The paper presents experimental studies of combinatorial algorithms designed to calculate the composition and order of application of vulnerability parry tools, conducted on the basis of a typical electronic repository of the cultural institution information system. Experimental studies of data protection algorithms with the use of a neural network composition calculation unit and the order of application of vulnerability parry tools have been carried out. The proposed approach makes it possible to obtain a gain in time and efficiency (reliability) by using standard means of protecting software tools of large-scale electronic repositories, typical for large cultural institutions.