apple-on-device-ai
Pass
Audited by Gen Agent Trust Hub on Mar 9, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves as a documentation and code-reference tool for Apple's AI ecosystem. No malicious patterns or security risks were identified in the analyzed files.
- [EXTERNAL_DOWNLOADS]: The skill references downloading open-source models from Hugging Face's 'mlx-community' and installing the 'coremltools' Python package. These are well-known technology services and official tools from Apple, falling under the trusted/well-known source category.
- [PROMPT_INJECTION]: The skill actually provides defensive guidance against prompt injection, specifically warning developers to avoid placing untrusted user content within system instructions and explaining how Apple's built-in guardrails function.
- [COMMAND_EXECUTION]: Code snippets demonstrate the use of standard development tools like 'pip' for package management and 'torch' for model tracing, which are consistent with the skill's stated purpose of model conversion and deployment.
Audit Metadata