The Agent Skills Directory

[SAFE]: The skill functions as a security-focused educational and diagnostic tool for Docker containerization. It explicitly instructs the agent to identify and correct security anti-patterns such as running processes as root, using insecure tags, and leaking sensitive data into image layers.
[EXTERNAL_DOWNLOADS]: The skill references official base images (Node.js, Go, Python) and security tools from well-known services (Google's gcr.io, Yelp's GitHub, wagoodman's GitHub). These references are standard for the containerization domain and follow the trusted scope rules for well-known services.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided Dockerfiles for review and diagnosis. While this creates a theoretical attack surface for indirect prompt injection, the risk is mitigated by the skill's specific validation rules and its advisory role. (Ingestion points: User-supplied Dockerfiles; Boundary markers: None; Capability inventory: Purely advisory; Sanitization: Regex validation rules in references/validations.md).

docker-containerization