docker-expert
Pass
Audited by Gen Agent Trust Hub on May 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to inspect the local environment, including 'docker version', 'docker info', 'docker ps', and 'docker images' to gather metadata about the container runtime.
- [COMMAND_EXECUTION]: To validate configurations, the skill uses 'docker build' to create images and 'docker run' to execute them locally, along with 'docker exec' for runtime inspection.
- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it processes untrusted local configuration files.
- Ingestion points: The skill searches for and reads 'Dockerfile' and 'docker-compose.yml' files from the local project directory using 'find'.
- Boundary markers: No specific boundary markers or instructions are used to distinguish between user-provided configuration and agent instructions.
- Capability inventory: The agent has the capability to build and run arbitrary containers and execute commands within them via shell access.
- Sanitization: There is no explicit sanitization or validation of the content of the ingested Docker files before the agent processes them.
Audit Metadata