SEO/GEO Page-Level Keyword Optimization Skill

Overview

This skill drives page-level keyword optimization by combining Ahrefs MCP data with GEO (Generative Engine Optimization) principles. It targets both traditional search engines and AI-powered search (ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini).

Inputs Required

Workflow

Phase 1 — Keyword Intelligence (Ahrefs MCP)

1. keywords_explorer(seed=primary_keyword)
   → Extract: volume, KD, CPC, related_terms, questions, parent_topic

2. site_explorer(url=competitor_url, mode="organic_keywords", limit=50)
   → Repeat for each competitor
   → Collect all keywords competitors rank for on this topic

3. content_gap(target=our_url, competitors=[url1, url2, ...], mode="page")
   → Output: keywords competitors rank for that our page does NOT cover
   → Filter by: volume > 100, KD \x3C 60

Deliverable: A keyword matrix with three tiers:

Phase 2 — On-Page Keyword Mapping

Map keywords to page structure following this hierarchy:

\x3Ctitle>          Primary keyword (≤60 chars)
\x3Cmeta desc>      Primary + secondary keyword + CTA (≤155 chars)
\x3Ch1>             Primary keyword variant (natural language)
\x3Ch2> × N         One secondary keyword per sub-topic
First 100 words  Primary + 1 secondary keyword (natural, not forced)
Body             LSI / gap keywords distributed at ~1 per 200 words
\x3Cimg alt>        Descriptive keyword-inclusive alt text
Internal links   Anchor text = target keyword of linked page

Rules:

• No keyword stuffing. Density is not a metric. Relevance and coverage are.
• Each H2 owns exactly one keyword intent. Do not blend intents in a single section.
• Internal link anchors must use the exact target keyword of the destination page.

Phase 3 — GEO Layer (AI Search Citability)

This phase targets AI engines (ChatGPT, Perplexity, Gemini, AI Overviews).

CITE Framework (one application per major section):

C — Claim     : State the answer directly in the first sentence.
I — Insight   : Add a non-obvious detail or mechanism.
T — Trust     : Cite a source, statistic, or expert reference.
E — Evidence  : Back it up with data, example, or case.

Structural requirements:

1. Opening definition block: First 50 words must answer the primary keyword query in one authoritative sentence. This is what AI engines extract and cite.

2. FAQ Schema: Convert Ahrefs question-type keywords into FAQ markup.

   {
     "@type": "FAQPage",
     "mainEntity": [
       {
         "@type": "Question",
         "name": "[keyword question from Ahrefs]",
         "acceptedAnswer": { "@type": "Answer", "text": "[CITE-structured answer]" }
       }
     ]
   }
   

3. Table / List structures: AI engines prefer extractable formats. Where the content allows, convert prose into a table or bulleted list with a clear header.

4. Factual density: Include at least 3 verifiable facts (statistics, dates, named entities) per 500 words. Vague content is not cited.

Phase 4 — Monitoring (Ahrefs API)

# Weekly page health check
results = ahrefs_api.get_organic_keywords(url=target_page, limit=100)

for kw in results:
    if kw["position"] > 10 and kw["volume"] > 200:
        trigger_reoptimize_alert(kw)

# Monthly content gap re-run
ahrefs_api.content_gap(target=our_url, competitors=competitor_list)
→ Flag any new gap keywords with volume > 300

Alert thresholds:

• Primary keyword drops below position 10 → immediate review
• Click share drops >20% month-over-month → full content audit
• New gap keyword with volume >500 appears → add FAQ or new section within 2 weeks

Priority Order

Reference Projects

Ahrefs MCP Endpoints Used