This may be a nitpick of your comment, but I think very little of claims that a large program is “proven to be correct with mathematical precision”. We all know there are bugs in every large program, and if we aren’t smart enough to write the code without mistakes, then we aren’t smart enough to evaluate a proof of its correctness without mistakes.

Their Github page even has a section on reporting security vulnerabilities, suggesting that they themselves are not so certain of its correctness as you make it sound.