Skill UI

Enables Claude to configure neural network architectures through the neural-network-builder plugin, creating or modifying models and defining layers, activation functions, and training parameters whenever network design or tuning is requested.

PyTorch Geometric provides tooling to build and train graph neural networks over irregular, multi-type graph data, covering dataset loading, message passing layers, mini-batching, and multi-GPU support for social, citation, molecular, and geometric learning workloads.

TorchDrug offers PyTorch-native graph neural networks, datasets, and tasks tailored to drug discovery, protein modeling, knowledge graphs, molecular generation, and retrosynthesis, letting researchers define custom GNN architectures for property prediction and synthesis planning.

Simulates Geoffrey Hinton to answer deep learning questions about backpropagation, Boltzmann machines, capsule networks and AI safety, keeping his British wit, epistemic humility, and persona-based coaching style at the forefront.

Guide federal agencies through implementing CISA Zero Trust Maturity Model v2.0 across identity, devices, networks, applications, and data, covering assessment, gap analysis, and progressive automation to reach optimal maturity.