Request

The Helm chart setup scripts (set-up-test.sh, set-up-multi-dedicated-test.sh,
set-up-multi-shared-test.sh) currently have no way to test with a locally-built Docker image.
They always deploy with the default image from values.yaml (the published
ghcr.io/y-scope/clp/clp-package image).

When developing Helm chart changes, we often need to test with a local clp-package image built from the working tree. This requires manually editing the helm install command each time, which is error-prone and not scriptable.

Add a --clp-package-image CLI argument to all setup scripts so that a locally-built image can be loaded into the kind cluster and used for deployment.

Possible implementation

Add two helpers to .set-up-common.sh:

1. parse_common_args() — parses --clp-package-image <image> from the script arguments and sets CLP_PACKAGE_IMAGE.
2. get_image_helm_args() — if an image is specified, loads it into the kind cluster via kind load docker-image and prints the --set flags for image.clpPackage.repository, image.clpPackage.tag, and image.clpPackage.pullPolicy=Never.

Update all three set-up-*.sh scripts to call parse_common_args "$@" and pass
"${CLP_PACKAGE_IMAGE}" to get_image_helm_args.

Usage:

bash tools/deployment/package-helm/set-up-test.sh --clp-package-image clp-package:dev-junhao-a6bf

When --clp-package-image is omitted, the scripts behave exactly as before (using the default image from values.yaml).