StepGuard introduces a step-level guard model that audits and checks tool actions before execution, addressing security risks in LLM-based agents. It is trained using StepGen, an automatic engine that generates safe and unsafe trajectories, and employs Balance-GRPO to dynamically balance learning between safe and unsafe actions. Experiments show StepGuard achieves high accuracy comparable to GPT-5.4 and significantly reduces attack success rates while minimally impacting utility.
By Zhijie Zheng, Yu Li, Chen Qian, Yuqian Fu, Yanwei Fu, Lu Sheng, Jing Shao, Dongrui Liu
arXiv:2606. 29887v1 Announce Type: new Abstract: In real-world applications, guardrails are often expected to identify unsafe user-model interactions according to application-specific safety policies, rather than relying on predefined risk taxonomies.
By Jiacheng Zhang, Haoyu He, Sen Zhang, Shen Wang, Xiaolei Xu, Yuhao Sun, Meng Shen, Feng Liu
VEPA is a framework that enables the verified evolution of formal guardrails for lifelong large language model agents. It represents guardrails as formal policies of symbolic rules over typed agent events, and during deployment uses LLM‑guided feedback to propose policy patches that fix new failures while preserving correct behavior. A deterministic verifier accepts only structurally valid, conflict‑free policies, and an evidence‑aware authority layer activates verified rules only when sufficient supporting evidence exists, ensuring that uncertain updates do not directly affect guardrail decisions.
By Duy Nguyen, Lesly Miculicich, Bhavana Dalvi Mishra, Shuo Li, Krishnamurthy Dj Dvijotham, Mohit Bansal, Tomas Pfister, Long T. Le
arXiv:2607. 19356v1 Announce Type: new Abstract: Tool-using LLM agents increasingly execute high-impact actions, making runtime safety monitoring essential.
By Elias Hossain, Md Mehedi Hasan Nipu, Tasfia Nuzhat Ornee, Rajib Rana, Niloofar Yousefi
arXiv:2609.00015v1 Announce Type: new
Abstract: AI agents powered by large language models are evolving from isolated assistants into heterogeneous systems in which multiple agents, planners, control...
By Dongsheng Chen, Xiangyu Zhao, Xin Yao, Xuetao Wei
Guard models are used to prevent LLM agents from performing forbidden actions, but this alone does not guarantee safety. The paper shows that many agent trajectories contain unfulfilled safety‑critical obligations—actions that should have been performed—highlighting a new source of risk. To address this, the authors introduce ObligationBench, a benchmark of 240 expert‑validated trajectories, and develop ObligationGuard, a model that improves obligation detection recall and exact‑match rates over existing models.
By Youwei Feng, Yitong Zhang, Yuetong Liu, Jia Li
arXiv:2607. 13594v1 Announce Type: new Abstract: LLM agents act on real-world environments through tool calls, and a single misjudged action can cause irreversible harm.
By Tianyu Chen, Chujia Hu, Wenjie Wang
arXiv:2610.16838v1 Announce Type: new
Abstract: As language-model agents write code, move money and change records on behalf of users, deciding whether an action is authorized has become a central sa...
By Huihao Jing, Shaojin Chen, Wenbin Hu, Yuxuan Liu, Haochen Shi, Zhongwei Xie, Haoran Li, Yangqiu Song
arXiv:2609.22724v1 Announce Type: cross
Abstract: Mobile agents powered by foundation models now automate complex, multi-step workflows on real devices, but their trajectories can violate app-specifi...
By Changyue Jiang, Jiayi Wang, Xin Wen, Jiarun Dai, Geng Hong, Xudong Pan
PROBE is an algorithm that trains language‑model auditors to probe deployed tool‑using agents for policy violations through realistic, multi‑turn, tool‑mediated dialogue. By framing auditing as a two‑player Markov game and using a composite reward that balances violation discovery, behavioral plausibility, conversation progress, and policy diversity, PROBE outperforms zero‑shot adversaries in uncovering broader and higher‑yield policy violations. The study compares GRPO, DPO, and GFlowNet optimizations, demonstrating that multi‑turn behavioral auditing is a distinct and necessary axis of agent evaluation, complementing existing capability and safety benchmarks.
By Kshitij Mishra, Abhijith Sharma, Nils Lukas, Salem Lahlou
arXiv:2607. 13081v1 Announce Type: cross Abstract: We present nsfaguard, a guardrail framework for securing agentic AI systems against operational threats, such as prompt injection, sensitive information extraction, malicious code requests, dangerous tool misuse, and resource exhaustion.
By SingGuard Team
RouteScan is a non‑intrusive auditing framework that detects harmful behavior in Mixture‑of‑Experts (MoE) large language models by analyzing expert‑routing telemetry captured from GPU execution. It uses the number of active GPU threads during the prefilling phase as a micro‑architectural fingerprint to isolate cross‑domain risk indicators and precisely identify malicious prompts. Evaluations on four open‑source MoE LLMs show strong generalization with AUROC > 0.91 on unseen harmful domains, while privacy tests indicate that full prompts cannot be reliably recovered from aggregated telemetry.
By Bo Lv, Zhiheng Xu, KeDong Xiu, Ruyi Ding, Tianhang Zheng, Zhibo Wang, Kui Ren