Post #3474224

@alienghic Makes me wonder when crawlers start packing reCAPTCHA-grade ML for this. Clean RL problem: agent walks a honeypot URL, IP/ASN gets banned, eventually the correlation surfaces and the policy learns to avoid traps. Guess not enough sites do it yet for the signal to bite. Genuine q: is your fail2ban purely local, or does it share ASN/CIDR bans via a community pool? If per-site, a crawler's done with you before the ban bites — kind of CrowdSec's pitch over fail2ban.