A few thoughts:

 - On "no subtle put-downs", you're inviting moderators to overreach. Subtle things are hard to judge, and banning them tells moderators it is ok to dive into things reasonable people might disagree on. It'd be better to go the opposite way and refuse to act on subtle things. And the examples are things that probably should not be banned; sometimes in societies people get frustrated with each other. Expressing that should be okay. You're building this for humans.
 - On the other hand, the "no name-calling or personal attacks" seems workable and positive.
 - On "no bigotry", there's really no way to do this right. "Likely to offend" is not the right standard - people get offended about all sorts of things, and trying to control people's language will (rightly) offend them too.
 - On "No harassment", you should probably prefer a less expansive definition of harassment, focused on repeated unwanted contact after requests to stop. Other offenses should be discussed as separate things.