Skip to main content

Security checks overview

GuardMap runs 48 checks across six categories on every scan.

Categories

CategoryChecksDescription
Pod Security21Container-level checks — privileges, capabilities, root user, resource limits, plaintext secrets, public images
RBAC12Role and binding checks — wildcards, cluster-admin, exec/attach, secrets access, node access
Network5NetworkPolicy coverage, allow-all policies, host networking, public load balancers
IAM / IRSA6AWS IAM permission checks for EKS IRSA chains, unused IRSA bindings
Batch / Workload3CronJob and Job hygiene — TTL, concurrency, missing deadline
Secrets / Config1Orphaned Secrets (exist but are unreferenced by any workload)

Severity levels

SeverityMeaningMax score impact
CriticalImmediate risk — can lead to full cluster or node compromise−42 pts
HighSignificant risk — increases blast radius of other vulnerabilities−28 pts
MediumElevated risk — should be addressed but not immediately dangerous−14 pts
LowBest practice violation — low direct risk but increases attack surface−6 pts

System namespace exclusions

GuardMap skips checks in system namespaces by default:

  • kube-system
  • kube-public
  • kube-node-lease
  • ingress-nginx
  • cert-manager
  • guardmap (the agent itself)

This avoids noise from privileged system components you don't control.

Check count by category

CategoryCriticalHighMediumLow
Pod Security4746
RBAC3432
Network221
IAM / IRSA2121
Batch / Workload12
Secrets / Config1