deepstream-profile-pipeline

Pass

Audited by Gen Agent Trust Hub on Jul 2, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill provides detailed instructions for performance benchmarking and optimization of DeepStream applications. All operations described are standard procedures for developers working in this domain.
  • [COMMAND_EXECUTION]: The Python script scripts/capacity_report.py executes nvidia-smi using subprocess.run. This is performed with a fixed list of arguments (not a shell string) to retrieve hardware metadata such as GPU name and compute capability. This is essential for the script's core functionality of calculating hardware-specific ceilings.
  • [EXTERNAL_DOWNLOADS]: The skill references the official NVIDIA container registry (nvcr.io) and documentation (developer.nvidia.com). These are verified vendor resources consistent with the skill's authorship.
  • [DATA_EXPOSURE]: No exfiltration patterns or hardcoded credentials were found. The skill only processes performance metrics and hardware specifications.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 2, 2026, 12:52 AM
Security Audit — agent-trust-hub — deepstream-profile-pipeline