ENH: Skew symmetric/matrix cross product function

Proposed new feature or change:

Dynamicists frequently use the matrix representation of the cross product (https://en.m.wikipedia.org/wiki/Cross_product#Conversion_to_matrix_multiplication), and the lack of a skew symmetric operator in NumPy is a frequent source of annoyance.

I propose either adding a skew function to linalg that returns the skew symmetric matrix of a vector np.cross(a, np.identity(a.shape[0]) * -1) (from https://stackoverflow.com/questions/66707295/numpy-cross-product-matrix-function), or more radically, making the second argument b of np.cross optional and have np.cross(a) return the matrix representation of [a x]. For the latter, would need to consider the interaction with other kwargs and different input shapes.

Start by reviewing the proposed linalg skew function and the alternative np.cross(a) API, including their interaction with existing kwargs and different input shapes. The issue mentions no implementation files or tests; done would require a settled API direction before implementation and validation can be scoped.