Goodness-of-fit testing is a statistical methodology used to assess whether a dataset conforms to a hypothesized theoretical distribution or model. This process is critical across scientific and industrial domains - from validating normality assumptions in medical research to evaluating financial risk models - as it ensures the reliability of subsequent analyses and conclusions. However, the effectiveness of such testing depends on the choice of criteria, which vary in their sensitivity to sample size, significance level, and alternative hypotheses.To address this challenge, we propose a flexible, open-source framework designed for systematic comparison of goodness-of-fit criteria. The framework enables researchers to configure experiments by adjusting parameters such as sample size, significance level, and alternative distributions, while offering modular integration under any criterion. Its architecture decouples data generation, criterion application, and result analysis, ensuring reproducibility and scalability.Using this framework, we provide a comprehensive comparison of normality criteria, evaluating their performance under varying sample sizes and alternative distributions. The results demonstrate significant differences in criterion power and robustness, underscoring the importance of context-aware methodology selection. This work advances statistical practice and supports the development of new criteria.