20.500.12556/RUL-172597 Uporaba umetne inteligence za izvajanje varnostnih pregledov Using Artificial Intelligence to Conduct Security Penetration Tests Diplomska naloga se osredotoča na povezovanje velikih jezikovnih modelov s procesi etičnega hekanja na osnovi protokola konteksta modela (MCP), katerega uporaba se hitro širi po svetu. Tekom diplome so predstavljeni pojmi, kot so kibernetska varnost, etično hekanje in njegov pomen za varnost, metodologija etičnega hekanja ter pravni vidiki. Nato sledi opis avtomatizacije testov in umetne inteligence. Pred praktičnim delom je razložena še uporaba orodij z jezikovnimi modeli, kjer se prvič spoznamo z delovanjem MCP. V praktičnem delu smo pripravili integracijo velikega jezikovnega modela s strežnikom MCP za podporo fazam vdornega testiranja, ter analizirali prednosti in slabosti tovrstnega pristopa. Podani so tudi primeri uporabe - enostavni in težji. Na koncu so obravnavani pravni in etični vidiki uporabe umetne inteligence v kibernetski varnosti. The thesis focuses on integrating large language models with ethical hacking processes based on the Model Context Protocol (MCP), whose use is rapidly spreading worldwide. Throughout the thesis, concepts such as cybersecurity, ethical hacking and its importance for security, the methodology of ethical hacking, and legal aspects are presented. This is followed by a description of test automation and artificial intelligence. Prior to the practical part, the use of tools with language models is also explained, where the functioning of MCP is introduced for the first time. In the practical part, we prepared the integration of a large language model with an MCP server to support the phases of penetration testing and analyzed the advantages and disadvantages of this approach. Use cases are also presented—ranging from simple to more complex. Finally, the legal and ethical aspects of using artificial intelligence in cybersecurity are discussed. kibernetska varnost etično hekanje umetna inteligenca Claude Kali Linux MCP cybersecurity ethical hacking artificial intelligence Claude Kali Linux MCP true false false Slovenski jezik Angleški jezik Diplomsko delo/naloga 2025-09-09 13:25:01 2025-09-09 13:25:06 2025-10-16 03:55:40 0000-00-00 00:00:00 2025 0 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 63140 251689475 Kerlatec_Aljaz_-_Uporaba_umetne_inteligence_za_izvajanje_varnostnih_pregledov.pdf Kerlatec_Aljaz_-_Uporaba_umetne_inteligence_za_izvajanje_varnostnih_pregledov.pdf 1 C01E876FDA922C47878E6194FC41DC3D cdf3bfb70dd1d96586c72bbe955440e6add0c9204a79621178af5c139b3c7f4e 7bdd7c95-8d6f-11f0-9328-0050569b8976 https://repozitorij.uni-lj.si/Dokument.php?lang=slv&id=216507 Fakulteta za elektrotehniko 0 0 0