Multi-Operator, Multi-UAV (MOMU) environments impose substantial cognitive demands on human teams supervising autonomous swarms under dynamic and safety-critical conditions, with similar challenges arising in other autonomous swarming domains. Many documented failures involve breakdowns in human situational awareness (SA) and team coordination, triggered by previously unrecognized conditions (unknown unknowns) that lead to hazardous situations. This motivates a shift from reactive practices, where new requirements are discovered only when failures occur, towards proactive, safety-driven requirements discovery. Unfortunately, existing requirement elicitation practices offer limited support for proactively identifying SA and team coordination failures under high workload and time pressures. We address this challenge through introducing a novel Multi-Operator, Multi-UAV eXperimental platform (MOMUx), which provides a high-fidelity simulation environment that serves as a catalyst for identifying MOMU interaction failures. MOMUx provides a low-risk environment for collecting multi-modal human-factor data in simulation, integrates the data into a mission timeline, and uses this to identify SIGACTs as short, time-bounded operationally significant events capturing breakdowns in situational awareness and team coordination. These SIGACTs guide the analysis and specification of missing requirements. We use MOMUx to study 12 teams supervising swarms of UAVs, identify 35 SIGACTs which were distilled into 24 new and clarifying requirements, and demonstrate that these requirements are relevant to real-world settings.