Metraipolis › archive › dc_ffc04d5c4fc9
Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery — arXiv:2609.09647
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.09647
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-16T07:24:28.070Z
- SHA-256 of the archived text
18dcb0cd6428bcb0ae91ff073e40c59db60902fcfb90066d8cd3390ba6c92fe4- Stable URL of this page
https://metraipolis.com/documento/dc_ffc04d5c4fc9
How to cite
Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery — arXiv:2609.09647 — arXiv (en), https://arxiv.org/abs/2609.09647 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_ffc04d5c4fc9 (Metraipolis archive dc_ffc04d5c4fc9, sha256 18dcb0cd6428bcb0ae91ff073e40c59db60902fcfb90066d8cd3390ba6c92fe4)
BibTeX
@misc{dcffc04d5c4fc9,
title = {Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery — arXiv:2609.09647},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.09647},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_ffc04d5c4fc9; sha256 18dcb0cd6428bcb0ae91ff073e40c59db60902fcfb90066d8cd3390ba6c92fe4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.09647. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
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. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-10 in cs.AI). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2609.09647 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-16T07:00:16.201Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-09-10?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).