Yes, it's not the intended algorithm but I think it would be an interesting valid rule for an algorithm in a different game as ABORT does not have separated consonants and falls on the most frequent branch. It might still raise some objections depending on each player's view on the game rules.
Maybe the consistency rule isn't working quite as I intended it. When writing that I was thinking in terms of operations which are used repeatedly (as in the first game). But here each operation applies only once within each clue and the 50% rule pretty much rules out any IF A THEN B ELSE C operation if applied only once per clue (since if B occurs more than half the time, C won't, and vice versa). That's probably too restrictive. What matters is that the hackers get to see enough examples of the rule that they can deduce it, so I wouldn't get hung up on that too much. It probably needs to be worded differently.
As to whether brute force attacks are appropriate, my inspiration for this game is the real-life problem of creating secure passwords as a hash of website names. Such a system needs to be resistant to a little guessing, since a hacker who has obtained some of your passwords may well try a few likely-looking combinations without knowing the entire system. IMO that is in the spirit of the game as it's part of the challenge to build in resistance to that. Plus, it would take enormous restraint not to make guesses knowing the password could only be one of four! It is, after all, a race between the hackers. The brute force applicable is limited by the fact that they only get one guess for every 3 states you destroy.
Unrelated: the 40 seconds restriction (and also the fact that you don't see the password letters so you need to recompute previous variables/letters) does impose serious constraints on the obscurity of a rule. Again, IMHO, it should be even simpler to decide these booleans if they are used more than once.
There is no need to recalculate anything you can keep track of in your head. For most people that would not be very much, but still, reusing a couple of booleans is fine, and cuts down on thinking time.
