Background. In modern digital environments, cryptographic mechanisms are essential for ensuring data confidentiality, integrity, and authenticity. One of the most widely used public-key cryptosystems is RSA, which underpins digital signatures, secure communication protocols (TLS/HTTPS), electronic document management systems, and financial transactions. Despite its robust mathematical foundation rooted in the challenges associated with factoring large integers, the practical security of RSA largely depends on accurate implementation and proper usage. Vulnerabilities typically arise from improper parameter selection, flawed protocol design, or insecure key management practices. Objective. The study aims to identify the conditions under which cryptanalytic attacks on RSA become feasible and to develop practical recommendations for improving system security. From an applied perspective, the research focuses on three types of attacks – keyless reading, digital signature attacks in notary-based schemes, and chosen-ciphertext attacks – in order to evaluate their feasibility, limitations, and impact on information system security. Methods. A sequential research methodology was applied. First, the RSA algorithm was analysed, including key generation, encryption, decryption, and digital signature mechanisms. Next, three attack models were examined: the keyless reading cycle attack, the notary-based digital signature attack, and the chosen-ciphertext attack. These methods were implemented in Python to experimentally evaluate their effectiveness under different parameter configurations. Finally, a comparative analysis was conducted considering both theoretical assumptions and practical implementation constraints. Results. The results indicate that the keyless reading attack is effective only when RSA parameters are weak and short cycles exist; with key lengths of 1024 bits or more, it becomes practically infeasible. In contrast, digital signature attacks demonstrate significant effectiveness in the presence of protocol-level flaws, such as blind signing, lack of input validation, or key reuse. The chosen-ciphertext attack demonstrates moderate effectiveness and depends on violations of secure key usage practices. Overall, the findings confirm that real-world vulnerabilities stem primarily from implementation weaknesses rather than inherent mathematical flaws of RSA. Conclusions. The study confirms that RSA remains secure when modern cryptographic best practices are properly applied. Most vulnerabilities arise from improper implementation, insecure configurations, or protocol design flaws rather than theoretical weaknesses of the RSA algorithm itself. To enhance security, it is recommended to use key lengths of at least 2048 bits, apply cryptographic hashing before signing, avoid key reuse across different cryptographic functions, and incorporate random parameters such as salts or timestamps. These measures significantly strengthen the robustness of RSA-based systems against cryptanalytic attacks.