Fix e2e CI deploy step by skipping Helm CRD install

[Copilot]

The e2e GitHub Actions job failed during operator deployment because the workflow pre-applies CRDs with kubectl, then helm install attempted to install/adopt the same CRDs and hit Helm ownership-metadata validation errors.

• Root-cause alignment

  • Keep CRD lifecycle in the existing pre-install step (kubectl apply -f config/crd/bases/) and prevent duplicate CRD handling in Helm.

• Workflow change

  • Updated .github/workflows/ci.yml in the e2e job’s Deploy operator command.
  • Added --skip-crds to helm install drop charts/drop ... so Helm does not try to install or claim already-existing CRDs.

• Behavioral impact

  • Operator chart deployment proceeds using pre-installed CRDs, avoiding Helm’s CRD ownership conflict in CI.

helm install drop charts/drop \
  --namespace drop-system \
  --create-namespace \
  --skip-crds \
  --set image.repository=controller \
  --set image.tag=ci \
  --set image.pullPolicy=Never