Writing a new distribution (2)

To elaborate, I express my codes in terms of Numpy library. This is my target function: y = x^-2/3., and I want to draw samples from it, so we have:

np.random.power(a = 1/3)

note that in Numpy, the power law function is: y = ax^(a-1) (we can ignore the coefficient a!)
The above code implements what I want except that the default range for x is [0,1]. I want to extend it to [0,100].

x in my problem is a physical quantity which can vary between 0 ,100 and I have to consider this condition