The fallacy consists in claiming that a person's conclusion should be rejected because they have a bad character or have an ulterior motive. This is a fallacy because I don't have good grounds for saying that the conclusion is false. A bad person can occasionally offer a good argument, and a conclusion can be plausible even if it's argued for by someone of suspect character.
It can't be quite true that this is what the fallacy consists in, because suppose the person's conclusion is:
I have a good character
or
I have no ulterior motive.
Then arguing that this conclusion should be rejected because, in fact, such-and-such shows that they have a bad character or an ulterior motive, is not fallacious: you do have good grounds for saying the conclusion is false. And, of course, it doesn't have to be anywhere near this obvious: exactly the same thing will happen if the conclusion is something else entirely but can in context reasonably be taken to imply either one of these two propositions, or any proposition of a similar sort.
Should we be in a situation in which the opponent's claim does not imply anything like these two propositions, then we would indeed have the fallacy; but what the fallacy would consist in would simply be irrelevance: the objection does not affect the truth value of the claim being made.
There's a deeper issue here than it might seem. It's true, of course, that a bad person can occasionally offer a good argument and that a conclusion can be plausible even if it's argued for by someone of bad character. But it's equally true that invalid arguments can occasionally be good arguments and that conclusions can be plausible even if the argument used to reach it is an incoherent argument. What is important in those cases is that validity and coherence of the underlying argument is relevant to assessing conclusions in general. And thus the questioner, I think, has a legitimate question: why wouldn't one say that a person's character is relevant to assessing their conclusions? We know, for instance, that there are biases. We know also that, when faced with a very ingenious dishonest person, we ourselves would factor that into our assessment of whether any argument they are likely to produce can be accepted as a good argument, even if we ourselves can't discover the flaw -- because any of us might, if put in the right circumstances, refuse to accept an argument that seems sound, for no other reason than that the character of the source makes us suspect a trap. That's not a formal reason for rejecting an argument, of course; but we aren't in the realm of formal logic, anyway. Rather than simply repeating the cliche, we should perhaps ask ourselves the question: Why wouldn't the character of the arguer always be relevant to evaluating the argument?
My own (rough and incompletely developed) view on this question is that, in fact, it always is to the particular instance of the argument, but not in the same way in every instance; and this is a rather important distinction. What allows us to say that a particular attack on someone's character as a response to an argument is bad, for instance, is not that people with bad characters can have good arguments, but that people with bad characters can give arguments that people with good characters could also give. It is this that lets us abstract from the character of the particular arguer in question. But it's a much trickier question than it seems on the surface. As with most of what we call 'informal logic' we are dealing not with a systematic account of reasoning but with a widespread philosophical folklore patched together from many different sources; a folklore so widespread that it is cliched. And it's immensely easier to repeat cliches about the irrelevance of character to arguments than to stop and say: But what is the underlying account?
