containerization
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the Bash tool to perform legitimate operations including Docker image building, AWS ECR authentication, and security scanning with Trivy and Grype. These commands are necessary for the skill's functionality and follow standard industry patterns.
- [PROMPT_INJECTION]: The skill utilizes templates that interpolate variables like ${APP_NAME} and ${VERSION} into shell commands and cloud manifests.
- Ingestion points: Data extracted from project documents and local application paths.
- Boundary markers: Not present in the shell command templates.
- Capability inventory: The agent has access to Bash, Write, and Edit tools.
- Sanitization: The skill does not explicitly define sanitization steps for the interpolated variables, creating a surface where malicious project metadata could influence command execution.
- [EXTERNAL_DOWNLOADS]: The skill references and downloads base images from well-known and reputable sources such as gcr.io/distroless and eclipse-temurin. These are recognized as safe, established services.
Audit Metadata