Sometimes I see projects with massive code problems that I feel I should rewrite in Common Lisp for sanity's sake.

But then I come across the problem of them either using some C cryptographic library or worse, having homebrewed something.

I am not a cryptographer nor any sort of relevant expert, so even discounting for the side-channel related things difficult to address in CL (e.g. timing attacks), I still don't feel confident at all even attempting to reimplement that.