It's a difficult thing to test, so it's doubtful there is any meaningful way to prep. Bring some printed pictures of work you've done and be ready to explain how you started and what steps you took to get there.
I would not expect a company to allow any medium to touch a computer - no jump drive, no CD, nada. Even if they believe you are a good person, which is why they invited you in, there's only risk to them and no benefit to possible exposure to malware that could have been loaded without your knowledge. Should they have anti-virus in place and all that? Sure. But you can't catch what you don't contact.
***
On the general topic of training and skills evaluations:
It would be nice if PTC supplied an end-point list and end-point tracking. An end-point is a place where the software has enough information to perform a task. Since it is the user who is supplying the required information, if you know the end-points, you know all the tasks you can get the software to perform. While the combinations of tasks is infinite, at least as far as humans are concerned, the number of tasks is fixed.
If you had a list of end-points, you can verify whether or not you know enough about the information requirements to get to them and have the software function. If PTC tracked end-point accomplishment, then they could dynamically report on what you successfully know vs what you either don't know or haven't tried. I suppose the software could differentiate between failed attempts and not visiting an endpoint.
Imagine you are asked if you know how to use a scientific calculator.
Each function button is an end-point and it's pretty easy to take each button one at a time and determine if you know about what it does and what you have to put into the calculator before pressing it. With tracking, one might find that SIN() gets used a lot, but COSH(), not ever. Then you could more easily determine what you were good at and whether there were things you should know. If it also tracked the number of times the result was an error - ASIN(5), for example, it could further tell when unacceptable conditions were present at the time the function was called.
This won't prove that you're a mathematician or engineer, but if there are functions that mathematicians and engineers use a lot and you match or exceed those uses then it's possible you know something. The inverse conclusion is also true, but it gives concrete support to proceed.
I've done military, nuclear, and public/consumer stuff and it's been a mixed bag. As a contractor or candidate they can be even tighter.
