Study on Artificial Intelligence Agents Applied to Cybersecurity and Penetration Testing in Systems
Científica · 20.06.2026 · 30 (1) · 1-29
DOI: 10.46842/ipn.cien.v30n1a09 ↗Kurz gefasst
Fachartikel in der Zeitschrift Científica, online erschienen am 20.06.2026. Verfasst von Gerardo Cruz Espinosa und 1 weiteren Autor:innen.
Abstract (Originalsprache)
This investigation analyzes the technical functioning and implementation of artificial intelligence agents oriented toward specific domains, particularly in the field of cybersecurity from an offensive perspective. These agents enable a certain degree of automation in the execution of penetration tests on computer systems due to this type of proofs are made through tools operated by cybersecurity professionals. The objective of this study is to examine their architecture and operational capabilities. The conducted tests demonstrate their ability to reason and perform actions through the use of large language models. The results suggest that these emerging technologies can support cybersecurity professionals in various operation tasks, it is important to note that these tasks are already being done manually. The methodology employed is structured and developed in stages, integrating information acquisition, contextual analysis, action planning, controlled execution, and continuous feedback.
Zitieren
APA
Espinosa, G. C., & Trejo, M. V. (2026). Study on Artificial Intelligence Agents Applied to Cybersecurity and Penetration Testing in Systems. Científica, 30(1), 1-29. https://doi.org/10.46842/ipn.cien.v30n1a09
Vancouver
Espinosa GC, Trejo MV. Study on Artificial Intelligence Agents Applied to Cybersecurity and Penetration Testing in Systems. Científica. 2026;30(1):1-29. doi:10.46842/ipn.cien.v30n1a09
BibTeX
@article{Espinosa2026,
author = {Espinosa, Gerardo Cruz and Trejo, Mariana Velasco},
title = {Study on Artificial Intelligence Agents Applied to Cybersecurity and Penetration Testing in Systems},
journal = {Científica},
year = {2026},
volume = {30},
number = {1},
pages = {1--29},
doi = {10.46842/ipn.cien.v30n1a09}
}
EUVIMED