arXiv:2605.00245v1 Announce Type: new Abstract: Large language models (LLMs) are now being explored for defense applications that require reliable and legally compliant decision support. They also hold significant potential to enhance decision making, coordination, and operationa