cluster-api

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs fetching and applying remote, public manifests (e.g., downloading clusterctl from https://github.com/..., running clusterctl init which downloads provider components, and kubectl apply -f https://raw.githubusercontent.com/projectcalico/.../calico.yaml), so untrusted third‑party content is ingested as part of the workflow and can materially influence tool behavior.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill includes explicit runtime commands that fetch and execute external artifacts required for operation — notably the clusterctl binary (https://github.com/kubernetes-sigs/cluster-api/releases/latest/download/clusterctl-linux-amd64) and a remote CNI manifest applied with kubectl (https://raw.githubusercontent.com/projectcalico/calico/v3.27.0/manifests/calico.yaml) — which are downloaded and executed during setup and are required dependencies.

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

• Attempt to modify system services in skill instructions detected (high risk: 0.90). The prompt includes commands that require sudo (e.g., moving binaries into /usr/local/bin), mounts the host Docker socket into containers, and instructs creating/deleting Docker containers and system-wide clusters—operations that modify host system state and can bypass host security, so it should be flagged.