Skip to main content

Security & DevSecOps

A policy engine and a secrets workflow, then OPA, Kyverno, Falco, Vault and cloud security posture.

Security work in operations is mostly the same few mechanisms applied at different layers: something that holds a secret and hands out short-lived leases, something that says no before an object exists, something that watches system calls after it runs, something that decides which pod may talk to which, and something that turns a scanner's thousand findings into the five that matter. Earlier phases already built some of them (capabilities and seccomp, RBAC, IAM evaluation, workflow and supply-chain security). This phase builds the rest from the standard library and checks each against the tool teams run in production.

How this phase is built​

LessonTypeWhat you buildWhat you then use
AdmissionBuildminiadmit: a validating and mutating webhook, policies as data, a CEL subset like ValidatingAdmissionPolicyValidatingAdmissionPolicy, Kyverno, Gatekeeper
Runtime detectionBuildminifalco: events from strace, a Falco-style rule language, detections and rule tuningFalco with the modern eBPF driver, Tetragon
NetworkPolicyBuildmininetpol: the NetworkPolicy evaluator, a connectivity matrix and explainNetworkPolicy on kind, Cilium and Calico, AdminNetworkPolicy
Secret scanningBuildminileaks: provider rules, entropy, Git history, allowlists, baselines and a pre-commit modegitleaks, push protection, git filter-repo
Workload identityBuildminiid: SPIFFE IDs, X.509 SVIDs from a small CA, peer attestation, rotation and mutual TLS authorised by IDSPIRE, service meshes, csi-driver-spiffe
BenchmarksBuildminibench: CIS Kubernetes checks as data against a node, with evidence and reportskube-bench, kubescape

More lessons for this phase are being written; the plan is in ROADMAP.md.

Reference layer​

The long-form guides, cheat sheet and interview questions for this phase live under DevSecOps and Cloud Security. The CKS exam-prep guide maps several of these lessons onto its domains.