CoreWeave Security Basics
Instructions
Secrets for Model Access
# HuggingFace token
kubectl create secret generic hf-token --from-literal=token="${HF_TOKEN}"
# Container registry credentials
kubectl create secret docker-registry regcred \
--docker-server=ghcr.io \
--docker-username=$USER \
--docker-password=$TOKEN
Network Policy for Inference Pods
apiVersion: networking.k8s.io/v1
kind: NetworkPolicy
metadata:
name: inference-isolation
spec:
podSelector:
matchLabels:
app: inference-server
policyTypes: [Ingress, Egress]
ingress:
- from:
- podSelector:
matchLabels:
role: api-gateway
ports:
- port: 8080
egress:
- to: [] # Allow all egress for model downloads
ports:
- port: 443
Security Checklist
- Kubeconfig stored securely, not in repos
- Secrets used for model tokens (not env vars in YAML)
- Network policies restrict inference endpoint access
- RBAC limits namespace access per team
- Container images scanned for CVEs
- PVCs encrypted at rest
Resources
Next Steps
For production readiness, see coreweave-prod-checklist.