The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of the kubectl and flux command-line interfaces to manage system resources and cluster state. It also utilizes wget and curl to interact with internal service APIs.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8) because it processes untrusted data from the cluster environment.
Ingestion points: Untrusted data enters the context through kubectl logs, kubectl get events, and flux get commands as described in SKILL.md.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded commands within the fetched data.
Capability inventory: The skill has access to shell execution, cluster authentication secrets, and internal network tools (curl, wget).
Sanitization: No validation or sanitization is performed on the output of cluster commands before processing.
[DATA_EXFILTRATION]: The skill accesses sensitive cluster credentials stored in the local file system at ~/.kube/dev.yaml, ~/.kube/integration.yaml, and ~/.kube/live.yaml to establish connectivity.
[SAFE]: The skill provides references to well-known technology documentation sites, such as grafana.github.io and prometheus-io, for researching service configurations.

k8s