The Agent Skills Directory

[COMMAND_EXECUTION] (SAFE): The skill contains multiple shell scripts and markdown documentation that use the kubectl CLI. These are standard operations for managing Kubernetes clusters (e.g., getting pods, viewing logs, scaling deployments) and do not pose a risk outside of their intended use for cluster administration.
[DATA_EXFILTRATION] (SAFE): While the skill demonstrates how to view Kubernetes secrets and logs, it does not contain any logic to transmit this data to external or untrusted domains. Commands like kubectl get secret are intended for local administrative debugging.
[CREDENTIALS_UNSAFE] (SAFE): Several files (SKILL.md, references/manifests.md) contain placeholder secrets and example API keys (e.g., sk-secret-key, secret123). These are clearly marked as examples and do not represent actual hardcoded credentials.
[EXTERNAL_DOWNLOADS] (SAFE): SKILL.md mentions kubectl apply -f https://example.com/manifest.yaml as a syntax example. This is a standard Kubernetes usage pattern and does not point to a malicious source.
[REMOTE_CODE_EXECUTION] (SAFE): The skill includes a Python script (scripts/health-checker.py) that uses subprocess.run to call kubectl. This is a legitimate way to automate cluster health checks and does not use untrusted input.

kubernetes