GEO-SEO skill for Claude

GEO-SEO checks how AI assistants such as ChatGPT, Perplexity and Google's AI Overviews can reach and quote a website, then looks at the usual SEO groundwork underneath. Its README calls the approach GEO-first and SEO-supported. It's a community skill by Zubair Trabzada under the MIT license, built for Claude Code, and Anthropic doesn't publish it. For the basics, see how Claude Skills work.

What the skill does

The main skill installs as geo. The repository holds 15 sub-skill folders and five subagents. A full audit (/geo audit) reads the homepage, guesses the business type, pulls up to 50 pages from the sitemap and runs the five agents at once. You get a GEO Score from 0 to 100 across six weighted areas, where AI citability and visibility count most. The parts a site owner will use:

• /geo citability grades each passage on whether it answers a question in its first sentences, makes sense without the text around it and carries specific facts, then suggests rewrites.
• /geo crawlers reads robots.txt, meta tags and HTTP headers to map which AI crawlers, such as GPTBot, ClaudeBot and PerplexityBot, can reach the site.
• /geo llmstxt checks an existing llms.txt or writes one ready to upload.
• /geo brands scores the brand's presence on YouTube, Reddit, Wikipedia and LinkedIn, then Quora, Stack Exchange, GitHub, forums, news and podcasts.

Schema, technical SEO and E-E-A-T checks complete the audit, and /geo quick prints a short snapshot without writing a file. The rest is agency tooling: /geo report-pdf builds a client PDF, while /geo prospect and /geo proposal keep a small sales pipeline and draft proposals in ~/.geo-prospects/.

When to use it

• You suspect robots.txt or a CDN rule keeps AI crawlers off your pages.
• A page ranks in Google, and you'd like its passages to be easier for AI answers to quote.
• You need a first llms.txt draft to edit.

For a wider audit with Search Console data and its own AI search command, use Claude SEO. To pair an audit with copywriting, install the Marketing Skills set. With no install at all, paste one of our SEO prompts into a chat.

How to install it

The README's quickest install pipes a remote script straight into bash. The manual route lets you read install.sh before it runs:

git clone https://github.com/zubair-trabzada/geo-seo-claude.git
cd geo-seo-claude
./install.sh

You'll need Python 3.10 or newer, Git and Claude Code. The installer copies the skills to ~/.claude/skills/ and the agents to ~/.claude/agents/, and keeps its Python packages in a separate environment at ~/.claude/skills/geo/.venv/, away from your system Python. On Windows the README asks for Git Bash and ./install-win.sh. GEO-SEO is a plain skill folder, so there's no /plugin command for it.

Example request

/geo audit [your site URL]
The site is a [type of business] for [audience]. After the audit, tell me which AI crawlers are blocked and whether that looks deliberate. Then pick the three passages on [key page URL] that most need work for citability and give a rewrite for each.

Limits

• Its description tells Claude to load it on words like "seo", "audit" or "optimize", or on any URL, so it can start on requests meant for another skill. Name the /geo command you want.
• The crawler skill recommends allowing GPTBot, though its own entry says content GPTBot reads may be used to improve OpenAI's models. Whether to let training crawlers in is your call.
• The README and SKILL.md open with market figures on AI traffic and conversion that the repository doesn't link to any source. Read them as the author's framing; the audit doesn't measure them.
• /geo report-pdf needs pandoc and Google Chrome at the standard macOS path.
• The uninstaller leaves ~/.geo-prospects/ in place, so delete that folder by hand if you tried the sales tools.
• The README closes with a link to the author's Skool community, pitched at people who want to sell GEO audits to clients.