How to recognise it.
| User agent | Mozilla/5.0 (compatible; SneezeAgentReady/1.0; +https://agentready.sneeze.it/bot) AgentReadyBot |
| robots.txt token | AgentReadyBot |
| Signed requests | Web Bot Auth (HTTP Message Signatures, Ed25519). Signature-Agent: "https://agentready.sneeze.it".
Public keys: https://agentready.sneeze.it/.well-known/http-message-signatures-directory |
| Operator | Sneeze It, 277 Fairfield Rd, Fairfield, NJ 07004. info@sneeze.it |
What one scan requests.
About fifteen requests, most of them in parallel, then it stops. It reads public pages only, never logs in, never submits a form and never places an order.
/robots.txtfirst. If it disallowsAgentReadyBot(or*) for the page asked about, the scan ends there and nothing else is requested.- The page asked about, a few of its same-origin scripts,
/llms.txtand/sitemap.xml. - Discovery documents:
/.well-known/mcp.json,/.well-known/model-context.json,/.well-known/mcp-manifest.json,/.well-known/ai-plugin.json,/.well-known/oauth-protected-resource,/.well-known/oauth-authorization-server,/.well-known/openid-configurationand/.well-known/agents.json. - The sign-in page linked from the homepage, read only, when robots.txt allows it.
- The page again, once unsigned and once signed, and a third time if the second is refused, to tell a block from a rate limit.
- On stores only, a handful of commerce discovery and catalogue feed paths.
One identity, always. Every request carries the user agent above, including pages rendered in headless Chromium (used when a page is empty until its scripts run, because many agents drive a real browser). If your site refuses us, we stop and report that. We never retry with a different identity or a disguised browser.
How to stop it.
Add this to robots.txt. It is obeyed before any other request is made:
User-agent: AgentReadyBot
Disallow: /