Dynamic safety cases (DSCs) offer a promising approach to ensuring the safety of autonomous and highly automated systems in complex and evolving operational contexts. While the theoretical underpinnings of DSCs have been explored in the recent past, practical applications remain scarce, and existing frame-works lack domain-specific clarity. This work tackles this gap by leveraging collaborative robots (cobots) as a real-world application for instantiating and examining DSCs. In this paper, we investigate how cobot-specific standards inform the structure and content of DSCs, and identify safety performance indicators that are both extractable from the safety case and meaningful for runtime assurance. Furthermore, we explore how the dimensions of dynamism -namely monitorability, assessibility and updateability- manifest in cobot systems, exploring tailored techniques based on system features like force limitation, speed limitation, and hand-guiding. We introduce a set of foundational dynamic safety case patterns applicable to cobot deployments. Finally, we propose integrating a digital twin within the DSC framework to reduce uncertainty, enhance coverage, and increase confidence in runtime safety assurance. This work contributes to a practical, standards-aligned, and technically grounded pathway for operationalizing DSCs in real-world cobot system.