The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or persistence mechanisms were detected. The instructions are aligned with best practices for cluster security and reliability.
[EXTERNAL_DOWNLOADS]: Fetches configuration assets and infrastructure operators (Kyverno, Gatekeeper, Velero, OpenCost, Prometheus stack) from trusted and official repositories managed by established organizations in the Cloud Native ecosystem.
[CREDENTIALS_UNSAFE]: Includes connection string templates in references/secrets-external-secrets.md. Analysis confirms these are non-functional placeholders (e.g., {{ .username }}, {{ .password }}) intended for use with the External Secrets Operator's template system and do not represent hardcoded credentials.
[COMMAND_EXECUTION]: Provides guidance on executing administrative commands using standard Kubernetes tools (kubectl, helm, velero) for cluster management and auditing.
[DATA_EXFILTRATION]: Documents standard procedures for backing up cluster state and volume data to external object storage for disaster recovery, which is the intended functional purpose of the backup section.
[PROMPT_INJECTION]: No attempts to subvert AI behavioral constraints were identified. Regarding Indirect Prompt Injection (Category 8): Ingestion points: User-provided project context and constraints (SKILL.md); Boundary markers: Absent; Capability inventory: Administrative shell execution (kubectl, helm) and cluster configuration modification; Sanitization: Absent.

kubernetes-production