5 papers indexed
Vincent Siu, Jingxuan He, Kyle Montgomery, Zhun Wang, Chenguang Wang, Dawn Song
Agent security is widely treated as a question about action content. Defenses ask whether an instruction looks malicious. Benchmarks ask whether an agent performs a harmful sounding action. \textbf{We argue that agent security is fundamentally a contextual problem, and that the c…
Niels Mündler-Sasahara, Hristo Venev, Dawn Song, Martin Vechev, Jingxuan He
Languages with rich static semantics, such as Rust, provide stronger guarantees for AI-generated code, but their strictness makes generation more difficult. Off-the-shelf compilers can provide useful feedback post-generation, but does not guide intermediate generation steps, such…
Tianyuan Zhang, Zonglei Jing, Jiangfan Liu, Ligong Zhang, Ke Ma, Chengzhi Sun, et al.
Vision-language agents (VLAs) are increasingly used to interpret complex driving scenes and support safety-critical reasoning. This report presents the CVPR 2026@AdvML Workshop Challenge on adversarial multimodal attacks against autonomous-driving VLAs. Built on DriveLM-style mul…
Yuqi Chen, Vincent Siu, Yang Liu, Dawn Song, Chenguang Wang
Tool-augmented large language models extend their capabilities beyond parametric knowledge through external tools, but tend to invoke them unnecessarily. We investigate whether tool-use decisions have any stable internal representation that can be extracted and manipulated, a que…
Peizhi Niu, Wenjie Qu, Shangding Gu, Tianneng Shi, Yuankai Li, Ahmad Tawaha, et al.
Claw-like AI agents (e.g., OpenClaw) are always-on processes with persistent access to credentials, files, tools, and external services. They take on system-level responsibilities -- installing packages, maintaining state, scheduling subtasks, and mediating I/O -- making security…