Add RHDH skills volume support to LCORE container

[yangcao77]

What does this PR do?

PR title

Add RHDH skills volume support to LCORE

Summary

• Adds an emptyDir volume for RHDH skills.
• Uses an init container with quay.io/redhat-ai-dev/utils:latest to fetch the main branch of redhat-developer/rhdh-skills.
• Copies the repository’s skills/ contents into the shared volume.
• Mounts the volume at /app-root/skills in lightspeed-core.
• Keeps the mount writable so skill scripts can create or modify files when required.
• Makes the PostSync patch idempotent across repeated Argo CD syncs.
• Adds CI coverage for initial deployment and repeated synchronization.

Testing

• bash tests/helm/test-skills-volume.sh
• helm lint charts/rhdh --set okp.enabled=false
• ShellCheck
• git diff --check

No new LCORE image or long-running sidecar is introduced.

Which issue(s) does this PR fix

https://redhat.atlassian.net/browse/RHIDP-16676

How to test changes / Special notes to the reviewer

[qodo-code-review]

ⓘ Qodo reviews are paused because the subscription is no longer active. Ask your workspace admin to reactivate the subscription to resume reviews. Manage billing

[Jdubrick]

Do we need to use utils image for this, or is there a really small image we can use that has git that could do the job?

[yangcao77]

I use quay.io/redhat-ai-dev/utils:latest because it is already used by this workflow, and provides Git and Bash. This avoids introducing another third-party image. I’m happy to switch if we have a preferred approved lightweight Git image.

[Jdubrick]

I'm good with keeping it as utils for now, maybe something we can revisit when we onboard to the newer rhdh chart that lets us natively add the volumes?

[yangcao77]

Latest commit d101568:

• Fixed the CI rollout failure by removing runAsNonRoot: true; the utils:latest image defaults to root and Kubernetes previously rejected it.
• Changed missing lightspeed-core handling from an error/exit to a warning.
• Skips skills and GCP mounts when LCORE is absent, while still applying the feedback sidecar.
• Updated the chart-upgrade TODO to include both skills and GCP mounts.
• Added regression tests for the security-context fix and missing-LCORE behavior.

[Jdubrick]

lgtm