In our latest Cato Connect webinar, the API & DevOps team explored how to secure AI agents; starting with local agents that can access tools, MCP servers, skills, and sensitive data on a user’s device. We covered key risks like indirect prompt injection, unverified tools, excessive permissions, and data exposure, along with practical ways to discover, monitor, govern, and enforce controls without slowing AI adoption.
Stay tuned for the next session on managed and custom agents!
