ChatGPT, Claude, Perplexity, Gemini are answering your customers' questions right now. Is your website part of the answer?
geoskills is an open-source Agent Skill suite that audits, fixes, and monitors your website's visibility across AI-powered search engines. It tells you exactly what AI engines can't see — and helps you fix it.
Compatible with Claude Code, OpenCode, OpenClaw, Codex CLI, Cursor, GitHub Copilot, and any AgentSkills-compatible agent.
Traditional SEO tools measure backlinks and rankings. But when users ask ChatGPT "what's the best X?", backlinks don't decide who gets cited — content structure, schema markup, and AI accessibility do.
Research shows the impact is massive:
| Signal | Impact |
|---|---|
| Content optimization strategies | 115-415% visibility improvement (Princeton/Georgia Tech, 2023) |
| Expert quotations in content | 41% more AI citations |
| Statistics and data inclusion | 30% higher citation rate |
| Cross-source brand consistency | 2.5x more likely to be cited |
# Install all skills (recommended)
npx skills add Cognitic-Labs/geoskills
# Run a full GEO audit on your website
/geo-audit https://your-website.com| Skill | What it does for you |
|---|---|
| geo-audit | Find out why AI engines are ignoring your website — get a full score breakdown and prioritized fix plan |
| geo-fix-content | Rewrite your content so AI engines quote it instead of your competitors |
| geo-fix-schema | Generate the structured data that helps AI engines understand what your business does |
| geo-fix-llmstxt | Create an llms.txt file — the robots.txt equivalent for AI engines |
| geo-compare | See how your GEO score stacks up against 2-3 competitors, side by side |
| geo-monitor | Track your GEO score over time and catch regressions before they cost you traffic |
geoskills evaluates your website across 4 dimensions:
| Dimension | Weight | What it measures |
|---|---|---|
| Technical Accessibility | 20% | Can AI crawlers access your content? (robots.txt, llms.txt, HTTPS, speed, sitemaps) |
| Content Citability | 35% | Will AI engines quote your content? (answer blocks, statistics, expertise signals, structure) |
| Structured Data | 20% | Does AI understand your entities? (JSON-LD schema, Organization, FAQ, Product markup) |
| Entity & Brand Signals | 25% | Does AI trust your brand? (cross-platform consistency, knowledge graph, authority signals) |
You get a composite GEO Score (0-100) with sub-dimension breakdowns, severity-ranked issues, and a prioritized fix plan — then use the fix skills to act on it.
npx skills add Cognitic-Labs/geoskillsVia ClawHub
clawhub install geoskillsManual install
# Claude Code
git clone https://github.com/Cognitic-Labs/geoskills.git ~/.claude/skills/geoskills
# OpenCode
git clone https://github.com/Cognitic-Labs/geoskills.git ~/.config/opencode/skills/geoskills
# OpenClaw
git clone https://github.com/Cognitic-Labs/geoskills.git ~/.openclaw/skills/geoskillsInstall a single skill only
# Via skills.sh
npx skills add Cognitic-Labs/geoskills --skill geo-audit
# Manual
git clone --depth 1 --filter=blob:none --sparse \
https://github.com/Cognitic-Labs/geoskills.git \
~/.claude/skills/geoskills
cd ~/.claude/skills/geoskills
git sparse-checkout set skills/geo-audit# Full GEO audit
/geo-audit https://example.com
# Fix skills — generate missing assets
/geo-fix-content https://example.com/blog/post
/geo-fix-schema https://example.com
/geo-fix-llmstxt https://example.com
# Compare competitors
/geo-compare https://mysite.com https://competitor-a.com https://competitor-b.com
# Track progress over time
/geo-monitor https://mysite.com
# In other agents, describe the task naturally:
# "Run a GEO audit on https://example.com"
# "Compare my site with competitor-a.com"Generative Engine Optimization (GEO) is the practice of optimizing content for AI-powered search engines and assistants. Unlike traditional SEO which targets link-based rankings, GEO focuses on making content discoverable, citable, and recommendable by large language models.
geoskills is the diagnostic layer — it tells you what to fix. AIvsRank.com is the measurement layer — it tracks how visible you actually are across AI platforms over time. All skills work fully standalone, no API key required.
Which AI platforms does geo-audit cover? geo-audit checks access for 11 AI crawlers including GPTBot (OpenAI), Google-Extended (Gemini), ClaudeBot (Anthropic), PerplexityBot, Bytespider (ByteDance), Applebot-Extended, CCBot, Cohere, Amazonbot, FacebookBot, and Meta-ExternalAgent.
Do I need an API key? No. All skills work fully without any API key.
What does the GEO Score measure? The composite GEO Score (0-100) weights four dimensions: Technical Accessibility (20%), Content Citability (35%), Structured Data (20%), and Entity & Brand Signals (25%).
How is geoskills different from Ahrefs or Semrush? Traditional SEO tools measure backlinks and keyword rankings. geoskills measures whether AI systems can read, understand, and cite your content — a fundamentally different signal set that determines your visibility in AI-powered answers.
All skills upgraded to align with the Scoring Model v2 introduced in geo-audit v1.1.0.
Scoring alignment (geo-compare, geo-monitor)
- Fixed sub-dimension scores to match
scoring-guide.md— Technical now correctly uses 5 sub-dimensions (35/22/18/13/12) including Multimedia Accessibility - Expanded all 4 dimension breakdowns (Citability, Schema, Brand) with full sub-score tables
New capabilities (geo-compare, geo-monitor)
- Business Type Weight Adjustments — scoring now adapts to SaaS, E-commerce, Publisher, Local, and Agency profiles
- Technical Gate Check — warns when AI crawlers are blocked, providing context for interpreting other scores
GEO-AUDIT-METAmachine-readable block — enablesgeo-monitorto reliably parse historical baselines and support chained monitoring- AIvsRank integration section in reports
- Error Handling for edge cases (unreachable URLs, timeouts, robots.txt blocks)
geo-monitor enhancements
- Parses
GEO-AUDIT-METAblock from baseline reports (with Markdown fallback for older reports) - Outputs its own META block, enabling chained re-audits without manual setup
- Scoring model version check — warns when comparing v1 vs v2 reports
geo-fix-content
- Defined explicit Citability scoring rubric (6 metrics, 100-point scale with clear thresholds)
- Added GEO Score impact context (Content Citability = 35% of composite score)
geo-fix-llmstxt
- Added GEO Score impact context (llms.txt = 7 points under Technical → Rendering)
geo-fix-schema
- Added GEO Score impact context (Structured Data = 20% weight, 4 sub-dimensions detailed)
All skills
- Added Error Handling sections for common failure modes
- Version bumped to 1.1.0
Contributions welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/improvement) - Commit your changes
- Push to the branch (
git push origin feature/improvement) - Open a Pull Request
- Additional business type profiles and scoring adjustments
- Language-specific citability heuristics (non-English hedge word dictionaries)
- New AI crawler detection rules
- Schema template improvements
- New fix skills (e.g., geo-fix-robots, geo-fix-meta)
Apache-2.0 — see LICENSE for details.
Built by AIvsRank.com — AI Visibility Measurement Platform