What Happened
A developer published geo-lint, an open-source content linter that runs 92 deterministic checks across four categories: traditional SEO, Generative Engine Optimization (GEO), content quality, and technical issues. It works on Markdown and MDX files, and ships with a Claude Code skill that can auto-fix violations in a loop.
The tool is built around GEO - the practice of optimizing content so AI search engines like ChatGPT, Perplexity, and Claude are more likely to cite it. The rules cover things that matter for LLM citation: question-formatted headings, FAQ sections, entity density, E-E-A-T signals, and citation-ready statistics. These are distinct from what Google's traditional algorithm rewards, and most content teams aren't tracking them yet.
The project is available on GitHub under an open-source license and can be integrated directly into Claude Code as a skill, letting it lint and fix content automatically.
Why It Matters
If you publish content and care about traffic, you're now playing two games at once. Google still sends the most clicks, but a growing share of discovery happens through AI chat interfaces. When someone asks ChatGPT or Perplexity a question, the model pulls from sources it considers authoritative and well-structured - but the signals it looks for aren't identical to what Google rewards.
Most SEO tools like Clearscope and Ahrefs focus entirely on traditional search. They'll tell you about keyword density and backlink profiles, but they won't flag that your headings aren't phrased as questions (which LLMs prefer when generating cited answers) or that your statistics lack proper attribution (which makes them less likely to be quoted).
Having 92 rules codified into a linter means you can catch these issues before publishing rather than wondering why your well-ranked content never shows up in AI-generated answers. The Claude Code integration is particularly useful - instead of manually fixing each violation, the skill loops through them automatically.
Our Take
GEO as a concept has been floating around SEO circles for about a year now, but tooling has lagged behind. Most advice amounts to blog posts listing best practices you're supposed to remember while writing. A linter that enforces rules mechanically is a better approach.
The 92-rule count is worth scrutinizing. Some of these rules are well-established (proper heading hierarchy, meta descriptions) while GEO-specific rules are still based on early research and practitioner observation rather than confirmed ranking factors. AI search engines don't publish their citation algorithms, so some of these checks are educated guesses.
That said, the approach is sound. Even if individual GEO rules evolve, having a pluggable linting framework means you can update rules without rebuilding your workflow. And the auto-fix loop through Claude Code is the right pattern - nobody wants to manually restructure 50 headings into question format.
If you're already using Claude Code for content work, this is worth adding to your setup. It's free, the rules are transparent, and the worst case is your content gets more structured - which helps traditional SEO anyway.