Check a prompt

What do Claude SEO skills need to run?

By GEO Search Console Published

Claude SEO skills need three things to run: an assistant that reads SKILL.md files, a place where it can run code, and whatever the skill's script calls.

Our two free skills need Python and, for the Common Crawl check, internet access to Common Crawl. This page is part of our free Claude SEO skills guide.

Python for our Claude SEO skills

Our Claude SEO skills need Python 3.8 or newer for their helper scripts, and nothing else from Python.

Both scripts use only Python's standard library, such as urllib.request for web requests, so there is no package to install and no API key to set.

Where python isn't Python 3, the skills tell the assistant to use python3, or py on Windows. If the organization schema skill can't run Python at all, it follows the same rules by hand from its reference notes.

Code execution in Claude

Code execution is what lets Claude run a skill's script.

In Claude on the web and desktop, turn on "Code execution and file creation" under Settings, Capabilities before you upload a skill. On Team and Enterprise plans the setting sits in the organization's settings, so an owner or admin turns it on.

Claude Code and the Codex CLI run scripts on the computer they run on, so Python has to be installed there.

Internet access for the Common Crawl check

Internet access matters most for the Common Crawl check, because its script asks Common Crawl's index at index.commoncrawl.org and reads its archive at data.commoncrawl.org.

  • On claude.ai and in the Claude apps, the code sandbox's internet access is set under Settings, Capabilities: it has to allow those two hosts, or all domains.
  • On a Team or Enterprise plan, an admin may have to allow them.
  • In Claude Code, you can allow the two hosts in its sandbox or permission settings.

When it can't reach them, the skill gives you a link to our Common Crawl checker, which runs the same check in your browser. The organization schema skill reaches the web only to read your site's current schema, and it can work from a saved copy of the page or from your answers instead.

Free tools for this guide

  • 1 Reach In your browser

    Crawl Record

    Is your page in Common Crawl?

    Open the Common Crawl checker

References

  1. Use skills in Claude, Claude Help Center.
  2. urllib.request: Extensible library for opening URLs, Python documentation. docs.python.org/3/library/urllib.request.html
  3. Extend Claude with skills, Claude Code documentation.