Give an example of a two-dimensional random vector whose components have 0 covariance but are not independent.

Many solutions are possible. A simple one is a discrete random vector X with PF

You can confirm that the covariance between the two components is 0 but that condition conditional PFs and fail to satisfy