This is interesting. I've wanted something like this but it might be difficult to implement reasonably. Depending on the relationships and topographical distances between HM, SMTP server and POP3 server (not necessarily all close - and POP3/IMAP and SMTP may be on different boxes).
For example, delivery delays can be introduced by a number of normal/expected factors - eg mail queue processing interval. So, once HM sends test email, how long or how many POP3 checks occur before we consider test failed? In my environment, which can involve several commercial email providers, delays in delivery of 1-2 hours is not unheard of.
Also, inverting results may be useful for monitoring 3rd party relay policy adherence. For example, if email IS received, then a problem (3rd party relay) exists.
Yeah, implementation is not trivial.
Also SMTP and POP3 servers can be located in different networks. In this case HostMonitor should request agent A to send e-mail and after a while HostMonitor should request agent B to check mail...