I also wanted to make pages for &, { and ( but those symbols were so multipurpose that I couldn't come up with a concise way to describe them.

Do note that some of these are only reachable by enclosing the symbol in quotes like tldr "<"

Not sure how I'd fix that

In #12091, :.md was renamed to colon.md. Renaming the files is the only possible solution, as there is no way to force Windows to allow these characters in filenames.

tldr-pages/tldr-lint#352
As a reminder for the future.

Would it be reasonable to have hardcoded exceptions in tldr, to redirect requests with forbidden windows filesystem characters to their complementing characters.
As in if I type tldr :, the program searches for tldr colon or tldr ">" gets converted to tldr greater-than
I ask this because I like consistency and being consistent with other symbol based queries like tldr ! would be nice.

Would it be reasonable to have hardcoded exceptions in tldr, to redirect requests with forbidden windows filesystem characters to their complementing characters.

As in if I type tldr :, the program searches for tldr colon or tldr ">" gets converted to tldr greater-than

I ask this because I like consistency and being consistent with other symbol based queries like tldr ! would be nice.

This is something that our clients should implement, which we need to write in the client-specification. Maybe open a separate discussion for this?

After you've removed those unnecessary parts of the urls, add something to tldr bot that automatically detects when specific URLs contains something extra. gnu.org for example always adds www to its urls when you open the link and manned always adds /man/ when navigating there manually.