What problem does it solve? AI assistants often cite competitor pages or outdated URLs because sites lack a curated, machine-readable map of their most important content. This Skill audits existing llms.txt files against a 100-point rubric or generates new ones from sitemap and navigation data, reducing hallucination risk and improving citation accuracy. ## Core Features & Use Cases - Validation & Scoring: Fetches an existing llms.txt, checks HTTP status, format structure, link integrity, description quality, freshness, and crawler reachability, then scores it against a 100-point rubric with severity levels. - Greenfield Generation: Discovers URLs from sitemaps and homepage navigation, prioritizes pages with a P0-P3 matrix capped at 30 curated links, and writes citation-ready descriptions using a factual template. - Companion File & Deployment Strategy: Decides when llms-full.txt is warranted and provides deployment paths for Next.js, WordPress, Nginx, Cloudflare Pages, and Vercel. - Use Case: A B2B SaaS docs site with 200 pages has no llms.txt. Run this Skill to crawl the sitemap, select the top 30 pages across Getting Started, API Reference, Product, and Company sections, and produce a deployable file plus a competitor benchmark table. ## Quick Start Audit the llms.txt file at example.com and generate an improved version with a validation score and competitor comparison.
