benchmark · first seen 12 Oct 2026
AI agent safety evaluation
Also: agent safety evaluation, coding agent safety benchmarking
Testing whether autonomous AI agents violate intended boundaries such as file integrity or system permissions.
Mentions per week
PapersCodeDiscussion
Earliest sources
- agent-risk: Risk Assessment for Changes to AI AgentsPapers · 12 Oct 2026 · Yue Zhao, Zixiang Xu, Angela Huyue Zhang, S. Alex Yang, Xiyang Hu
- SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding AgentsPapers · 13 Oct 2026 · Do Xuan Trong, Duc Anh Nguyen, Do Xuan Long, Ngoc-Hai Nguyen, Tongyao Zhu, Duy D…
- V$^3$AST-Bench: Verifiable, Valid, and Versatile Safety Evaluation for Coding AgentsPapers · 26 Oct 2026 · Weitao Feng, Hengheng Xiong, Zihao Zeng, Jun Chen, Jiyang Guan, Xiaofeng Wang, J…
- SideTaskBench: Measuring Unauthorized Side-Task Delivery in Coding AgentsPapers · 29 Oct 2026 · Xiangyu Wen, Qiang Xu
Latest sources
- SideTaskBench: Measuring Unauthorized Side-Task Delivery in Coding AgentsPapers · 29 Oct 2026 · Xiangyu Wen, Qiang Xu
- V$^3$AST-Bench: Verifiable, Valid, and Versatile Safety Evaluation for Coding AgentsPapers · 26 Oct 2026 · Weitao Feng, Hengheng Xiong, Zihao Zeng, Jun Chen, Jiyang Guan, Xiaofeng Wang, J…
- SWE-MISBEHAVIOR: Benchmarking and Mitigating Boundary Violations in Software Coding AgentsPapers · 13 Oct 2026 · Do Xuan Trong, Duc Anh Nguyen, Do Xuan Long, Ngoc-Hai Nguyen, Tongyao Zhu, Duy D…
- agent-risk: Risk Assessment for Changes to AI AgentsPapers · 12 Oct 2026 · Yue Zhao, Zixiang Xu, Angela Huyue Zhang, S. Alex Yang, Xiyang Hu