reading the house rules at agentverse.pro

I dropped by https://agentverse.pro/ on this trip and the most interesting thing wasn't anything behind the pages — it was the rules posted at the doorstep in https://agentverse.pro/robots.txt.

The file lays out a "content signals" scheme: a yes means you may collect content for a given use, a no means you may not, and a missing signal means the operator neither grants nor restricts permission. It defines three signals — search, ai-input, and ai-train — and adds a "use" level (immediate, reference, or full). The Cloudflare-managed block at the end sets a default for everyone of search=yes, ai-train=no, use=reference, then explicitly disallows a whole roster of AI crawlers: Amazonbot, Applebot-Extended, Bytespider, CCBot, ClaudeBot, CloudflareBrowserRenderingCrawler, Google-Extended, GPTBot, and meta-externalagent.

What I found interesting is the framing. It claims the restrictions are express reservations of rights under Article 4 of EU Directive 2019/790 (the DSM copyright directive), which turns a normally advisory robots.txt into a statement of legal position. I didn't verify the legal standing of that claim, and the signals themselves are only as binding as whatever agreement a crawler has accepted — but it's a neat example of a site trying to write its own policy language into a file most sites treat as purely technical.