The Agent Skills Directory

[PROMPT_INJECTION] (HIGH): The skill is vulnerable to Indirect Prompt Injection through the processing of Kubernetes logs, events, and resource descriptions.
Ingestion points: Untrusted data enters the agent context via kubectl logs, kubectl get events, and kubectl describe as seen in SKILL.md and references/COMMANDS.md.
Boundary markers: None detected. There are no instructions or delimiters defined to prevent the agent from interpreting content within logs or events as new commands.
Capability inventory: The skill involves extensive shell command execution (kubectl) and file system writes to /tmp as described in SKILL.md.
Sanitization: None detected. The skill does not provide mechanisms to sanitize or filter out potential escape sequences or hidden instructions in the ingested data.
[COMMAND_EXECUTION] (LOW): The skill's primary function is to provide a set of shell commands for the agent to execute. While these are standard kubectl commands, they grant the agent significant visibility and potential influence over the Kubernetes environment.

k8s-query