The Agent Skills Directory

[SAFE]: The skill serves as a comprehensive reference for Docker best practices, focusing heavily on security hardening, such as using non-root users, minimizing image size, and avoiding the storage of secrets in image layers. It explicitly warns against anti-patterns like hardcoding credentials.
[EXTERNAL_DOWNLOADS]: The skill references Docker base images from trusted and well-known sources, including Alpine Linux, Google Container Registry (Distroless), Chainguard (Wolfi), and official language repositories for Node.js, Python, Go, Java, and Rust. These references are standard for container development and originate from reputable vendors.
[COMMAND_EXECUTION]: The skill suggests the use of established security scanning tools such as Docker Scout, Trivy, and Grype, as well as native Docker diagnostic commands. These tools are recommended for identifying vulnerabilities and debugging container configurations.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze user-provided code and Docker configurations.
Ingestion points: User-provided application source code, Dockerfiles, and Docker Compose configurations provided to the agent.
Boundary markers: None explicitly requested by the skill for the agent to wrap around analyzed content.
Capability inventory: The skill provides instructions for creating and auditing Docker configurations, involving file analysis and suggested command execution.
Sanitization: The skill mitigates risks by instructing the agent to strictly follow security principles such as non-root execution and secret management, reducing the impact of potentially malicious inputs.

docker