realitykit
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill contains legitimate technical documentation and Swift code snippets for iOS development using Apple's RealityKit and ARKit frameworks.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill references documentation on an external domain (sosumi.ai). While this is not a whitelisted service, these are static hyperlinks for documentation and do not involve active network operations or data exfiltration.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided code and plans for review, creating an ingestion point for untrusted data. However, the skill does not have capabilities such as network access or file system modification, which significantly mitigates the potential impact of indirect injection.
- [METADATA_POISONING]: The skill refers to future software versions (Swift 6.3, iOS 26+). This appears to be a minor metadata discrepancy or placeholder rather than a malicious attempt to deceive users or bypass security controls.
Audit Metadata