Welcome to the community!
That’s a really tough question. It really depends on what you’re trying to achieve.
But I would say that if you want to go down the science route, the best option would be design a proper test array with a variety of plausible user inputs - the ‘.’ (and empty string! ‘’, oof, empty string cost one project a lot of money lol) could be a valid edge case, but it probably won’t really inform real life behavior.
Personally, I would tell you that the system prompt is probably considerably less important than you may think.
Are you generally trying to get a feel for how prompting works, or are you trying to proof your system against malicious input, or perhaps just trying to ensure you get a stable, structured output?
While the scientific approach can certainly be useful and I would absolutely not want to discourage it if you have the time and energy for it, I’d personally rather take the engineering approach: use tried methods to design and construct a prompt to get the outcomes I’m looking for.