Hi Blanco,
First... are you certain it is a "shooting star?" There are many things that *look like a "shooting star"* but actually are not.
Tumbling satellites (and their trails) can look like them. So can cosmic rays.
Ironically the entire idea of image integration is to generally reject these things so they do not show up in your image.
If the shooting star is actually very bright- there is a fancy way to do it- but it is likely that without some processing practice you will think it is a bit tedious in effort.
What you would do is integrate your images in the normal way and have this thing rejected and not appear. Then you subtract this integrated image from your single image that contains the "shooting star" (I will continue to be dubious...lol). This may require the application of Linear Fit before you do the subtraction. You will then end up with an image that has just the "shooting star."
Finally you can blend/combine this image with your integrated result using a number of ways including using "maximum" or blend with "lighten."
This is the method I used to create the image below showing all the satellite trails running through my data (that were subsequently rejected).
You can also be a little more brute force with it an generate a mask (with a line of some kind)... but the above technique gets closer to keeping the "real" relationship of the values.