Docker
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides high-quality security recommendations, explicitly instructing users to avoid dangerous practices such as running containers as root, hardcoding secrets in image layers (ENV/ARG/COPY), and using the
--privilegedflag. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to process and analyze untrusted external data, including user-provided Dockerfiles, Compose YAML files, and container logs. This creates an inherent attack surface for indirect prompt injection if those files contain malicious instructions intended to manipulate the agent's behavior.
- Ingestion points: Analyzes user-provided Docker configuration files and runtime logs at the user's request.
- Boundary markers: None explicitly defined in the instruction files; relies on the underlying agent's standard content delimiters.
- Capability inventory: The skill provides guidance and command suggestions but does not contain scripts for automatic execution of arbitrary shell commands beyond standard Docker CLI usage.
- Sanitization: No explicit sanitization of Dockerfile content is mentioned, which is typical for static analysis and educational tools.
Audit Metadata