Project honeypot integration?

Botnets are constantly trying to brute force into this or that e-mail account in my server. The limits help, but they just keep trying slowly and with different IP addresses every time. It's aoying to receive notifications and to have to block them manually.

It just so happens that I noticed many of these IPs are already known as malicious by project honeypot. CSF can retrieve the list of IPs from their RSS feed, but it only provides the 25 latest entries (likewise for the user interface; 50 if you're logged in). But they do seem to have an API with various example implementations.

Has anyone been able to integrate their DA/dovecot install with PH in order to automatically check and block addresses that are blacklisted before they attempt to log in? Any ideas?