In my opinion to be sure that protection mechanisms work, you should know what attacks can be performed against them. Only if you know techniques that can be used to attack you, you can be sure whether you are protected. If you go to Sec.SE asking about attack protection without knowing full attack profile, you can't judge whether those protection mechanisms work. So before knowing protection measures, you should know attack itself.
Based on this, I think that questions asking about how to perform an attack seem to be good for me as this knowledge is needed to be able to judge protection measures.
For example to be sure that my protection measures catching scan anonymizers work, I should know techniques that can be used to anonymize scans.
To figure out how to build smartphone that will be able to notify user in case of attack, I should know how those attacks work.
Arguments like this can be devised in most of situations.
I vote to allow content like this to be posted on site.