Moltbot, an open‑source agentic AI that can act autonomously across apps, has surged in popularity and reignited debates over campus cybersecurity. The agent’s capabilities—accessing files, credentials and persistent memory—prompted warnings from Palo Alto Networks and other security firms that Moltbot-style agents could enable novel prompt‑injection and delayed‑execution attacks. Campus IT leaders and research computing units face immediate choices about policy, endpoint controls and training as faculty and students experiment with agentic tools. The episode highlights tradeoffs between productivity gains from autonomous assistants and institutional obligations to protect research data, student information and infrastructure.