tao-run-on-slurm
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill installs the TAO SDK from NVIDIA's internal GitLab server (
gitlab-master.nvidia.com) or public registries. These are legitimate resources owned by the skill author. - [COMMAND_EXECUTION]: Uses standard utilities such as
ssh,sbatch,srun, andsacctto submit and monitor jobs on a SLURM-managed cluster. These operations are essential for the skill's intended purpose. - [INDIRECT_PROMPT_INJECTION]: The skill processes user-defined cluster hostnames and filesystem paths. It includes mandatory pre-execution validation steps (e.g., using
sshwithtest -e) and security warnings to ensure these inputs do not lead to unexpected command execution. - [CREDENTIALS_SAFE]: Provides rigorous instructions for handling SSH keys and NGC API keys. It specifically recommends using piped input to avoid secrets appearing in shell history and mandates setting restrictive file permissions (
chmod 600) for private keys.
Audit Metadata