kubernetes
Pass
Audited by Gen Agent Trust Hub on Sep 19, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill provides instructions for reading logs and resource descriptions, creating a surface where untrusted data from the cluster could be presented to the agent. 1. Ingestion points: 'kubectl logs' and 'kubectl describe' in SKILL.md. 2. Boundary markers: No delimiters or ignore-instructions are specified for the command outputs. 3. Capability inventory: Modification of cluster resources, interactive shell access (kubectl exec), and container deployment (kubectl run) are present in SKILL.md. 4. Sanitization: Not mentioned in the implementation.
- [COMMAND_EXECUTION]: The skill documents the use of 'kubectl exec' to perform interactive debugging inside containers.
- [REMOTE_CODE_EXECUTION]: The skill includes instructions for deploying utility containers using 'kubectl run' with standard images like busybox for network debugging.
Audit Metadata