geo-audit

Diagnose why AI can't find, cite, or recommend your website.

geo-audit is an open-source Claude Code skill that runs a comprehensive Generative Engine Optimization (GEO) audit. It tells you exactly what's blocking ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews from discovering and citing your content — and gives you a prioritized fix plan.

Repository: github.com/Cognitic-Labs/geoskills

Installation

Via skills.sh

# Install geo-audit only
npx skills add Cognitic-Labs/geoskills --skill geo-audit

# Or install all geoskills
npx skills add Cognitic-Labs/geoskills

Via ClawHub

clawhub install geoskills

Manual Install

git clone https://github.com/Cognitic-Labs/geoskills.git ~/.claude/skills/geoskills

Usage

In Claude Code, run:

/geo-audit https://example.com

No configuration required. All four analysis agents launch in parallel automatically.

AIvsRank Integration

AIvsRank API integration is coming soon. All skills work fully without any API key.

How It Works

geo-audit evaluates your site across four dimensions using a 3-layer model:

+---------------------------------------------+
|              SIGNAL LAYER (25%)             |
|         Entity & Brand Signals              |
|   Wikipedia · LinkedIn · Reddit · Reviews   |
+---------------------------------------------+
|             CONTENT LAYER (35%)             |
|           Content Citability                |
|  Answer Blocks · Stats · Structure · E-E-A-T|
+---------------------------------------------+
|              DATA LAYER (40%)               |
|   Technical Access (20%) + Schema (20%)     |
|  AI Crawlers · SSR · llms.txt · JSON-LD     |
+---------------------------------------------+

Composite formula: GEO Score = Technical×0.20 + Citability×0.35 + Schema×0.20 + Brand×0.25

What Gets Analyzed

Technical Accessibility (20%)

  • AI crawler access: GPTBot, ClaudeBot, Google-Extended, PerplexityBot, and 7 more
  • robots.txt and X-Robots-Tag headers
  • Server-side rendering (SSR) vs. client-side rendering (CSR)
  • llms.txt presence and completeness
  • HTTPS, response time, compression
  • Sitemap, meta tags, Open Graph, canonical URLs

Content Citability (35%)

  • Answer block quality (Q&A patterns, definitions, FAQ)
  • Self-containment (passages understandable in isolation)
  • Statistical density (numbers, sources, data recency)
  • Structural clarity (headings, lists, tables, paragraph length)
  • Expertise signals (author bylines, expert quotes, publication dates)

Structured Data (20%)

  • Organization / LocalBusiness schema with sameAs
  • Article / BlogPosting schema with author and dates
  • Speakable property for AI voice assistants
  • FAQPage, HowTo, BreadcrumbList schemas
  • JSON-LD format and syntax validation
  • Auto-generated fix templates for missing schemas

Entity & Brand Signals (25%)

  • Wikipedia / Wikidata entity presence
  • LinkedIn, Crunchbase, industry directory listings
  • Reddit discussions, YouTube presence
  • Cross-source brand name and description consistency

Sample Output

GEO Audit: example.com
   Business type: SaaS (detected)
   Pages to analyze: 8

Running 4 parallel analyses...
   Technical Accessibility: 72/100 (3 issues)
   Content Citability: 58/100 (5 issues)
   Structured Data: 45/100 (4 issues)
   Entity & Brand: 81/100 (2 issues)

GEO Score: 62/100 (Grade C: Developing)

Full report: GEO-AUDIT-example-com-2026-03-12.md

The generated report includes:

  • Score breakdown with sub-dimensions
  • Prioritized issue list (Critical to Low)
  • Specific fix instructions per issue
  • Ready-to-use JSON-LD templates
  • Top 5 quick wins with expected point gains
  • 30-day improvement roadmap

Score Grades

Grade Range Meaning
A 85-100 AI-optimized. Likely cited by major AI engines.
B 70-84 Solid foundation. Targeted fixes push to A-tier.
C 50-69 Significant gaps. Structured 30-day plan needed.
D 30-49 Major issues. Prioritize critical fixes first.
F 0-29 Fundamental problems blocking AI discovery.

Business Type Adjustments

geo-audit auto-detects your business type and adjusts scoring weights accordingly:

Type Key Adjustments
SaaS Rendering +10%, Answer Blocks +10%, FAQ/HowTo schema +15%
E-commerce Product schema +20%, Statistical Density +15%, Reviews +15%
Publisher All citability +10%, Article schema +15%, Entity Recognition +10%
Local LocalBusiness schema +25%, Directories +20%
Agency Entity Recognition +15%, Expertise Signals +15%

Research Foundation

The scoring methodology synthesizes findings from 101+ sources:

  • Aggarwal et al. (2023) — "GEO: Generative Engine Optimization" (Princeton / Georgia Tech)
  • BrightEdge (2024-2025) — AI citation correlation studies
  • Google Search Central — Schema.org and rich results documentation
  • SparkToro / Zyppy — Zero-click search and AI answer source analysis
  • Backlinko, Ahrefs, Semrush, Moz, and 90+ industry publications

Diagnostic vs. Measurement

geo-audit identifies what to fix. AIvsRank.com measures how visible you actually are across AI platforms over time — tracking real mentions in ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews.

Together, they give you the complete picture.

Previous geoskills