Resilient multi-agent output containment under undisclosed leaders and actuator attacks

arXiv·low signal

arXiv:2606.27257 (June 25) presents control-theoretic methods for keeping multi-agent systems coordinated when leader dynamics are hidden and actuators are under adversarial attack. It's from the classical MAS/control side rather than LLM agents, but speaks to robustness of physical multi-agent deployments (robotics, swarms). Adjacent signal for teams bridging LLM orchestration with real-world actuated agents.