SEO Aeo

SEO + AEO Agent Skill — SERAPH

Philosophy: Research first. Deploy second. Validate always. Core Principle: Human-first content, machine-optimized structure.

🎯 Selective Reading Rule (MANDATORY)

Read REQUIRED files always, OPTIONAL only when needed:

🔴 aeo-strategies.md = ALWAYS READ. Others = only if relevant to the task.

Memory System

Load Memory (Start of Every Task)

cat memory/project-memory.json

If file doesn't exist, run the intake flow and create it:

cp .agent-os/skills/seo-aeo/memory/project-memory.template.json memory/project-memory.json

Save Memory (End of Every Task)

Update memory/project-memory.json with:

• Keywords researched and their intent classification
• Pages optimized and what schema was applied
• Submission status (GSC, Bing, GBP)
• Pending tasks
• Competitor insights discovered
• Core Web Vitals scores if measured

Phase Flow — Follow This Order Every Time

PHASE 0: Intake (if project-memory.json is empty or new)

Ask for or extract from existing files:

• Business name, URL, industry, geo, language, primary goal
• Top 3 competitor URLs
• Target audience (ICP)
• Any existing GSC/Analytics data

Then research:

1. Keyword universe (seed → LSI → long-tail → question-based → entity)
2. Search intent per keyword (Info / Nav / Commercial / Transactional)
3. Competitor schema audit — what types are they using?
4. AEO question mapping — what AI engines should cite this site for?

PHASE 1: Technical Files

Generate in this order:

1. robots.txt → see references/technical-seo.md §robots
2. sitemap_index.xml → see references/technical-seo.md §sitemap
3. Full <head> meta block → see references/technical-seo.md §meta
4. Schema markup (JSON-LD) → see references/schema-library.md

PHASE 2: AEO Content

Apply AEO rules from references/aeo-strategies.md to every page:

• Inverted pyramid structure
• Question-based H2/H3 headers
• FAQ schema on every page
• Entity coverage map
• Direct answer blocks

PHASE 3: Submission & Verification

Follow checklist in references/submission-checklist.md:

• GSC property setup + sitemap submission
• Bing Webmaster + IndexNow API
• Google Business Profile
• Structured data validation commands

PHASE 4: Monitoring Plan

Output weekly + monthly automation tasks. For n8n workflows, generate the JSON trigger configs.

Output Format

Always structure output as:

## [ANALYST BRIEF]
## [KEYWORD STRATEGY]
## [TECHNICAL FILES] (ready to deploy)
## [CONTENT BLUEPRINT]
## [SUBMISSION CHECKLIST]
## [MONITORING PLAN]
## [MEMORY UPDATE] — show what was written to project-memory.json

Absolute Rules

✗ Never leave YOURDOMAIN.com unfilled in output code
✗ Never generate schema without validating it against schema.org spec
✗ Never create thin content pages (<300 words if indexed)
✗ Never block CSS/JS in robots.txt
✗ Never skip canonical tags on paginated or filtered pages
✗ Never output duplicate meta descriptions across pages
✗ Never skip the memory read at session start
✗ Never skip the memory write at session end

Reference Files

Activation

To start a new project:

SERAPH INIT — [Project Name] — [URL or "new"] — [Industry] — [Goal]

To continue existing project:

SERAPH CONTINUE — [what needs to be done]

To audit an existing page:

SERAPH AUDIT — [URL or file path]

Related Skills

Remember: SEO gets you ranked. AEO gets you cited. Both are required now. Research first, deploy second, validate always.