hf-cloud-serving-image-selection
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The scripts/mirror_image.py script executes aws and docker CLI commands to manage container images. It uses subprocess.run with list-based arguments, which is a standard and secure practice for calling external tools.
- [EXTERNAL_DOWNLOADS]: The skill fetches container image metadata and model configurations from official AWS and Hugging Face repositories. These are well-known, trusted organizations, and the operations are performed for legitimate deployment verification purposes.
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to fetch config.json from Hugging Face model repositories to determine model architecture. * Ingestion points: SKILL.md (via curl to huggingface.co). * Boundary markers: Absent. * Capability inventory: scripts/mirror_image.py (subprocess and docker calls), SageMaker deployment operations. * Sanitization: Absent; however, the skill provides a strong security recommendation to set SM_VLLM_TRUST_REMOTE_CODE to 'false' to prevent the execution of arbitrary code from model artifacts.
- [DYNAMIC_EXECUTION]: The skill uses a Python script (scripts/mirror_image.py) to dynamically construct and execute shell commands for container management. This is a standard utility function and uses secure implementation patterns avoiding shell=True.
Audit Metadata