provisioning-cluster-for-production

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). This skill’s required workflow includes runtime “curl” of publicly hosted Kubernetes operator CRDs/operator YAMLs (outsider-authored free text fetched from raw.githubusercontent.com) that would be ingested by the agent/automation and thus can carry indirect prompt-injection content.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill contains runtime install/apply commands that fetch and execute remote content from https://binaries.cockroachdb.com/cockroach-v.linux-amd64.tgz, https://raw.githubusercontent.com/cockroachdb/cockroach-operator/master/install/crds.yaml, https://raw.githubusercontent.com/cockroachdb/cockroach-operator/master/install/operator.yaml, and https://charts.cockroachdb.com/ (curl | tar, kubectl apply -f, and helm repo add/helm install), so these URLs are runtime external dependencies that execute remote code/manifests.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 0.80). The skill includes commands that modify system-wide locations (notably copying the cockroach binary to /usr/local/bin) and other operations that typically require elevated privileges, so it encourages changes to the host system state that may need sudo/root access.