skills/nvidia/skills/doca-rdma/Gen Agent Trust Hub

doca-rdma

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the user on how to run various diagnostic and configuration commands, such as dmesg, mlxconfig, and ibv_devinfo. Several of these commands are correctly identified as requiring elevated privileges (sudo), which is standard practice for interacting with low-level network hardware and kernel drivers.
  • [EXTERNAL_DOWNLOADS]: The skill references external documentation and sample code located on nvidia.com and official NVIDIA GitHub organizations (github.com/NVIDIA-DOCA, github.com/NVIDIA). These are trusted vendor sources and do not present a security risk.
  • [INDIRECT_PROMPT_INJECTION]: The skill serves as a technical assistant, processing user-provided code and environment details. It generates build and execution instructions that incorporate user-specified variables (e.g., device IDs or paths). The attack surface is minimal as the guidance is restricted to the local development environment for the DOCA SDK.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 11:46 PM
Security Audit — agent-trust-hub — doca-rdma