nvidia-cuda-performance
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides a structured framework for NVIDIA CUDA performance optimization, focusing on technical modeling and evidence-based engineering.
- [SAFE]: No evidence of prompt injection or attempts to bypass safety filters was found. The instructions are purely technical and domain-specific.
- [SAFE]: All external links point to official NVIDIA documentation, reputable academic sources (ArXiv), or well-known open-source repositories (NVIDIA, Dao-AILab, IST-DASLab) on GitHub. None of these involve automated execution of untrusted code.
- [SAFE]: The skill does not contain any obfuscated commands, Base64-encoded payloads, or hidden text patterns.
- [SAFE]: There are no persistence mechanisms, privilege escalation attempts, or data exfiltration patterns. The skill focuses on analyzing performance metrics (latency, throughput, occupancy) rather than accessing sensitive system data.
Audit Metadata