model-authoring
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFE
Full Analysis
- Official Resource References: The skill directs users to official documentation and repositories, such as the coreai-models repository on GitHub. These references are provided to support the implementation of high-performance model primitives and are standard for specialized hardware development.
- Standard Tooling Integration: The instructions mention standard developer tools for Apple platforms, including the use of
xcrunfor model compilation. This is a routine part of the build process for creating on-device machine learning assets. - Architecture Optimization Patterns: The skill provides specific PyTorch implementation patterns, such as BC1S layout and Conv2d projections, which are technical requirements for energy-efficient execution on the Neural Engine. These are well-documented hardware-specific optimization techniques.
- Library Usage: The skill includes code snippets that interact with widely recognized libraries like PyTorch and HuggingFace Transformers. These are used within their intended scope for model configuration and weight management.
Audit Metadata