SKIMMR is currently in a research prototype stage and thus admittedly suited more for 'power users' ... Yet no knowledge of Python should be required to run the tool locally on your own corpus of PubMed abstracts (or any other texts you may posses) - the downloadable package is built so that spinning up a custom instance of SKIMMR requires only launching a couple of command line scripts if you have Python installed. A bit geeky, I agree, but possible with not so much technical knowledge. We're currently working with my group on a couple of more user friendly and comprehensive technologies based on SKIMMR that will require no technical knowledge at all, but there is no public version out yet.
Another (easier and possibly more immediate for you) way is to give me a list of PubMed IDs or actual texts you're interested in and I can set up an experimental instance of SKIMMR for you on my server. My resources for that are limited as I can't obviously do this for infinite number of people on my relatively small research server, and it may take a while (like about 1 week) to squeeze it into my schedule, but I'd be happy to do that in general.