← ClaudeAtlas

ai-search-auditlisted

Audit a website or its codebase for what AI assistants and search crawlers can read, and fix it, covering AI crawler access in robots.txt, text that only appears after JavaScript runs, indexing and snippet rules, titles, headings, canonical links, structured data, sitemap and llms.txt. Use when someone asks whether ChatGPT, Claude, Perplexity, Gemini or Google's AI Overviews can read their site, for an AI SEO, GEO or AEO audit, or why their pages never appear in AI answers. Reports observed facts and a fix list, never an invented score.
visibian/seo-geo-skills · ★ 0 · AI & Automation · score 75
Install: claude install-skill visibian/seo-geo-skills
# AI search audit Find out what an AI assistant or search crawler can actually read on a site, and fix what stops it. Work from observed facts: every finding names the page, what was seen and why it matters. ## Before you start - Ask for the site's address and the 3 to 10 pages that matter most (home, pricing, main product or service pages, the best guide). If you are in the site's codebase, find the routes yourself and ask only which pages matter most. - Tell the user what this covers: whether the pages can be read and understood. Whether an assistant then names the brand depends on what the pages say and on other sites, which this audit does not measure. ## Steps 1. Check each key page, from this skill's folder: `python3 scripts/page_check.py https://example.com/pricing` It reports the answer code, indexing and snippet rules, title, description, canonical link, headings, how many words are in the HTML before any script runs, signs that the text is added by scripts, structured data and dates, and, for a live page, which AI crawlers robots.txt allows on that path, plus the sitemap and llms.txt. In a codebase you can also point it at a built page with `--file`. If the script cannot run, fetch the raw HTML yourself (without running scripts) and check the same things by hand. Say that you did. 2. Check crawler access across the site with the `ai-crawler-access` skill's method if any search crawler or user fetcher is closed, or if robots.txt could not be read. 3. Ch