Read my latest writings

Exploring the dangerous world of prompt injection attacks in modern AI systems, from MCP servers to agentic browsers, and how to defend against them.