The Agent Skills Directory

[SAFE]: The skill provides detailed guidance on container hardening, including the use of non-root users (adduser, USER appuser), multi-stage builds to minimize image size, and the use of specific image digests for reproducibility.
[SAFE]: Explicitly identifies and warns against critical security anti-patterns such as privileged: true, network_mode: host, and the dangerous practice of bind-mounting /var/run/docker.sock.
[SAFE]: Recommends secure secret management techniques, such as using BuildKit secrets (RUN --mount=type=secret) to avoid leaking credentials in image layers, and advises against committing .env files to source control.
[SAFE]: Provides a robust container supply chain playbook in references/supply-chain.md, recommending the use of official security scanners (Trivy, Grype), SBOM generation, and image signing (Cosign/Sigstore).

docker-patterns