Reverse mail proxy conceptual question

Hello everyone,

I have a system running freebsd in a DMZ (A), accessing a system in the internal network (B) as a reverse proxy via HTTPS. Now I wanted to use it as a proxy for mail on B aswell.
I wrote an auth script in Python which just does this:

print "Content-type: text/html\n"
print "Auth-Status: OK"
print "Auth-Server: 10.0.0.23"
print "Auth-Port: 993"

I tried to test access 993 via telnet but it does not work at all, when I try to authenticate it just throws an error and exits, no authentication failure or anything.
I dont want nginx to do authentication, this can happen on the server that it proxies to, doesnt it?!

So far this has been quite some pain, because the documentation does not explain that you even need this auth script and what for.
Can someone explain the sense behind the conceptual decision to not just ask for IP and port in mail proxy config and instead force users to write some weird auth script? Does not make any sense to me at this point.

Hope someone can shed light upon this