ovphysx-host-runtime-boundary

Pass

Audited by Gen Agent Trust Hub on Jul 22, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill establishes a controlled workflow for bridging a sandboxed execution environment with a host-based physical simulation runtime.
  • [COMMAND_EXECUTION]: The skill uses the runpy module to execute a pre-installed host script (ovphysx_host_helper.py). This is a legitimate part of the deployment topology for the NVIDIA Physical AI environment and uses a specific, authorized control channel.
  • [SAFE]: The instructions contain clear security directives to avoid passing unsanitized user input as file paths and to maintain strict separation between sandbox and host filesystem operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 22, 2026, 03:09 PM
Security Audit — agent-trust-hub — ovphysx-host-runtime-boundary