"Varied" or "widely distributed" seem like better terms. To my mind, "skew" is what we have in the original problem (with a zero-bound) - it only describes asymmetric distributions.