helping someone with a task that could turn out to be, somewhat of a malicious task, is that something we should answer here in our community?
“That could turn out to be (…) malicious” is not a reasonable criterion. Any question could turn out to have a malicious purpose. As an example, let me take a random question¹:
E-tender system
Gee, a tender system. This could be used to handle bids for a criminal act. Should we ban that question on this basis? Of course not!
I presume that you really meant “a task that is likely to be malicious”. The problem is, how do you determine that a task is likely to be malicious?
You can't rely on the asker's intent. You can't rely on their intent because you can't know what their intent is, only what they've stated. And even if the intent was clearly stated and you believed them, it wouldn't be relevant — maybe the next visitor who browse that question will have a different intent.
Stack Exchange used to have a site about firearms. (It closed due to lack of traffic.) A firearm is fundamentally designed to cause harm. Now firearms can surely be used for good, but that's the decision of the person wielding it. Any firearm recommendation would be likely to run afoul a “likely to be malicious” criterion, since it could be used straightforwardly to kill someone.
So we have a clear precedent that “likely to be malicious” isn't the right line to draw. Rather, the line is something like “unlikely to be useful for benign purposes”.
Challenge: find an otherwise suitable question for this site that is unlikely to be useful for benign purposes.
Where can I find a stealth keylogger for Windows? demonstrably does not fit the bill: the question explicitly lists a benign purpose — “I would like to test detection of a stealth keylogger …”
On Security Stack Exchange, “black hat” questions, i.e. questions on how to attack a system, are permitted. This is not surprising since many of the active users there are professional penetration testers — people whose job is to look for vulnerabilities in IT infrastructure and report it to said infrastructure's owner. But this isn't the only reason. I'm an application developer, and the products I make are security sensitive. My job is pretty fundamentally white hat: a big part of my job is to make a product that resists attacks. In order to do that, I need to understand the attack techniques! The black hat content on Security Stack Exchange is useful to me because I'm a white hat.
I have no idea whether the asker of the keylogger question is, in fact, testing its detection. But this is the sort of question I might ask (mutatis mutandis, I don't work on Windows administration). So I personally see this type of questions as useful, precisely because my intent is not malicious².
¹ Methodology: I drew random numbers between 1 and 26555 (the latest question as I write) until I found one that was the ID of an open question (i.e. excluding answers, closed questions, tag wikis, etc.).
² Or isn't it? For all you know, my job is to protect weapon control systems used by <insert your most-hated dictator here>.
