The replies in this thread have already said the biggest limitation of C2 - the users' design capabilities, which applies to pretty much every technology that has to do with creating something.
(I got carried away below... the paragraph above is really all I wanted to say. You can stop reading from here XD)
I'll move onto a lesser limitation instead: C2 uses technology that is largely experimental with unknown stability, namely HTML5 and "fiddly" javascript.
It's not a problem for everybody, especially if you stick to well-organised and systematic design practices (which should be technology neutral), which is what my first paragraph was talking about. However some people are not comfortable, or literally run into problem, with C2 game building where the root of the problem lies in javascript or HTML5 itself. The platform that C2 targets, which is HTML5, can be ever so flexible, but at the risk of not having every single use case covered already and end up having you to fiddle with the core aspects / backend of your game.
Some people may have particular ways of designing their games, be it prior knowledge from other engines or programming language, and suddenly they can't use the exact same practices C2 is designed for. That's more of a preference problem rather than limitation. People who came from programming background may also find the graphical program logic interface (aka the event editor) a bit weird and prefer typing them out instead of visually arranging the logic.