The Blind Spot of Agent Safety: How Benign User Instructions Expose Critical Vulnerabilities in Computer-Use Agents Paper • 2604.10577 • Published Apr 12 • 27
Structured Distillation of Web Agent Capabilities Enables Generalization Paper • 2604.07776 • Published Apr 9 • 21
CoAct-1: Computer-using Agents with Coding as Actions Paper • 2508.03923 • Published Aug 5, 2025 • 13
Feedback Friction: LLMs Struggle to Fully Incorporate External Feedback Paper • 2506.11930 • Published Jun 13, 2025 • 53
Efficient Reinforcement Finetuning via Adaptive Curriculum Learning Paper • 2504.05520 • Published Apr 7, 2025 • 11
