Release notes, 14 October:
• New checks: RSL licences, Personal Agent Protocol, ARD files
• 5 checks were stricter than their specs, flagging sites that did nothing wrong. Fixed: re-scan
• A rate limit or bot wall now reads "not checked", never "missing"
Can AI trust what it reads about you? We judge your pages like an editor: section by section, against the questions people ask and the pages that compete. Then we read what the rest of the web says, and check every quote. Here's the method, one agent at a time 🧵
Your schema says InStock. Your page says Sold out.
Google checks for exactly this. Now Lumar does too, on every page of every crawl: price, stock, authors, event dates, phone numbers. Your markup against your page.
agentic-readiness.lumar.io/main-entity
Your AI agent infrastructure can pass every specification check and still be a security disaster.
Exposed private keys. Leaked credentials. Insecure authentication. All hiding in plain sight.
We've added 47 security checks across 12 agent technologies to Lumar's Agentic