What you need to do is to find the general form of a matrix that satisfies A=-A^T. That is, make each of the 9 entries a variable, and use that equation to make as many substitutions as you can to simplify A.
Once you've done that, you should have found the minimum number of variables you can use to describe a matrix satisfying such an equation. That number is the dimension of the vector space W.
Then, you need to decompose the matrix A as a linear combination of linearly independent 3x3 matrices. This shouldn't be too hard, as all you need to do is to look at entries with the same variable, and make it so A is equal to a sum of 3x3 matrices, with a variable as a coefficient in front of each matrix. Then each of those 3x3 matrices are your basis vectors.
I hope that makes sense.
For example, the matrix {{a,2b},{b,-a}} (first row is a and b, second row is b and a) has dimension 2, because the variables a and b can freely vary. I can decompose the matrix as a{{1,0},{0,-1}} + b{{0,2},{1,0}}. And so my basis vectors are the matrices {{1,0},{0,-1}} and {{0,2},{1,0}}.
Edit: corrected something and used a slightly better example.
.
