arXiv AI By Md Jafrin Hossain, Mohammad Arif Hossain, Weiqi Liu, Nirwan Ansari

The Containment Gap: How Deployed Agentic AI Frameworks Fail Public-Facing Safety Requirements

Read the original on arXiv AI →

arXiv:2606. 12797v1 Announce Type: new Abstract: Agentic large language model systems that autonomously invoke tools, maintain persistent memory, and execute multi-step plans are increasingly deployed in public-facing domains, including government services, healthcare triage, and financial advising.

Summary generated by The Flow from the publisher's feed. The full article lives at arXiv AI.

arXiv AI
Jun 18

SafeClawBench: Separating Semantic, Audit-Evidence, and Sandbox Harm in Tool-Using LLM Agents

arXiv:2606. 18356v1 Announce Type: cross Abstract: Tool-using language-model agents introduce security failures that go beyond unsafe text: they can disclose protected objects, write persistent memory, send messages, modify databases, or trigger harmful code and tool effects.

By Yuchuan Tian, Mengyu Zheng, Haocheng Mei, Ye Yuan, Chao Xu, Xinghao Chen, Hanting Chen, Yu Wang