Distill classical agent skills into lightweight MLX LoRA adapters for local execution on Apple Silicon. - View it on GitHub