Quantum Key Distribution Secured Federated Learning for Channel Estimation and Radar Spectrum Sensing in 6G Networks

Collection of recent security-related arXiv papers spanning AI/ML robustness, agent security, privacy, and hardware trust. Highlights include: ClawWorm, a demonstrated self-propagating worm targeting LLM-agent ecosystems (OpenClaw) that hijacks configurations and propagates autonomously; a large-scale red-team study showing widespread vulnerability of LLM-based agents to indirect/hidden prompt injection attacks (attack rates up to ~8.5%); analysis of state-dependent multi-turn jailbreaks that cause rapid, reproducible safety collapse; BadLLM-TG, an LLM-driven trigger generator that reduces NLP