ios-ai-ml
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns, obfuscation, or insecure practices were detected across the analyzed files. The skill provides legitimate developer guidance for CoreML, Vision, and Natural Language frameworks.
- [DYNAMIC_EXECUTION]: The skill documents the standard iOS pattern for dynamic model loading (
MLModel.compileModelandMLModel(contentsOf:)). While this allows execution of binary model data downloaded at runtime, it is the idiomatic and documented approach for on-device AI in the Apple ecosystem. - [EXTERNAL_DOWNLOADS]: Mentions the management of models downloaded from remote URLs. This is documented neutrally as a common architectural requirement for iOS ML apps and does not point to specific untrusted sources.
- [DATA_EXFILTRATION]: No data exfiltration patterns were found. The skill explicitly promotes privacy-preserving patterns (Section 8), such as on-device personalization and avoiding cloud uploads for inference.
Audit Metadata