eks

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

• Third-party content exposure detected (high risk: 0.80). The skill includes commands that fetch and apply public, user-hosted manifests (e.g., kubectl apply -f https://raw.githubusercontent.com/kubernetes/autoscaler/... and curl https://raw.githubusercontent.com/aws-samples/... | kubectl apply -f -) and installs images from public registries (oci://public.ecr.aws/...), which require ingesting untrusted third-party content as part of the workflow.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill contains runtime commands that fetch and apply remote manifests—specifically "kubectl apply -f https://raw.githubusercontent.com/kubernetes/autoscaler/master/cluster-autoscaler/cloudprovider/aws/examples/cluster-autoscaler-autodiscover.yaml" and "curl https://raw.githubusercontent.com/aws-samples/amazon-cloudwatch-container-insights/latest/.../cwagent-fluent-bit-quickstart.yaml | ... | kubectl apply -f -"—which execute remote code/configuration and are used by the skill's example steps.