Not exhaustive by any stretch but rooted in observations here over the past few years...
1. Don't ever lie to your student. Lying is not pedagogy or mentorship. When the student finds out about the lie, your relationship with that student is forever poisoned. They may not tell you they found out, but they will from then on regard you with suspicion.
2. Be direct and crystal clear at all times about all things money related. KNOW the university regulations around graduate funding and apply them openly and honestly.
If your student finds out they are doing work for you over and above their contractual funding, you are exploiting them. It doesn't matter how many times you tell yourself or them that you're giving them opportunity.
Be proactive in explaining that any extra work you might want them to do is over and above what their funding covers. Explain very clearly that if they turn down this work, it will have no impact on your supervisory relationship and you will not take petty punitive measures against them.
Pay them according to the rules. If your department is under-paying its gradstudents contrary to regulations, have the moxy to challenge your department.
3. Do not disparage your departmental colleagues because you are trying to attract a particular graduate student who has expressed interest in working with said colleagues. "Career suicide" is not a term to be used on 22 year old gradstudent who has just described their research and life aspirations to you. Doing such makes you look far worse in the student's eyes than the other prof (or chair with a CV several orders of magnitude greater than yours) you've just stabbed in the back.
4. If your student has spent an unknown number of hours completing project work for your comments, have the decency to read it, engage with it, and comment critically. The student can tell from your comments whether you read it. If your comments are shallow and uncritical, the student will think the same of you and lose faith in your capacity to mentor. They may not tell you this and you may be surprised to find they've sought an alternative mentor.
5. Bullying your graduate students is unacceptable. Some will cower and keep quiet because they are afraid to jeopardise their degree. Others will take action.
6. If you take on graduate students because they are feathers in your in tenure cap, do not tell them you need them to "hurry up and finish" so you can get your tenureship. This makes you appear like a self-interested asshole with no interest in them or their work other than what they can do for your career prospects. They will not be happy to continue working for you and some will leave your employ and your tenure prospects will be fucked.
7. Do not say things like "you're not my best student so don't expect as much support from me as I give my other student," or "I'm going to be a holy tyrant and fuck with your life," or "you cannot go away for the holidays because you must work for me."
8. Anything more than 2.5 years, full time study, is an unacceptable amount of time to complete a 2 year masters degree. If you are planning a 3+ year masters for them, tell them this when they start and be prepared to lose them. Anything less is dishonest - see the first point.
9. If you plan a 3 year degree, and the student knows this, fund them for three years or be very clear if you cannot and help them seek alternative sources.
10. Always keep in mind that you have a significant amount of control over your students' lives. Treat them with courtesy and respect at all times. Failure to do so will leave you wondering why they aren't interested in publishing with you post-degree: you treated them like shit for several years and they want nothing more to do with you. Ever.
11. Not every masters student is interested in a PhD. Or publishing. Or joining academia in any form. Understand this and do not let it impede your supervision. They may well live a much happier life outside of the ivory tower and indeed, their experience with you may well be one of the reasons why they opted out of academia...
12. If the student pays full-time tuition, they are paying to complete their degree, not be your multi-project indentured labour. If you tell them their degree is secondary to the other work they're doing for you, you are in the wrong. If you want a research assistant, hire a research assistant, not a student trying to get out of here in two years or less.
13. Remember what you told your student the last time you met. If need be, keep a journal of your meetings so you can refresh your memory. It annoys your students to no end when you don't remember telling them what you've told them (particularly if it involves copious amounts of work), and your present suggestions are the diametric opposite. Students might start to think that your doctorated mind contains mostly inconsistent bodily waste product.
14. Meet with your students. As often as is necessary. This will vary with students. Deal with it.
15. If you don't know what your students' research is about, you are not supervising them.