Metraipolis › archive › dc_19913fc3a423

ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents — arXiv:2608.11878

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.11878
License (as archived)
CC0-1.0
Archived
2026-10-19T05:31:01.048Z
SHA-256 of the archived text
ff3cc673f84cda712c21e384c00aebebc46454293738dfaafa6414145b9cd82d
Stable URL of this page
https://metraipolis.com/documento/dc_19913fc3a423

How to cite

ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents — arXiv:2608.11878 — arXiv (en), https://arxiv.org/abs/2608.11878 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_19913fc3a423 (Metraipolis archive dc_19913fc3a423, sha256 ff3cc673f84cda712c21e384c00aebebc46454293738dfaafa6414145b9cd82d)
BibTeX
@misc{dc19913fc3a423,
  title = {ToolHazard: Scaling Adversarial Environments for Security Evaluation and Alignment of LLM-based Agents — arXiv:2608.11878},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.11878},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_19913fc3a423; sha256 ff3cc673f84cda712c21e384c00aebebc46454293738dfaafa6414145b9cd82d},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.11878. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large language model (LLM) agents integrated with external tools are vulnerable to indirect prompt injections embedded in environmental states. However, existing studies largely rely on manually implemented or reused environments, stochastic LLM-based tool simulation, and predefined injection locations, limiting scalable security research across broader domains. To bridge this gap, we propose **ToolHazard**, a scalable adversarial environment synthesis framework that reduces human engineering and supports expansion with additional seed domains and compute. Through an Environment Simulator, an Attacker Agent, and a User Simulator, ToolHazard synthesizes executable stateful environments, discovers viable injection points and generates environment-specific payloads, and constructs state-grounded long-horizon tasks. Based on ToolHazard, we build **ToolHazard-Bench** for stress-testing agents under complex workflows and diverse environmental attacks. Experiments reveal substantial agent vulnerabilities and show that injection timing and placement affect attack effectiveness. Moreover, ToolHazard-generated alignment data improves security on both ToolHazard-Bench and AgentDojo while preserving benign task utility.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-13 in cs.CL).
Categorie arXiv: Cryptography and Security (cs.CR) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.11878 — 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 2026-10-19T05:15:39.258Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-13?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).