ovphysx-drop-contact-acceptance
Pass
Audited by Gen Agent Trust Hub on Aug 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill directs the agent to run a versioned local Python script, scripts/run_ovphysx_drop_probe.py, to interface with the native OVPhysX runtime. This is an intended functionality for simulation verification.
- [PROMPT_INJECTION]: The skill processes external 3D files, creating a surface for indirect prompt injection. Ingestion points: .blend and USD files specified in SKILL.md. Boundary markers: Not present for descriptive data. Capability inventory: Execution of scripts/run_ovphysx_drop_probe.py. Sanitization: Source file hashing and validation of pose data (rejection of NaN/Inf) in SKILL.md.
- [SAFE]: The skill requires working on file copies rather than original data and implements source integrity hashing before and after simulation runs.
Audit Metadata