I'd say, along with some quick interview of the student, asking what game he wanted/intended to do, eventually why and how he did what he did you'd have a solid overview of the student's potential.
Depending on the people and their current programming knowledge half an hour can be enough, some might need more time.
It's a bit awkward to me though to "push the tool" before solid paper planning. But I don't know how your course is structured, and possibly you can get more infos observing a person focused on its work than in any paper design.
I don't think it's outrageous to use C2 for such a test though. You just have to take it as a tool to demonstrate potential than a full disclosure of one people's skill (especially in a so short discovery period).