AI Security Course: OWASP LLM Top 10 2026 and Agentic AI
A hands-on course in securing LLM applications and AI agents, built around the OWASP Top 10 for LLM Applications 2026 and the OWASP Top 10 for Agentic Applications. You build SupportBot, a deliberately vulnerable assistant with retrieval and tools, and attack and fix it lesson by lesson: direct and indirect prompt injection, sensitive information disclosure, hidden context exposure, excessive agency, agent goal hijack, tool and identity abuse, code execution, memory poisoning, supply chain and data poisoning, vector store weaknesses, improper output handling, misinformation and unbounded consumption. The second half teaches AI red teaming with garak, PyRIT and promptfoo as a CI gate, then production defenses, monitoring, incident response and governance with MITRE ATLAS, the NIST AI RMF and the EU AI Act. Each lesson is a short, hands-on read with tested code and review questions.