Parser could allow for different behaviour of tokens if they are not needed in syntax at their position

[s4ke]

For example < and > are only needed for capturing groups. Escaping them outside of capturing gruops is a bit weird, but can be kept for simplicity of the parser.