Securing Generative AI Agentic Workflows: Risks, Mitigation, and a Proposed Firewall Architecture

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Bahadur, Sunil Kumar Jang, Dhar, Gopala
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911016162426880
author Bahadur, Sunil Kumar Jang
Dhar, Gopala
author_facet Bahadur, Sunil Kumar Jang
Dhar, Gopala
contents Generative Artificial Intelligence (GenAI) presents significant advancements but also introduces novel security challenges, particularly within agentic workflows where AI agents operate autonomously. These risks escalate in multi-agent systems due to increased interaction complexity. This paper outlines critical security vulnerabilities inherent in GenAI agentic workflows, including data privacy breaches, model manipulation, and issues related to agent autonomy and system integration. It discusses key mitigation strategies such as data encryption, access control, prompt engineering, model monitoring, agent sandboxing, and security audits. Furthermore, it details a proposed "GenAI Security Firewall" architecture designed to provide comprehensive, adaptable, and efficient protection for these systems by integrating various security services and leveraging GenAI itself for enhanced defense. Addressing these security concerns is paramount for the responsible and safe deployment of this transformative technology.
format Preprint
id arxiv_https___arxiv_org_abs_2506_17266
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Securing Generative AI Agentic Workflows: Risks, Mitigation, and a Proposed Firewall Architecture
Bahadur, Sunil Kumar Jang
Dhar, Gopala
Cryptography and Security
Generative Artificial Intelligence (GenAI) presents significant advancements but also introduces novel security challenges, particularly within agentic workflows where AI agents operate autonomously. These risks escalate in multi-agent systems due to increased interaction complexity. This paper outlines critical security vulnerabilities inherent in GenAI agentic workflows, including data privacy breaches, model manipulation, and issues related to agent autonomy and system integration. It discusses key mitigation strategies such as data encryption, access control, prompt engineering, model monitoring, agent sandboxing, and security audits. Furthermore, it details a proposed "GenAI Security Firewall" architecture designed to provide comprehensive, adaptable, and efficient protection for these systems by integrating various security services and leveraging GenAI itself for enhanced defense. Addressing these security concerns is paramount for the responsible and safe deployment of this transformative technology.
title Securing Generative AI Agentic Workflows: Risks, Mitigation, and a Proposed Firewall Architecture
topic Cryptography and Security
url https://arxiv.org/abs/2506.17266