The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions for executing system commands for Kubernetes cluster management and GitOps operations using kubectl and flux.- [DATA_EXFILTRATION]: The instructions direct the agent to access sensitive Kubernetes configuration files located at ~/.kube/*.yaml. These files contain authentication tokens and cluster certificates. Additionally, the skill interacts with internal service endpoints on the tomnowak.work domain.- [REMOTE_CODE_EXECUTION]: The skill describes patterns for executing remote commands within cluster containers via kubectl exec combined with network operations like wget.- [PROMPT_INJECTION]: The skill exhibits vulnerability to indirect prompt injection. 1. Ingestion points: The agent is instructed to read untrusted data from the cluster using kubectl logs, kubectl describe, and kubectl get events. 2. Boundary markers: No boundary markers or instructions to ignore embedded content are provided. 3. Capability inventory: The skill provides high-privilege access via kubectl and flux commands. 4. Sanitization: There is no evidence of sanitization or filtering applied to the data retrieved from the cluster API.

k8s