Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery

A new arXiv paper (2609.09647v1) presents a black-box red teaming framework that found 56.25% average governance risk, 65% privacy risk in multi-agent configurations, and agent behavior vulnerabilities reaching 85% across two agent architectures (CrewAI and AutoGen) and four base models. The framework combines a seven-domain taxonomy, the fully automated SAGE-RT red teaming method generating 120 adversarial scenarios per domain, and human-validated evaluation using LLM judges. The authors report the approach identifies critical architectural vulnerabilities without privileged access, offering a scalable path toward safer agent deployments.

arXiv:2609.09647v1 Announce Type: new Abstract: Agentic systems are rapidly moving to production, where they read untrusted inputs, call tools with real permissions, and act autonomously, expanding the security surface beyond chat-only models. Yet standard evaluations remain single-turn and fail to capture multi-step agent vulnerabilities. We present a systematic black-box framework for risk-aware agent evaluation requiring only basic system descriptions. Our approach introduces: 1 a seven-domain taxonomy mapping observable behaviors to risk categories, 2 fully automated SAGE-RT red teaming producing 120 adversarial scenarios per domain, and 3 human-validated evaluation using LLM judges. Empirical validation across two agent architectures CrewAI and AutoGen with four base models reveals alarming patterns: 56.25\% average governance risk, 65\% privacy risk in multi-agent configurations, and agent behavior vulnerabilities reaching 85\%. Our black-box approach effectively identifies critical architectural vulnerabilities without privileged access, providing a scalable path toward safer agent deployments.