gpu-use

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

• Third-party content exposure detected (high risk: 0.80). The skill SSHes into a user-specified or default remote host (e.g., "ssh {SSH_TARGET}" and the default ssh felix@124.158.103.16 -p 10022) and ingests untrusted runtime outputs (nvidia-smi, docker ps/inspect, /proc/*/environ) that the agent must read and which directly influence its reporting and resource-allocation decisions.

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

• Potentially malicious external URL detected (high risk: 1.00). Flagged because the skill performs runtime SSH to ssh felix@124.158.103.16 -p 10022, which executes commands on a remote host and is a required external dependency for the skill.