model-datasets

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). The workflow downloads and organizes AI models and datasets from HuggingFace and ModelScope registries, which involves fetching community-published repository content.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (medium risk: 0.30). The skill instructs the execution of a remote shell script via curl -LsSf https://astral.sh/uv/install.sh | sh during setup, introducing an external dependency from Astral Software.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). The skill prompt instructs the agent to perform low-level disk partitioning, formatting, and modifying system configuration files like /etc/fstab, which require root/sudo privileges and can compromise the machine's state.