tao-run-on-slurm

Pass

Audited by Gen Agent Trust Hub on Jun 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes ssh to run commands on a remote SLURM login node. This includes job submission via sbatch, task execution with srun, and monitoring/management using squeue, sacct, and scancel. It also performs remote directory management and permission checks using standard shell tools.
  • [EXTERNAL_DOWNLOADS]: The skill instructions specify installing the nvidia-tao-sdk package. It includes an installation option from an internal vendor repository at gitlab-master.nvidia.com. As the skill is authored by NVIDIA, this reference to their own infrastructure is considered a trusted vendor resource.
  • [SAFE]: The skill prescribes secure handling of sensitive information, such as using a printf | ssh heredoc to install NGC_KEY credentials on the cluster, which prevents the secret from appearing in shell history or log files.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 30, 2026, 05:48 PM
Security Audit — agent-trust-hub — tao-run-on-slurm