When a site scores 4/6
Usually discovery is present but one machine-readable delivery path is missing, such as Link headers or markdown negotiation.
Sample reports
AI agent readiness examples
A concrete demo report for AgentScan itself, followed by common score patterns. Use it to understand what a passing result looks like and what evidence should sit behind each check.
https://agentscan.devmahmoud.me/
AgentScan homepage self-audit
Demo report updated May 29, 2026
This demo report is based on the current AgentScan production configuration and links each passing check to the shipped resource or behavior.
robots.txtpass/robots.txt returns text/plain with wildcard and AI crawler groups.
sitemap.xmlpass/sitemap.xml is generated from the current public routes.
Link headerspassThe homepage response advertises api-catalog, llms.txt, sitemap, RSS, and Atom resources.
Markdown negotiationpassRequests with Accept: text/markdown are rewritten to a markdown homepage response.
AI bot rulespassrobots.txt declares explicit rules for GPTBot, ChatGPT-User, ClaudeBot, anthropic-ai, and PerplexityBot.
Content Signalspassrobots.txt declares Content-Signal: ai-train=no, search=yes, ai-input=no.
Common score patterns
When a site scores 2/6
The site is often basic-search-crawlable, but it has not declared AI crawler policy or agent-readable references.
When a site scores 6/6
The site exposes crawl policy, canonical URL discovery, machine-readable references, alternate markdown delivery, and explicit AI usage signals.
How to read the score
Discovery
Can agents find robots.txt, sitemap.xml, and machine-readable references without guessing?
Fetch behavior
Do public pages respond consistently for crawlers and support useful alternate formats such as markdown?
Policy clarity
Are AI crawler access, training preferences, and content-use signals declared explicitly?
Check your own site
Run the same six checks on a live public URL and get implementation prompts for failed items.