The skew Brownian motion is a strong Markov process which behaves like a Brownian motion until hitting zero and exhibits an asymmetry at zero. We address the following question: what is a natural counterpart of the skew Brownian motion in the situation that the noise is a stable Lévy process with finite mean and infinite variance. We define a skew stable Lévy process $X$ as the limit of a sequence of stable Lévy processes which are perturbed at zero. We point out a formula for the resolvent of $X$ and show that $X$ is a solution to a stochastic differential equation with a local time. Also, we provide a representation of $X$ in terms of Itô`s excursion theory.