What problem does it solve? Writing animations with ManimGL (Grant Sanderson's 3Blue1Brown engine) is error-prone because it differs significantly from Manim Community Edition in imports, CLI commands, class names, and camera APIs, and its interactive development workflow is poorly documented. ## Core Features & Use Cases - ManimGL-specific guidance: Covers InteractiveScene, ShowCreation, Tex with t2c color mapping, self.frame camera control, and fix_in_frame(), with a clear comparison table against Manim Community Edition. - Interactive development workflow: Explains the -se flag, checkpoint_paste(), and self.embed() for IPython-based iterative animation development. - 3D and camera control: Documents frame.reorient(), parametric surfaces, textured spheres, lighting, and 3D axes. - Use Case: A user asks to animate a transformer attention mechanism visualization; the skill provides working examples like attention pattern grids, softmax masking, and embedding vectors adapted from 3Blue1Brown's video code. ## Quick Start Ask the AI to create a ManimGL scene that animates a mathematical equation with colored terms using Tex and t2c, then render it with the manimgl CLI.