Docker-hardened
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill is designed to guide the agent in executing Docker CLI commands, which can perform system-level operations on the host. This capability is enabled through the metadata requirement for the docker binary, allowing the agent to manage containers, images, and host resources.
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted user data such as Dockerfiles, Compose files, and operational instructions.
- Ingestion points: User-provided Docker configurations and command parameters enter the agent's context during task execution.
- Boundary markers: The skill includes a 'Security Guardrails' section in SKILL.md that explicitly instructs the agent to confirm destructive actions and reject unsafe configurations like broad host path mounts.
- Capability inventory: The skill facilitates container management, including network creation, volume mounting, and image building via the docker tool.
- Sanitization: The skill provides logic for the agent to identify dangerous flags and promotes the use of secure credential injection methods, although it relies on the underlying model's adherence to these instructions.
Audit Metadata