unable to build kubernetes client config: invalid configuration when running helm install --dry-run

Goal: Use helm install --dry-run to validate a chart locally without providing production Kubernetes credentials, aiming for zero network traffic.

Constraint: The documentation does not explicitly state whether Helm will perform a server-side dry‑run (requiring API access) or fall back to client‑side rendering when a valid kubeconfig is missing, and this behavior depends on the cluster version and the dry‑run feature gate.

Questions: Does helm install --dry-run attempt to contact the API server when KUBECONFIG is unset or points to an empty file? If a server-side dry‑run is not available, does Helm automatically render locally without any cluster call? How can one determine which path was taken without enabling verbose debug logs that might still trigger a request?