Lessons from the Front-line: What Research Reveals about MCP Security

The Model Context Protocol (MCP) is quickly emerging as a key standard for linking LLMs to external data and tools. While it unlocks new capabilities, MCP also introduces security risks that feel all too familiar, echoing challenges from past waves of innovation.

This session is a chance to connect those dots. We’ll explore how known risks like supply chain vulnerabilities, remote server exposure, and client-side attacks show up in the context of MCP. More importantly, we’ll share practical guidance to help developers and AI teams treat MCP servers with the care they deserve, as privileged, high-value components in your stack.