AI coding agents don't need to be attacked to expose your data. In this session, James from Latio and Gil Friedman, Field CTO at Backslash Security, demonstrate exactly how it happens: a basic prompt that reaches further than anyone intended, malicious rules hidden in legitimate-looking configuration, and prompt injection arriving through GitHub issues and Context7. All live, not theoretical.
Watch to see what your existing tooling misses on developer workstations, and how Backslash Security discovers every agent, MCP server, and Skill running there, then blocks risky actions in real time - without slowing developers down.
