The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses aws CLI and kubectl to perform diagnostic checks on email infrastructure, SES identities, and Kubernetes workload status. These operations are scoped to discovery and troubleshooting.
[DATA_EXPOSURE_AND_EXFILTRATION]: The instructions include robust defensive measures to prevent credential leakage. It explicitly prohibits printing, logging, or echoing secret values, and mandates the use of an external Bitwarden-based helper skill (bws-secrets) for handling sensitive values. It further instructs against inspecting Kubernetes Secret contents directly.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data via kubectl logs and repository searches using ripgrep. Although this creates an ingestion surface for potentially untrusted strings, the skill mitigates this by enforcing a read-only-by-default posture and requiring explicit user authorization for any state-changing actions.

email-setup