The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill is vulnerable to instructions embedded in the data it processes.
Ingestion points: The skill reads external content through kubectl logs, kubectl describe, and kubectl get events (Workflow and Symptom sections).
Boundary markers: Absent. There are no instructions to the agent to distinguish between its own logic and instructions that might be found in logs.
Capability inventory: The skill executes modifying commands like kubectl set resources, kubectl rollout undo, and kubectl delete (Interactive Fix and Anti-Pattern sections).
Sanitization: Absent. No filtering or escaping is applied to the data retrieved from the cluster before it is used to determine actions.
[Command Execution] (HIGH): The skill provides a high-privilege interface to Kubernetes. Even with user confirmation steps, the risk of the agent suggesting or executing harmful commands based on manipulated or misinterpreted cluster state is substantial.

troubleshooting-kubernetes