A formal model of agentic AI vulnerabilities in containerised cloud environments
Cyber Security: A Peer-Reviewed Journal · 02.08.2026 · 10 (1) · 45
DOI: 10.69554/jhix5712 ↗Kurz gefasst
Fachartikel in der Zeitschrift Cyber Security: A Peer-Reviewed Journal, online erschienen am 02.08.2026. Verfasst von Advait Patel und 3 weiteren Autor:innen.
Abstract (Originalsprache)
In today’s evolving digital landscape, a new kind of AI has emerged: agentic artificial intelligence (AI) systems. These are not just tools but intelligent entities that act on their own, make decisions, and carry out tasks proactively. Once placed inside containerised cloud environments such as Kubernetes, however, these powerful agents open the door to a whole new category of security concerns. They often interact on the fly with application programming interfaces, data channels, and other AI agents, sometimes all at once, and not always in predictable ways. Recognising the risks, this paper introduces a formal framework designed to capture and analyse these unique vulnerabilities. Going beyond theory, the study offers a full threat taxonomy focused on agent behaviours, uses temporal logic to trace how agents and environments interact over time, and introduces a practical measure called the Agentic Vulnerability Exposure Metric (AVEM). To put this model to the test, researchers ran simulated cyberattacks on a Kubernetes-based orchestration system that was managing large language model (LLM)-powered agents. The results were eye-opening: AVEM was able to expose security holes such as weak isolation between agents, overlooked permissions, and routes for privilege escalation that had gone undetected before. To test the practicality of the model, real-world attack scenarios are simulated on a Kubernetes-managed AI platform utilising LLM agents. The findings reveal that AVEM effectively uncovers weaknesses such as poorly defined isolation boundaries, unexpected capabilities, and opportunities for privilege escalation between agents. This article is also included in The Business & Management Collection which can be accessed at https://hstalks.com/business/.
Zitieren
APA
Patel, A., Gudur, V., Upadhyay, C., & Sudarsan, S. (2026). A formal model of agentic AI vulnerabilities in containerised cloud environments. Cyber Security: A Peer-Reviewed Journal, 10(1), 45. https://doi.org/10.69554/jhix5712
Vancouver
Patel A, Gudur V, Upadhyay C, Sudarsan S. A formal model of agentic AI vulnerabilities in containerised cloud environments. Cyber Security: A Peer-Reviewed Journal. 2026;10(1):45. doi:10.69554/jhix5712
BibTeX
@article{Patel2026,
author = {Patel, Advait and Gudur, Vaishnavi and Upadhyay, Charit and Sudarsan, Shalini},
title = {A formal model of agentic AI vulnerabilities in containerised cloud environments},
journal = {Cyber Security: A Peer-Reviewed Journal},
year = {2026},
volume = {10},
number = {1},
pages = {45},
doi = {10.69554/jhix5712}
}
EUVIMED