Sorry, but I'm not clear on what part of "You may not make threats of violence or promote violence" specifies it must be "illegal" violence. Part of the President's job may be to make threats and conduct violent actions against other nations. That doesn't change the fact that it violates the terms of use.
That kind of talk has no place on Twitter based on Twitters own rules that say no threats are allowed.
I get the theory, but you're over-interpreting the specific words of one part of the policy.
The words are an expression of the ideas in people's heads, and are written with a particular context in mind. That context is fighting the normal kinds of platform abuse they see. The policy was definitely not meant to encompass the typical actions of governments and heads of state. Nobody was even thinking about that at the time they were written.
If they were to use that to ban Trump, it would be pretty obvious rules lawyering [1]. Which many people certainly feel is justified, and I understand why. But from Twitter's perspective, rules lawyering, no matter how much people feel it's justified, undermines trust. So if they want to get rid of Trump, they'd be better off writing a specific policy that Trump (and others) are clearly violating, not bending something else to fit.
That kind of talk has no place on Twitter based on Twitters own rules that say no threats are allowed.