The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run to execute the kubectl CLI and other system probes such as ibstat, nvidia-smi, and lsmod. These executions are performed using list-based argument passing (avoiding shell=True), which follows security best practices for preventing shell injection vulnerabilities.
[SAFE]: All executed commands are read-only diagnostics designed to probe hardware and environment configurations. The skill includes explicit instructions and code logic to avoid mutating the cluster or exposing sensitive secrets.
[DATA_EXFILTRATION]: While the skill reads configuration manifests and pod environment data, this information is only used for local analysis and reporting to the agent/user via standard output. No network exfiltration or unauthorized data transfers were identified.

dynamo-interconnect-check