Visualize training metrics, debug models with histograms, compare experiments, visualize model graphs, and profile performance with TensorBoard - Google's ML visualization toolkit
Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF,...
Best practices and workflows for the Spotify Web API. Use when building apps that interact with Spotify to: (1) Handle rate limits and 429 errors, (2) Implement efficient polling strategies, (3)...
Expert guidance for Fully Sharded Data Parallel training with PyTorch FSDP - parameter sharding, mixed precision, CPU offloading, FSDP2
Brief description of what this skill does and when to use it.
Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable...
Sentry error monitoring and performance tracing patterns for Next.js applications.
Guides database migration projects including engine changes (MySQL to PostgreSQL, Oracle to PostgreSQL, SQL Server to PostgreSQL), version upgrades, cloud migrations (on-premise to RDS/Cloud...
Designs comprehensive software solution architectures including system components, technology stacks, integration patterns, scalability strategies, and deployment models. Produces architecture...
Extract subtitles/transcripts from YouTube videos. Triggers: "youtube transcript", "extract subtitles", "video captions", "视频字幕", "字幕提取", "YouTube转文字", "提取字幕".
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or...
Evaluates LLMs across 100+ benchmarks from 18+ harnesses (MMLU, HumanEval, GSM8K, safety, VLM) with multi-backend execution. Use when needing scalable evaluation on local Docker, Slurm HPC, or...
Leverage OpenAI Codex/GPT models for autonomous code implementation. Triggers: "codex", "use gpt", "gpt-5", "gpt-5.2", "let openai", "full-auto", "用codex", "让gpt实现".
Replace with description of the skill and when Claude should use it.
Comprehensive React and Vite SPA performance, architecture, and deployment guidelines. Use this skill when building, reviewing, or refactoring React applications built with Vite (SPA). Covers...
Provides comprehensive KeyCloak administration guidance including realm management, user/group administration, client configuration, authentication flows, identity brokering, authorization...
Interactive feature development workflow from idea to implementation. Creates requirements (EARS format), design documents, and task lists. Triggers: "kiro", ".kiro/specs/", "feature spec",...
Control WezTerm terminal emulator via CLI. Manage panes, tabs, workspaces, and execute commands in running terminals.
Refactor CLAUDE.md files to follow progressive disclosure principles. Use when CLAUDE.md is too long or disorganized.
Provides guidance for enterprise-grade RL training using miles, a production-ready fork of slime. Use when training large MoE models with FP8/INT4, needing train-inference alignment, or requiring...