geo-audit

Run a GEO audit that checks for llms.txt and llms-full.txt files, AI-crawler directives in robots.txt, JSON-LD citation metadata, and snippet-extractable content structure. The audit reports PASS, WARN, or FAIL per check with severity ratings.

GEO (Generative Engine Optimization) targets LLM chatbots like ChatGPT and Perplexity citing your pages in answers. AEO (Agentic Engine Optimization) targets coding agents like Claude Code and Cursor preferring your docs over training data when writing code against your product.

llms.txt is a discovery file at your site root listing key URLs for LLM crawlers, with llms-full.txt as its companion. Missing llms.txt is a high-severity FAIL in GEO audits because it is the primary entry point for AI discovery.

No. The skill is gated to web projects with a public content surface such as pages, docs directories, or web-framework markers like Next.js or Astro. Backend-only APIs report PASS immediately since there is nothing for LLM crawlers to index.

Token counts are estimated as character count divided by 4, a cross-vendor heuristic. Thresholds flag Quick Starts over 15K tokens, API references over 25K, and conceptual guides over 20K as failures, since oversized pages exceed practical LLM context budgets.

No. The audit only reports findings to chat with severity grades; it does not generate llms.txt, scaffold AGENTS.md, or split oversized pages. The operator decides which gaps to address.