As a maintainer of GNU Parallel I often get bugs reports from people using my software in ways I had never imagined. The simple reply to that could be:
Do not do that. My tool was never intended to do that.
Or if you are in a really helpful mood:
Here is a work-around that does not trigger the bug.
You can especially be tempted to give such a reply if you feel this bug is not going to affect many people or if it is a lot of work. But this kind of thinking is limiting.
Programs should aspire to be without error and work even if they get input that is out of the ordinary. The reason for this is that if users can rely on programs not breaking no matter the input, then users can computer generate input, and the program can then be used in novel ways that you never foresaw.
Therefore a much better answer is:
That looks like a bug. Please file it in the bug tracking system, so we can track how many people are affected by this.
And if you feel it is a low priority bug:
I do not have the time to fix the bug, but you can help by analyzing the bug so we know exactly what is wrong. A patch is also welcome – just be aware that a patch that fixes this bug, may trigger other bug or may be written so poorly that it will not be applied.
This also has a social benefit. The bug finder has taken time to make an example that is MCVE, and you acknowledge him for the work he has done. By putting it in the tracker, you are also saying: “We are not going to simply forget this bug.” And by giving him the option of sending a patch, you may even get some of the work done – if he (or another affected user) feels it is that important.