The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents
Paper • 2604.10577 • Published • 22
Zhe Xu
zhexu3
AI & ML interests
None yet
Recent Activity
upvoted a paper about 15 hours ago
The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents upvoted a paper about 1 month ago
Heterogeneous Agent Collaborative Reinforcement Learning upvoted a paper about 1 month ago
Video-Based Reward Modeling for Computer-Use AgentsOrganizations
None yet