Metraipolis › archive › dc_362cb22686f3

Benchmarking Jailbreak Guardrails for Embodied Agents — arXiv:2610.06122

Source: arXiv
Original: https://arxiv.org/abs/2610.06122
License (as archived): CC0-1.0
Archived: 2026-10-06T05:25:25.857Z
SHA-256 of the archived text: e944fcbd20aa0fd9dd350f87ce5aebbb1510be3b9ae568c173d18637570167aa
Stable URL of this page: https://metraipolis.com/documento/dc_362cb22686f3

How to cite

Benchmarking Jailbreak Guardrails for Embodied Agents — arXiv:2610.06122 — arXiv, https://arxiv.org/abs/2610.06122 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_362cb22686f3 (Metraipolis archive dc_362cb22686f3, sha256 e944fcbd20aa0fd9dd350f87ce5aebbb1510be3b9ae568c173d18637570167aa)

BibTeX

@misc{dc362cb22686f3,
  title = {Benchmarking Jailbreak Guardrails for Embodied Agents — arXiv:2610.06122},
  howpublished = {arXiv},
  url = {https://arxiv.org/abs/2610.06122},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_362cb22686f3; sha256 e944fcbd20aa0fd9dd350f87ce5aebbb1510be3b9ae568c173d18637570167aa},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.06122. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Embodied agents powered by large language models and vision-language models are increasingly deployed in physical environments, but jailbreak attacks can induce these agents to perform physically harmful actions. A growing number of guardrail methods have been proposed to intercept dangerous behavior before it is executed, yet existing safety benchmarks evaluate the embodied models themselves, leaving it unclear how well these guardrails actually defend an embodied agent in practice. We present the first systematic evaluation of jailbreak guardrails for embodied agents. To compare guardrails under identical conditions, we build a pluggable evaluation framework that treats the embodied agent as a fixed backend and each guardrail as a module that can intervene at the perception, planning, or control stage. We subject six representative guardrails to template-based and automated jailbreak attacks as well as safe instructions, and assess them at the system level along three dimensions: defense effectiveness, measured by the bypass rate and the hazard success rate in the simulator; usability, measured by the false-positive rate and the task completion rate on safe instructions; and efficiency, measured by the latency overhead added at runtime. Experiments on guardrails that span different intervention stages, decision mechanisms, and input modalities reveal a clear trade-off among the three dimensions, and show that no single guardrail dominates in all settings. We further analyze how intervention stage, decision mechanism, and input modality shape safety outcomes, and we offer practical guidance for selecting and designing guardrails for embodied agents.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.06122 — 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 2026-10-06T05:24:05.649Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).