The rise and fall of MoltBot is a sign that AI Agents are being rushed to deployment. How much of your data can you trust to AI Agents? In this episode of the Tech Field Day podcast, Girard Kavelines and Aunudrei Oliver join Alastair Cooke to cast an eye over AI agent security and data security. ClaudeBot exploded into popular consciousness as an AI helper that could gain new skills and help organize and simplify your life. Following the name changes to MoltBot and OpenClaw, some issues came to light. OpenClaw learned new skills by installing software from a highly open repository. Malware authors jumped on board and put data-stealing software into popular skills. The cautionary tale here is that your data needs protection from bad actors, and basic security practices are vital. The rush to deploy an agent or any new technology often tramples on good governance and security practices, with predictable poor outcomes. AI agents will undoubtedly bring great value to businesses and individuals, provided that robust controls are built in from the start.