Skewed distribution is one where one tail is longer than the other. A distribution could be positively or negatively skewed as shown below
Couple of outliers closer to the center of the distribution may not affect skew much but the farther they are away say 3–6 Standard deviation(SD), such data points can have a great impact as these data point lies outside the overall pattern of a distribution.
Best approach would be to plot the data and decide the cutoff for many SD should be considered for identifying an outlier.
Let me know if this works