In this data article, a database is presented and described that can be used to solve multiskilled personnel assignment problems (MPAP) under uncertain demand. This database contains simulated datasets along with a real dataset taken from a Chilean retail store. Information about the store such as the number of departments and workers, the type of labor contract, the cost parameter values, and the average demand in all store departments, are presented in the real dataset. While information related to stochastic demand of the store departments was created with a Monte Carlo simulation, and is presented in the simulated dataset, consisting of 18 text files categorized by: (i) Type of sample (in-sample or out-of-sample). (ii) Type of truncation method (zero-truncated or percentile-truncated). (iii) Demand coefficient of variation (5, 10, 20, 30, 40, 50%). Academics and practitioners may utilize this dataset to benchmark the performance of diverse methods to optimize under uncertain demand and, therefore, obtain robust multiskilling levels to the same (or similar) MPAP. Additionally, it is provided an Excel workbook that generates up to 10,000 demand scenarios with different coefficients of variation.