Learn
Learn AI security
Technical, defender-focused lessons on how LLM applications, agents and AI coding tools get attacked and how to secure, test and monitor them. Each lesson cites the standard, specification or original research behind every claim and links into the ThreatDeck playbooks, detections and labs where you can apply it.
- 6 paths
- 35 lessons
- 5.3 hours
- Free, no account
- Progress saved only in this browser
- Path 1Foundation AI security fundamentals How LLM applications and agents work as attack surface, the trust-boundary tests practitioners use, and the frameworks (OWASP, MITRE ATLAS, NIST) that give you shared vocabulary. 5 lessons 46 min · 20 questions
- Path 2Practitioner Securing LLM applications Prompt injection and the architectures that contain it, safe handling of model output, permission-aware RAG, keeping secrets out of context, what guardrails can and cannot do, and security regression tests in CI. 6 lessons 54 min · 24 questions
- Path 3Practitioner Agents, tools and MCP Design agent tools and identities so a steered agent can do little harm, review MCP servers and their authorization like any other third-party code, and put approvals and logs where they change outcomes. 6 lessons 55 min · 24 questions
- Path 4Practitioner Securing AI coding assistants in an organisation What Claude Code, Codex, Copilot, Cursor and Gemini CLI can reach on a developer machine, how their approvals and sandboxes work and fail, and how to enforce policy, MCP allowlists, supply chain controls and telemetry across a fleet. 6 lessons 57 min · 24 questions
- Path 5Practitioner Red teaming AI systems Plan and run security red team engagements against LLM applications and agents: scoping and rules of engagement, agent and indirect-injection testing, open-source tooling, impact-based findings and continuous regression testing. 6 lessons 50 min · 24 questions
- Path 6Practitioner Detecting and responding to AI abuse LLMjacking, the cloud and agent telemetry that records AI use, detection engineering for AI abuse, agent incident response, AI-enabled attackers, and using AI in the SOC without trusting it blindly. 6 lessons 55 min · 24 questions
How the lessons are made
- Written for practitioners: how the attack works, which control stops it, how to test and detect it. No vendor pitches.
- Every factual claim has a numbered reference to a primary source: OWASP GenAI, MITRE ATLAS, NIST, the MCP specification, vendor documentation or the original research. Each lesson shows the date its links and facts were last checked.
- Self-check answers are explained in the page. Quiz scores and completion are stored in your browser's localStorage (
td.learn) and never sent anywhere. - Hands-on practice: use the labs only on systems you are authorised to test.