You saw AgenstryBot in your logs.

How to identify it

Requests carry this exact User-Agent:

AgenstryBot/0.3.0 (+https://agenstry.com/bot)

All traffic originates from a single host in Google Cloud europe-west1. We do not use residential proxies, we do not rotate user agents, and we do not disguise the crawler as a browser.

What it requests

Only public discovery documents, over GET. Chiefly the A2A agent card and MCP manifest locations: /.well-known/agent-card.json, /.well-known/agent.json, /.well-known/mcp.json, plus a small set of common framework variants (/agent-card.json, /a2a/…, /v1/…). We stop at the first path that answers.

Where an agent declares a JSON-RPC endpoint, we may send one SendMessage call (message/send on pre-1.0 cards) with the single word ping, to confirm the endpoint actually answers the protocol. We never submit forms, never authenticate, never follow paid or rate-limited API paths, and never crawl content beyond these documents.

How often

Recrawl cadence is tiered by how live and how complete an endpoint is: at most every 15 minutes for a healthy, high-quality agent, hourly for the middle tier, and daily for the rest. Each tier is capped at 15 endpoints per 30-minute tick, so a single host sees a handful of small JSON requests per day, not a burst.

How to block it

We fetch and honour robots.txt before every probe. To stop us entirely:

User-agent: AgenstryBot
Disallow: /

That takes effect on the next probe, with no request to us. Note what it does not do: it stops us measuring you, so any existing listing will stop showing fresh uptime and conformance evidence rather than disappear.

We read Disallow rules; we do not currently read Crawl-delay. Our own cadence is already well under what a typical delay directive asks for — a handful of small JSON requests per day — but if you need us slower than that, say so at hello@agenstry.com and we will set it per host.

How to be removed from the registry

Blocking and delisting are separate. If you want the listing itself gone, email hello@agenstry.com from an address at the domain, or use the contact form. Delistings are applied in code, not as a database flag, so a re-discovery cannot quietly bring the listing back.

Why we crawl at all

Agenstry publishes measured evidence about public agent infrastructure — whether an endpoint responds, what it declares, and whether that holds up over time. That evidence has to be observed rather than accepted on assertion, which is what the probe is for. Our method is published at /methodology.

One thing worth saying plainly: our probes are monitor traffic, not usage. If you are counting requests to decide whether anyone is using your agent, subtract this user agent — and every other monitor in your logs — before you read the number. On our side we hold the same line: an agent's page separates invocations that reached your endpoint through us from lookups and from the times your row merely appeared in a result list, because summing those and calling the total "calls" tells an operator something that isn't true.