chain risk. The git clone approach lets you inspect claude-seo\install.ps1 before running it.
Quick Start
# Start Claude Code
claude
# Run a full site audit
/seo audit https://example.com
# Analyze a single page
/seo page https://example.com/about
# Check schema markup
/seo schema https://example.com
# Generate a sitemap
/seo sitemap generate
# Optimize for AI search
/seo geo https://example.com
Commands
