The Agent Skills Directory

[SAFE]: The skill provides standard configuration templates for Docker and Docker Compose. It follows security best practices by demonstrating multi-stage builds to minimize image size and implementing non-root user execution (USER appuser) to reduce the attack surface.
[SAFE]: The Docker Compose file contains hardcoded credentials (POSTGRES_PASSWORD: pass, DATABASE_URL=postgres://user:pass@db:5432/mydb). These are clearly dummy placeholders for local development and do not represent actual sensitive data exposure.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided application code and package manifests to generate Dockerfiles and Compose configurations.
Ingestion points: Project source code, package.json, requirements.txt, go.mod, and go.sum files in SKILL.md examples.
Boundary markers: None explicitly defined in the provided templates to delimit untrusted input from build commands.
Capability inventory: The skill facilitates command execution within containers via npm ci, go mod download, go build, and uv pip install (via SKILL.md).
Sanitization: No specific sanitization or validation of input file content is implemented within the provided templates.

docker