
This is a highly opinionated architectural guide for ML projects that use factory and registry patterns. It's built around a specific template structure with Hydra configs, and it's very explicit about when to invoke it: only when you're creating new registrable components with decorators like @register_model or @register_dataset. The skill includes concrete patterns for auto-importing modules, config-driven model initialization (models take only a cfg parameter, never individual args), and a standardized directory layout separating data_module, model_module, and trainer_module. It's useful if you're working in this exact architectural style, but it's narrow by design. The "when not to use" section is longer than most skills, which tells you everything about its intended scope.
$ npx -y skills add galaxy-dawn/claude-scholar --skill architecture-design --agent claude-codeInstalls into .claude/skills of the current project.
This skill defines the standard code architecture for machine learning projects based on the template structure. When modifying or extending code, follow these patterns to maintain consistency.
The project follows a modular, extensible architecture with clear separation of concerns. Each module (data, model, trainer, analysis) is independently organized using factory and registry patterns for maximum flexibility.
Use this skill when:
@register_dataset@register_model__init__.py factory wiringDo not use this skill when:
Sponsors