apple-on-device-ai
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues were identified in the skill. The content consists of legitimate technical documentation, architectural patterns, and code samples for Apple Silicon machine learning development.
- [EXTERNAL_DOWNLOADS]: The skill references standard machine learning tools and community resources, such as the
coremltoolsPython package and Hugging Face model identifiers (e.g.,mlx-community/Qwen3-4B-4bit). These are well-known and expected resources in the machine learning domain. - [INDIRECT_PROMPT_INJECTION]: The skill demonstrates security awareness by providing explicit guidance to developers on how to handle untrusted user content and how to manage system-enforced guardrails and refusals when using Foundation Models.
Audit Metadata