(1)
Robust Multi-Agent Collaboration via Trace-Based Safety Monitoring and Reasoning Path Alignment. JDIAS 2026, 1 (1).