Also interesting. Of course, that's also what it would look like if an intelligent person wanted to make a specific threat: Create a burner account, make the threat, be specific, don't make errors, delete the account. They probably also conducted their activity through a random open proxy server so it is untraceable. It is also what it would look like if the victim was on an alternate twitter account that they use for safety and didn't want to reveal that account for privacy reasons, so when they saw this tweet they logged out and took the screen cap.
There are plenty of opportunities for a figure such as her to screencap perfectly legitimate threats without needing to make one up. It would be somewhat more surprising if this were false than true, but at this time I'm holding judgement on either possibility until we find out more information.