The Agent Skills Directory

[SAFE]: The skill prioritizes operational security by enforcing strict human-in-the-loop requirements for any high-stakes actions, such as production deployments or destructive resource modifications.
[SAFE]: Core operational rules provided in the skill, such as those for Kubernetes (rules/kubernetes-safety.md) and CI/CD (rules/ci-cd-reproducibility.md), emphasize security best practices including least privilege and environment pinning.
[SAFE]: A local validation script (scripts/validate-output.sh) is included to ensure the agent's output maintains the required safety structure, such as checking for human approval prompts when production is mentioned, and does not pose a security risk.

ops-engine