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

doca-dma

Pass

Audited by Gen Agent Trust Hub on Oct 16, 2026

Risk Level: SAFEPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [PRIVILEGE_ESCALATION]: The skill provides instructions to run applications as the root user on BlueField DPU hardware and to use sudo when executing the dmesg command to capture kernel driver logs for debugging purposes.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a vulnerability surface for indirect prompt injection as it is designed to ingest and process data from the user's local development environment, including source code and system configuration files.
  • Ingestion points: The skill reads files from the standard installation path /opt/mellanox/doca, queries system state via pkg-config and doca_caps, and analyzes user project files for modification guidance.
  • Boundary markers: No specific delimiters or safety warnings for the agent to ignore embedded instructions within user-provided code are documented.
  • Capability inventory: The skill maintains capabilities to execute shell commands (ls, cat, pkg-config, doca_caps, dmesg) to perform system discovery and debugging.
  • Sanitization: The instructions do not describe explicit sanitization or validation of the content retrieved from the user's filesystem.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 16, 2026, 05:17 AM