> browse/registry

> torchforge-rl-training

Provides guidance for PyTorch-native agentic RL using torchforge, Meta's library separating infra from algorithms. Use when you want clean RL abstractions, easy algorithm experimentation, or scalable training with Monarch and TorchTitan.

> azure-document-intelligence

Expert knowledge for Azure AI Document Intelligence development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building, debugging, or optimizing Azure AI Document Intelligence applications. Not for Azure AI services (use azure-ai-services), Azure AI Search (use azure-cognitive-search), Azure AI Vision (use azure-ai-vision), Azure Machine Learning (use a

> azure-health-data-services

Expert knowledge for Azure Health Data Services development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building, debugging, or optimizing Azure Health Data Services applications. Not for Azure Health Bot (use azure-health-bot), Azure Data Factory (use azure-data-factory), Azure Data Explorer (use azure-data-explorer), Azure Machine Learning (use azur

> azure-local

Expert knowledge for Azure Local development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building, debugging, or optimizing Azure Local applications. Not for Azure AI Foundry Local (use azure-ai-foundry-local), Azure Foundry Classic (use azure-foundry-classic), Microsoft Foundry (use azure-microsoft-foundry), Azure Machine Learning (use azure-machine-

> azure-microsoft-foundry

Expert knowledge for Microsoft Foundry development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building, debugging, or optimizing Microsoft Foundry applications. Not for Azure AI Foundry Local (use azure-ai-foundry-local), Azure Foundry Classic (use azure-foundry-classic), Azure Machine Learning (use azure-machine-learning), Azure Databricks (use azur

> azure-personalizer

Expert knowledge for Azure AI Personalizer development including troubleshooting, decision making, limits & quotas, security, configuration, and integrations & coding patterns. Use when building, debugging, or optimizing Azure AI Personalizer applications. Not for Azure AI services (use azure-ai-services), Azure Machine Learning (use azure-machine-learning), Azure AI Search (use azure-cognitive-search), Azure AI Metrics Advisor (use azure-metrics-advisor).

> azure-quantum

Expert knowledge for Azure Quantum development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building, debugging, or optimizing Azure Quantum applications. Not for Azure HDInsight (use azure-hdinsight), Azure Databricks (use azure-databricks), Azure Synapse Analytics (use azure-synapse-analytics), Azure Machine Learning (use azure-machine-learning).

> huggingface-accelerate

Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.

> pennylane

Hardware-agnostic quantum ML framework with automatic differentiation. Use when training quantum circuits via gradients, building hybrid quantum-classical models, or needing device portability across IBM/Google/Rigetti/IonQ. Best for variational algorithms (VQE, QAOA), quantum neural networks, and integration with PyTorch/JAX/TensorFlow. For hardware-specific optimizations use qiskit (IBM) or cirq (Google); for open quantum systems use qutip.

> pytorch-lightning

High-level PyTorch framework with Trainer class, automatic distributed training (DDP/FSDP/DeepSpeed), callbacks system, and minimal boilerplate. Scales from laptop to supercomputer with same code. Use when you want clean training loops with built-in best practices.

> quantizing-models-bitsandbytes

Quantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction with minimal accuracy loss. Use when GPU memory is limited, need to fit larger models, or want faster inference. Supports INT8, NF4, FP4 formats, QLoRA training, and 8-bit optimizers. Works with HuggingFace Transformers.

> gptq

Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.

> hqq-quantization

Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.

> llama-cpp

Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.

> tensorrt-llm

Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.

> shap

Model interpretability and explainability using SHAP (SHapley Additive exPlanations). Use this skill when explaining machine learning model predictions, computing feature importance, generating SHAP plots (waterfall, beeswarm, bar, scatter, force, heatmap), debugging models, analyzing model bias or fairness, comparing models, or implementing explainable AI. Works with tree-based models (XGBoost, LightGBM, Random Forest), deep learning (TensorFlow, PyTorch), linear models, and any black-box model

> splitting-datasets

This skill enables Claude to split datasets into training, validation, and testing sets. It is useful when preparing data for machine learning model development. Use this skill when the user requests to split a dataset, create train-test splits, or needs data partitioning for model training. The skill is triggered by terms like "split dataset," "train-test split," "validation set," or "data partitioning."

> azure-anomaly-detector

Expert knowledge for Azure AI Anomaly Detector development including troubleshooting, best practices, architecture & design patterns, limits & quotas, configuration, and deployment. Use when building, debugging, or optimizing Azure AI Anomaly Detector applications. Not for Azure AI Metrics Advisor (use azure-metrics-advisor), Azure Monitor (use azure-monitor), Azure Machine Learning (use azure-machine-learning).

> outlines

Guarantee valid JSON/XML/code structure during generation, use Pydantic models for type-safe outputs, support local models (Transformers, vLLM), and maximize inference speed with Outlines - dottxt.ai's structured generation library

> unsloth

Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization