Regularly priced at $249, now free for AEI members.
Master the cutting edge of AI security with AEI’s Agentic Security Engineering, a 4-hour practitioner-focused course. Move beyond static firewalls to design dynamic runtime trust boundaries, implement in-loop policy enforcement, and produce cryptographic evidence for autonomous agents. Built on AEBOP T1.2 with hands-on labs.
Escape static thinking. This module diagnoses why firewalls and checklists fail for autonomous agents and introduces the core principle of "runtime trust." You'll learn to see security as a moving boundary that must be embedded within the cognition loop, not wrapped around it.
Go from theory to tangible artifacts. Learn to map the agent's attack surface with trustmap.yaml and author enforceable guardrails with policy.yaml. Through hands-on labs, you'll replace static credentials with dynamic, task-scoped identity and establish your first runtime security boundaries.
Embed security into the cognitive loop. Build and deploy a runtime interceptor (security.py) that validates every action before execution. You'll then generate cryptographically signed evidence for each decision, streaming immutable proof of control to a Trust Ledger.
Advance your systems methodically. Apply the L0-L5 maturity model to assess your current state and execute a structured upgrade path. Implement enterprise patterns for policy-as-code and prepare for cross-domain federation, moving from contained agents to a verifiable trust fabric.
Transition from project to practice. Establish the metrics, rituals, and automated playbooks needed for ongoing security assurance. Learn to detect drift, orchestrate recovery, and provide continuous proof of compliance—transforming security from an implementation into an intrinsic system behavior.