A Claude Code skill that generates SEO and GEO (Generative Engine Optimization) optimized content — blogs, FAQs, glossary terms, comparison pages, and pillar guides.
Built from real conversations with SEO/GEO practitioners and Google's official guide to optimizing for AI search.
| File | What it does |
|---|---|
SKILL.md |
The Claude Code skill — install it and use /seo-blog |
seo_geo_playbook.md |
Complete research, structure, writing, and post-publish rules |
content_architecture.md |
Website content map — what pages to build, where they go, how they connect |
humanization_prompts.md |
Copy-paste prompts for making AI content sound human + image generation templates |
learnings.md |
Pre-seeded insights that improve over time with --improve |
Copy the skill to your Claude Code skills directory:
mkdir -p ~/.claude/skills/seo-blog
cp SKILL.md seo_geo_playbook.md content_architecture.md humanization_prompts.md learnings.md ~/.claude/skills/seo-blog/The skill also reads a shared writing style guide. If you don't have one, it works without it — but creating one at ~/.claude/skills/content-shared/writing_style_guide.md will improve voice consistency.
/seo-blog [topic] # Full pipeline: research → outline → write → output
/seo-blog --outline [topic] # Research + outline only (review before writing)
/seo-blog --research-only [topic] # Keyword research + forum mining + competitor audit
/seo-blog --type comparison [topic] # Generate a specific content type
/seo-blog --type glossary [term] # Generate a glossary entry
/seo-blog --type pillar [topic] # Generate a pillar page
/seo-blog --improve "your feedback" # Save a learning for future posts
Works with any project. Pass --project [name] to inject project-specific context, or the skill will ask.
- Keyword research via Google Trends + Keyword Planner MCPs (if available)
- Forum mining — finds actual questions people ask on Reddit, HN, StackOverflow, Quora
- Competitor content audit — what top-ranking articles cover and what they miss
- Data sourcing — finds citable stats with specific numbers
- Classifies funnel position and recommends the right content type
- Question-based H2 headings from real search queries
- Self-contained sections (each answers its heading question — critical for GEO)
- TL;DR on top, tables, FAQs, interlinks, outlinks, data citations
- Two-pass technique: completeness first, then humanization (H.U.M.A.N. framework)
- Banned word list to avoid AI-detection signals
- MDX format (for Next.js blogs) or generic Markdown
- Full SEO meta tags, structured data (BlogPosting, FAQ, Breadcrumb)
- Verification checklist (17 checks across structure, SEO, humanization)
- Image generation prompts for Gemini/ChatGPT (copy-paste ready)
The skill is aligned with Google's AI optimization guide, which debunks several popular "GEO hacks":
| Popular advice | Google says | This skill's approach |
|---|---|---|
| FAQ schema = GEO must-have | "Structured data isn't required for AI search" | Adds FAQ schema (helps rich snippets), doesn't treat it as a GEO lever |
| Chunk content for AI | "No requirement to break into tiny pieces" | Self-contained sections for readability, not AI parsing |
| Write specifically for AI | "AI understands synonyms" | Writes for humans, structures for scannability |
| llms.txt files | "Don't need machine-readable files" | Doesn't prioritize |
What Google does emphasize:
- Unique point of view with first-hand experience
- Non-commodity content (case studies, proprietary data, specific numbers)
- Quality organization (clear sections, headings)
- Multimedia (images/video)
- Technical fundamentals (crawlability, speed)
The content_architecture.md file includes a complete website content map:
- Blog posts, pillar pages, comparison pages, glossary terms
- FAQ strategy (embedded sections, not standalone pages)
- Linking rules between content types
- Priority build order for any new project
- URL patterns and funnel mapping
Inspired by how reo.dev and firecrawl.dev structure their content.
This skill was built from:
- Google's guide to optimizing for generative AI in search — the official source on what actually matters for GEO
- AEO/GEO panel discussion featuring SEO/GEO practitioners — practical advice on content structure, research workflows, and what's working in 2026
- Content humanization research — techniques for burstiness, perplexity, voice consistency, and avoiding AI-detection signals
- Reference blogs — reo.dev (content ecosystem approach) and firecrawl.dev (minimal structure + programmatic SEO)
MIT