I have used/seen the convention elsewhere to use a $ prefix to denote a 'builtin' function or variable. It makes it easier when there is a possibility of some runtime defined variable to prevent naming conflicts.

i.e.

exp:="$pi * radius^2"

I can't think of a reason for $ to be a reserved character I would propose the parser recognizes it as a character and looks for variables and functions containing it.

Happy to do a PR if there isn't a counter argument