Expert integration patterns for Claude API and TypeScript SDK covering Messages API, streaming responses, tool use, error handling, token optimization, and production-ready implementations for building AI-powered applications
Guidance for counting tokens in datasets, particularly from HuggingFace or similar sources. This skill should be used when tasks involve counting tokens in datasets, understanding dataset schemas, filtering by categories/domains, or working with tokenizers. It helps avoid common pitfalls like incomplete field identification and ambiguous terminology interpretation.
Guidance for SAM-based cell segmentation and mask conversion tasks involving MobileSAM, mask-to-polygon conversion, CSV processing, and command-line interface design. This skill applies when working with Segment Anything Model (SAM) for biological image segmentation, converting binary masks to polygon coordinates, processing microscopy data, or building CLI tools that interface with deep learning models. (project)
Guidance for optimizing LLM inference request batching and scheduling problems. This skill applies when designing batch schedulers that minimize cost while meeting latency and padding constraints, involving trade-offs between batch count, shape selection, and padding ratios. Use when the task involves grouping requests by sequence lengths, managing shape compilation costs, or optimizing multi-objective scheduling with hard constraints.
Guidance for implementing minimal GPT-2 inference in constrained environments (code golf challenges). This skill should be used when implementing neural network inference from scratch, parsing binary checkpoint formats, implementing BPE tokenization, or working on code golf challenges involving ML models. Covers verification strategies and common pitfalls for checkpoint parsing and model inference.
Guidance for extracting weight matrices from black-box ReLU neural networks using only input-output queries. This skill applies when tasked with recovering internal parameters (weights, biases) of a neural network that can only be queried for outputs, particularly two-layer ReLU networks. Use this skill for model extraction, model stealing, or neural network reverse engineering tasks.
Guidance for extracting weight matrices from black-box ReLU neural networks using only input-output queries. This skill applies when tasks involve model extraction attacks, recovering hidden layer weights from neural networks, or reverse-engineering ReLU network parameters from query access.
Guidance for building and training with the Caffe deep learning framework on CIFAR-10 dataset. This skill applies when tasks involve compiling Caffe from source, training convolutional neural networks on image classification datasets, or working with legacy deep learning frameworks that have compatibility issues with modern systems.
Guidance for creating standalone CLI tools that perform neural network inference by extracting PyTorch model weights and reimplementing inference in C/C++. This skill applies when tasks involve converting PyTorch models to standalone executables, extracting model weights to portable formats (JSON), implementing neural network forward passes in C/C++, or creating CLI tools that load images and run inference without Python dependencies.
This skill provides guidance for retrieving and verifying information from dynamic ML leaderboards (MTEB, Scandinavian Embedding Benchmark, HuggingFace leaderboards, etc.) with specific temporal requirements. It should be used when tasks involve finding top-performing models, rankings, or benchmark results as of a specific date, especially when the data source is frequently updated.
Guidance for finding the largest eigenvalue of small dense matrices with performance optimization. This skill should be used when tasks involve computing eigenvalues (especially the dominant/largest eigenvalue), optimizing numerical linear algebra routines, or improving performance of numpy/scipy matrix operations for small matrices (typically 2-10 in size).
This skill provides guidance for implementing PyTorch pipeline parallelism for distributed training of large language models. It should be used when implementing pipeline parallel training loops, partitioning transformer models across GPUs, or working with AFAB (All-Forward-All-Backward) scheduling patterns. The skill covers model partitioning, inter-rank communication, gradient flow management, and common pitfalls in distributed training implementations.
This skill provides guidance for translating RStan (R-based Stan interface) code to PyStan (Python-based Stan interface). It should be used when converting Stan models from R to Python, migrating Bayesian inference workflows between languages, or adapting R data preparation logic to Python equivalents.
Comprehensive MLOps workflows for the complete ML lifecycle - experiment tracking, model registry, deployment patterns, monitoring, A/B testing, and production best practices with MLflow
Guidance for converting R-Stan (RStan) code to Python-Stan (PyStan). This skill applies when translating Stan models and inference code from R to Python, including API mapping between RStan and PyStan 3.x, hyperparameter translation, and handling differences in output formats. Use this skill for statistical model migration, Bayesian inference code conversion, or when working with Stan models across R and Python ecosystems.
This skill should be used when reconstructing PyTorch models from weight files (state dictionaries), checkpoint files, or partial model artifacts. It applies when the agent needs to infer model architecture from saved weights, rebuild models without original source code, or recover models from corrupted/incomplete saves. Use this skill for tasks involving torch.load, state_dict reconstruction, architecture inference, or model recovery in CPU-constrained environments.
Guidance for recovering PyTorch model architectures from state dictionaries, retraining specific layers, and saving models in TorchScript format. This skill should be used when tasks involve reconstructing model architectures from saved weights, fine-tuning specific layers while freezing others, or converting models to TorchScript format.
Creates summary for human review - and comprehensive implementation plan for LLM, complete code examples, and verification steps
Creates Skills for Claude. Use when users request creating/updating skills, need skill structure guidance, or mention extending Claude's capabilities through custom skills.
Install skills from github.com/oaustegard/claude-skills into /mnt/skills/user. Use when user mentions "install skills", "load skills", "add skills", "update skills", "refresh skills", or references a skill not currently installed.
Internal skill. Use cc10x-router for all development tasks.
Walk-forward analysis for trading strategies. Tests whether optimized parameters hold up on out-of-sample data. Use when checking parameter robustness, detecting potential overfitting, or validating a backtest.
Manages Database Seeders with advanced support for JSON data sources, idempotency checks, and relationship mapping.
Use when starting any conversation - establishes mandatory workflows for finding and using skills, including using Skill tool before announcing usage, following brainstorming before coding, and creating TodoWrite todos for checklists
