To answer the question: Yes, there is a way. Use a bright but unsaturated star. That might already be good enough, or you might want to do any number of corrections to this PSF model.
A better way is to figure out the parameters to use with TinyTim, and generate an approximately right PSF from those. Most parameters you know, or they are easy to look up, such as camera, chip, filter. For the pixel position, you need to look up the original exposures that went in to your (presumably) drizzled image; the object will be more or less in the same place in all of them, so you can use the average position.
With these, you can generate an overresolved PSF image. If your (presumably) drizzled image is north-aligned, you will need to rotate the PSF accordingly; ideally using drizzling, but a simple image rotation might work well enough if your PSF resolution is much higher than the HST resolution. The final result is a PSF that might not be useful for precision photometry, but should be good enough for modelling a galaxy. In any case, compare the PSF to the image of a star to make you it is roughly right (east-of-north angles in HST images have a way of being wrong).