seo-geo-audit-skill

name	seo-geo-audit-skill
description	Run a unified SEO and GEO audit for a website, page, or domain. Use when the user asks for a full SEO audit, GEO audit, AI visibility review, EEAT review, entity audit, authority audit, or wants one prioritized report that combines technical findings, content quality, trust signals, and AI citation readiness.
metadata	{"author":"GEO-SEO","version":"1.2.0","homepage":"https://github.com/GEO-SEO/seo-geo-audit","primaryEnv":"SERPAPI_API_KEY","requires":{"env":["SERPAPI_API_KEY"],"bins":["python3"]}}

SEO GEO Audit Skill

Use this skill to produce one audit that covers both search performance and AI visibility, including on-site readiness, off-site entity signals, and platform-specific AI search fit.

Overview

Use this skill when you need one audit that combines technical SEO, content quality, trust, entity clarity, off-site brand visibility, and platform-specific AI readiness into a single prioritized output.

Best For

• brands and SaaS teams that need one shared audit across SEO and GEO
• founders or operators who want a management-ready summary, not a pile of raw issues
• agencies that need a repeatable audit structure across multiple clients
• SEO teams that want to connect technical debt, content quality, off-site authority, and AI visibility in one workflow

Start With

Run a full SEO and GEO audit for https://example.com

Audit this homepage in boss mode: https://example.com

Give me an operator-style SEO GEO audit with P0, P1, and P2 actions

External Access And Minimum Credentials

This audit can run with direct page access only, but some environments may also use optional search or crawl integrations.

• SERPAPI_API_KEY: optional for broader search visibility checks or search-result enrichment
• python3: optional helper tooling in environments that pair this skill with local audit scripts

If no external API is configured:

• continue with direct page and site observations
• state clearly what was observed versus what was not verified
• do not imply access to search-console, private crawl logs, or third-party datasets unless the user provided them

Access Policy

This audit can run without private integrations.

• search-result enrichment is optional, not required
• local helper tooling is optional, not required
• do not claim access to Search Console, analytics, server logs, private crawlers, or proprietary datasets unless the user explicitly provides them
• when data is missing, mark it as Not verified and continue from observable evidence

This skill is designed for three common use cases:

• Homepage audit: quick diagnosis for a single URL
• Site audit: broader review across key templates and pages
• Domain visibility audit: strategic review including entity and authority signals

What This Skill Covers

1. Technical SEO

   • crawlability
   • indexability
   • performance and rendering
   • security and trust headers
   • schema and metadata
   • mobile and accessibility risks

2. On-page SEO

   • title, meta description, headers
   • search intent alignment
   • content depth and structure
   • internal linking
   • media optimization

3. GEO and AI visibility

   • answer-first formatting
   • extractability and quotability
   • semantic clarity
   • AI crawler access signals
   • machine-readable brand and content signals

4. Trust, entity, and authority

   • author and editorial transparency
   • about, contact, and policy signals
   • organization identity consistency
   • entity disambiguation
   • third-party credibility and authority indicators

5. Off-site entity and brand mentions

   • LinkedIn company and leadership presence
   • Reddit discussion and recommendation visibility
   • YouTube channel and transcript-level visibility
   • Wikipedia and Wikidata entity presence
   • GitHub, Product Hunt, Crunchbase, news, podcasts, and industry-community mentions

6. Platform-specific AI readiness

   • ChatGPT and Bing index alignment
   • Perplexity citation friendliness
   • Google AI Overviews answer formatting
   • Gemini entity, schema, and Google ecosystem fit
   • Bing Copilot freshness, IndexNow, and LinkedIn/GitHub support signals

Workflow

1. Define the scope.

   • single page
   • limited site crawl
   • domain-level strategic review

2. Start with the observable baseline. Collect factual findings from live pages, crawl data, page source, rendered output, and any available audit tooling.

   Recommended crawl presets:

   • single page: one-page diagnostic mode
   • fast site pass: capped multi-page sample for template-level review
   • broader site audit: larger sample for structural review
   • deeper review: larger crawl cap with heavier performance/rendering analysis

   Important: this workflow is page-capped rather than fixed-depth. It follows valid internal links until the configured page cap is reached.

3. Separate evidence from judgment. Label each item as:

   • Observed
   • Assessment
   • Not verified

4. Review the site in layers.

   • technical layer
   • on-page/content layer
   • GEO layer
   • entity/authority layer
   • off-site mention layer
   • platform-specific readiness layer

5. Review frontend evidence architecture.

   • Read references/frontend-evidence-audit.md for product, service, ecommerce, manufacturing, education, regulated, or local-service websites with a meaningful buyer journey.
   • Map the homepage, product/category, audience/use case, workflow/integration, pricing/value, comparison, risk/trust, proof, research/data, education, support/docs, and conversion layers.
   • Mark each layer strong, weak, fragmented, missing, or not applicable.
   • Check category clarity, buyer-decision coverage, claim-to-proof linkage, answer extractability, internal discoverability, conversion continuity, localization, freshness, accessibility, and responsive behavior.
   • Use competitor patterns as contextual evidence, not as mandatory templates.

6. Prioritize actions.

   • P0: blockers, trust failures, indexing failures, major performance issues
   • P1: meaningful ranking and citation improvements
   • P2: optimization and scale work

7. Present the result in the right audience mode.

Output Rules

• If the user writes in Chinese, answer in Chinese unless asked otherwise.
• Always begin with a short executive summary.
• Keep technical facts and strategic recommendations distinct.
• Do not invent data for backlinks, Search Console, server logs, or AI citation share.
• Mark unavailable inputs as Not verified.
• Do not recommend public pricing, free tools, comparison libraries, glossaries, programmatic pages, or llms.txt solely because a benchmark competitor uses them.

Audience Modes

Boss mode

Use when the user asks for 老板版, executive summary, or management briefing.

• short
• business impact first
• minimal jargon
• no long issue dumps

Read references/output-template-zh-boss.md before writing.

Operator mode

Use when the user wants a practical roadmap.

• balanced detail
• clear priorities
• implementation-oriented language

Read references/output-template.md before writing.

Specialist mode

Use when the user wants deep analysis.

• include assumptions
• include validation gaps
• include section-level scoring logic

Read references/scoring-framework.md before writing.

Minimum Deliverable

Every audit should include:

• scope
• overall technical view
• content and GEO view
• entity and authority view
• off-site mention and entity validation view
• platform-specific readiness view
• top priorities
• missing data and validation notes

References

• references/scoring-framework.md
• references/output-template.md
• references/output-template-zh-boss.md
• references/frontend-evidence-audit.md