realitykit
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill follows established best practices for iOS augmented reality development, including device capability checks, camera permission handling, and asynchronous resource loading.
- [SAFE]: The code examples for entity management, physics, and gestures are consistent with standard RealityKit patterns and do not contain hidden malicious logic or unauthorized system access.
- [SAFE]: External references are provided for documentation purposes. While these links point to a third-party domain (sosumi.ai) rather than official Apple documentation, they are static references and do not represent a security risk through active code execution or data exfiltration.
- [SAFE]: The skill defines an interaction surface for processing user-provided code and configurations for review. This represents a potential indirect prompt injection vector, but the skill itself operates as a set of static guidelines. 1. Ingestion points: User-provided descriptions of AR setups and code snippets for review (documented in evals/evals.json). 2. Boundary markers: Not present. 3. Capability inventory: Modification of Swift source code and iOS project settings (e.g., Info.plist). 4. Sanitization: Not applicable for these static instructions.
Audit Metadata