The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/collect_dynamo_debug_bundle.py executes system commands via kubectl to retrieve pod logs and resource statuses. The execution uses list-based arguments which mitigates shell injection risks.
[DATA_EXFILTRATION]: The skill includes a robust redaction mechanism in its collection script. It uses regular expressions to identify and mask secrets, Bearer tokens, and Hugging Face tokens within command outputs before they are saved or displayed, reducing the risk of accidental credential exposure.
[SAFE]: The skill's behavior aligns with its documented purpose. It explicitly avoids collecting Kubernetes secrets and provides a structured approach to troubleshooting without requiring elevated or destructive permissions.

dynamo-troubleshoot