In the new context of transitional, mobilised and globalised labour markets, an urgent need has emerged for meaningful assessment tools, methods and techniques to measure and recognize the workers' skills. This paper aims to present a novel approach in skill assessment developed under NADINE H2020 project, the indirect skill assessment, as an alternative or supplement to the traditional selection methods. AI technology is utilised to facilitate the indirect skill assessment via two agnostic content serious games (Tetris and 2048), based on the evaluation of an individual's performance in playing a game outside a situational judgment test (SJT) context. Novel datasets have been developed, comprised of game sessions and the corresponding skill assessments of the players through validated psychometric questionnaires, which were the basis for the algorithm training that would provide the estimation of a player's skillset. The trained neural models for both games proved to have strong skill assessment capabilities, indicating that there is indeed a correlation between a person's action sequence and his/her different skills.