All protocols
AI Security & Red Teaming
Attack and defend AI systems: injection, jailbreaks, agentic risks.
This protocol is still being written. Follow the Discord to hear when it drops.
Why AI Fails
Know your target: how and why models break.
0/5 lessons completed · ~7 min
Prompt Injection
The #1 vulnerability of LLM apps, and why it's still unsolved.
0/5 lessons completed
Jailbreaking & Red Teaming
Bypassing guardrails, and doing it responsibly.
0/5 lessons completed
Attacking the Model
Attacks on training data, weights, and privacy.
0/4 lessons completed
Agentic Security
Securing agents: tools, MCP, permissions, blast radius.
0/5 lessons completed
Defending AI Systems
The defender playbook: guardrails, filtering, evals, architecture.
0/5 lessons completed