hz-unity-meta-mixed-reality-utility-kit

Pass

Audited by Gen Agent Trust Hub on Sep 25, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill documentation includes methods for loading scene data from external JSON strings, which creates a potential surface for indirect prompt injection if ingested room data contains malicious instructions in anchor labels or metadata.
  • Ingestion points: MRUK.LoadSceneFromJsonString in SKILL.md.
  • Boundary markers: None identified.
  • Capability inventory: Unity scene manipulation, NavMesh generation, and prefab spawning.
  • Sanitization: No input validation for the scene JSON is described.
  • [DYNAMIC_EXECUTION]: To facilitate interaction with the MRUK assembly from limited environments like Unity MCP, the skill documents a reflection-based approach to resolve types and invoke methods at runtime.
  • Evidence: Implementation details in SKILL.md under the 'Calling SDK Methods via Unity MCP' section using System.Reflection patterns.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 25, 2026, 06:57 PM
Security Audit — agent-trust-hub — hz-unity-meta-mixed-reality-utility-kit