‘ skewness is a measure of the asymmetry of the probability distribution of a real-valued random variable about its mean’

Categorical variables are hence handled differently when used as features as compared to numeric variables which is why we do one-hot encoding or dummy coding.