If a_min > a_max, the results can get a bit strange:

>>> np.clip(0, 1, 0)
1
>>> np.clip(1, 1, 0)
0

Now this can arguably be blamed on faulty inputs, and a precise reading of the docs (it returns "An array with the elements of a, but where values < a_min are replaced with a_min, and those > a_max with a_max.") explains that behavior (once we realize that the two tests have to be made one after the other), but throwing a ValueError (possibly silenceable/activatable by a keyword switch) could be helpful.