Iron Harbor

Iron Harbor is the course I wished existed when I started: a local-first path into network security automation where you finish real systems instead of collecting slides. I author it, including the platform it runs on.

The structure is one integrated portfolio grown through seven evidence-gated checkpoints. A learner starts from a deliberately defective starter system and repairs it: an idempotent Linux and SSH baseline with Ansible, a reversible EVE-NG observation gateway, a tested Python assurance runner around Nmap and TLS, an OPA/Rego policy gate in CI, a Suricata sensor fed by deterministic PCAP replay, and a temporary AWS VPC mirror that must produce drift, cost, and destroy proof before it earns its place.

The gates are the point. A checkpoint does not advance because the code runs; it advances because the evidence exists: tests pass, rollback is demonstrated, failures are recovered and written down. AI may help write the implementation, but the architecture, authorization, blast radius, and release decision stay human, because those are the parts that make someone an engineer.

Building this has also been my own curriculum. Every checkpoint I designed is a skill I had to prove on myself first.