Model. Knowledge. Agent.
Test the trust boundaries.
- Prompt injection and model instructions
- RAG and knowledge sources
- Agents, MCP and tool permissions
Explore browser-based AI security training. The syllabus, demo and waitlist are on a separate website.

