The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8) by ingesting untrusted user data and interpolating it into generated artifacts.
Ingestion points: Requirements gathering in Stage 1 (SKILL.md), specifically application entry points, frameworks, and versions.
Boundary markers: The skill lacks explicit instructions for using delimiters or boundary markers when interpolating user-provided strings into the Dockerfile template.
Capability inventory: The skill generates executable file content (Dockerfiles) which will be processed by the Docker engine.
Sanitization: There is no evidence of sanitization, escaping, or validation of user-provided strings before they are written to the Dockerfile.
[COMMAND_EXECUTION]: The skill contains instructions to execute local shell scripts and external binaries as part of a fallback validation workflow.
Evidence: SKILL.md (Stage 5) provides commands for local execution: bash ../dockerfile-validator/scripts/dockerfile-validate.sh Dockerfile, hadolint Dockerfile, and checkov -f Dockerfile --framework dockerfile.
Context: While intended for security validation of generated artifacts, these represent the execution of subprocesses and local scripts based on relative file paths.

dockerfile-generator