nvshmem-select-remote-transport
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is designed for read-only system inspection and transport recommendation.
- [COMMAND_EXECUTION]: The skill executes a local diagnostic script 'scripts/collect-transport-facts.sh'. The instructions explicitly mandate that the agent must inspect this script before its first invocation to ensure safety. The script itself performs standard system discovery using common Linux diagnostic utilities.
- [DATA_EXFILTRATION]: No network operations or data exfiltration patterns were detected. The skill instructions specifically state that no network access is required for inspection.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes user-supplied CUDA kernel source code. While this presents a theoretical surface for indirect prompt injection, the risk is mitigated by the skill's structured output format which requires technical justification for recommendations based on observed patterns.
Audit Metadata