Spamming check and control interface
I would like suggest you to make a tool for restricting the spamming (both compromised account and injected scripts)
Well, the idea is good but it already exists:
For the spamming, well, there is no way, csf + scan outgoing mails could be enough. But it would be great for a possiblity to limit network though.....