Bump default CNI config cniVersion to 1.0.0

[tsorya]

What

Raise the cniVersion declared in the default CNI configuration from 0.3.1 to 1.0.0 in all shipped manifests and the Helm chart.

Why

OpenShift 4.23+ upgraded multus to CNI spec 1.1.0, which requires delegate plugins to declare cniVersion >= 0.4.0. Flannel's manifests still declared 0.3.1, causing multus to crash on startup:

failed to load the primary CNI configuration as a multus delegate
with error 'delegate cni version is 0.3.1 while top level cni version is 1.1.0'

Files changed

• Documentation/kube-flannel.yml
• Documentation/kustomization/kube-flannel/kube-flannel.yml
• Documentation/minikube.yml
• chart/kube-flannel/values.yaml

Why it's safe

• Every supported container runtime accepts 1.0.0 configs (containerd ≥ 1.6, CRI-O ≥ 1.24)
• The flannel CNI plugin (flannel-io/cni-plugin) already supports CNI spec 1.0.0
• Existing pod attachments are unaffected on upgrade: the runtime caches each attachment's config and replays the original cniVersion at DEL time

Reference

• Same fix applied to Calico: projectcalico/calico#12965, projectcalico/calico#13352
• Cilium already shipped 1.1.0 support: cilium/cilium#37028

Testing

• OpenShift 4.23+ with multus 1.1.0 — flannel as primary CNI
• Standard kubeadm cluster — no regression
• Upgrade path 0.3.1 → 1.0.0 — existing pods unaffected