Top Agent Skills

Framework for state-of-the-art sentence, text, and image embeddings. Provides 5000+ pre-trained models for semantic similarity, clustering, and retrieval. Supports multilingual, domain-specific, and multimodal models. Use for generating embeddings for RAG, semantic search, or similarity tasks. Best for production embedding generation.

Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy. Perfect for understanding GPT architecture from scratch. Train on Shakespeare (CPU) or OpenWebText (multi-GPU).

Control LLM output with regex and grammars, guarantee valid JSON/XML/code generation, enforce structured formats, and build multi-step workflows with Guidance - Microsoft Research's constrained generation framework

Optimizes transformer attention with Flash Attention for 2-4x speedup and 10-20x memory reduction. Use when training/running transformers with long sequences (>512 tokens), encountering GPU memory issues with attention, or need faster inference. Supports PyTorch native SDPA, flash-attn library, H100 FP8, and sliding window attention.

Extract structured data from LLM responses with Pydantic validation, retry failed extractions automatically, parse complex JSON with type safety, and stream partial results with Instructor - battle-tested structured output library

Distributed training orchestration across clusters. Scales PyTorch/TensorFlow/HuggingFace from laptop to 1000s of nodes. Built-in hyperparameter tuning with Ray Tune, fault tolerance, elastic scaling. Use when training massive models across multiple machines or running distributed hyperparameter sweeps.

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.

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.

Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.

Extend context windows of transformer models using RoPE, YaRN, ALiBi, and position interpolation techniques. Use when processing long documents (32k-128k+ tokens), extending pre-trained models beyond original context limits, or implementing efficient positional encodings. Covers rotary embeddings, attention biases, interpolation methods, and extrapolation strategies for LLMs.

Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace. Use when training large-scale models with limited compute (5× cost reduction vs dense models), implementing sparse architectures like Mixtral 8x7B or DeepSeek-V3, or scaling model capacity without proportional compute increase. Covers MoE architectures, routing mechanisms, load balancing, expert parallelism, and inference optimization.

Trains large language models (2B-462B parameters) using NVIDIA Megatron-Core with advanced parallelism strategies. Use when training models >1B parameters, need maximum GPU efficiency (47% MFU on H100), or require tensor/pipeline/sequence/context/expert parallelism. Production-ready framework used for Nemotron, LLaMA, DeepSeek.

Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.

Provides guidance for interpreting and manipulating neural network internals using nnsight with optional NDIF remote execution. Use when needing to run interpretability experiments on massive models (70B+) without local GPU resources, or when working with any PyTorch architecture.

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.

Accelerate LLM inference using speculative decoding, Medusa multiple heads, and lookahead decoding techniques. Use when optimizing inference speed (1.5-3.6× speedup), reducing latency for real-time applications, or deploying models with limited compute. Covers draft models, tree-based attention, Jacobi iteration, parallel token generation, and production deployment strategies.

Initiates a 6‑month trial of DHEA replacement in women with primary adrenal insufficiency (PAI) who have persistent low libido, depressive symptoms, or low energy despite otherwise optimized glucocorticoid and mineralocorticoid therapy. If the patient does not report a sustained beneficial effect after 6 months, discontinue DHEA.

Comprehensive toolkit for protein language models including ESM3 (generative multimodal protein design across sequence, structure, and function) and ESM C (efficient protein embeddings and representations). Use this skill when working with protein sequences, structures, or function prediction; designing novel proteins; generating protein embeddings; performing inverse folding; or conducting protein engineering tasks. Supports both local model usage and cloud-based Forge API for scalable inference.

Material-UI v7 component library patterns including sx prop styling, theme integration, responsive design, and MUI-specific hooks. Use when working with MUI components, styling with sx prop, theme customization, or MUI utilities.

This skill should be used for time series machine learning tasks including classification, regression, clustering, forecasting, anomaly detection, segmentation, and similarity search. Use when working with temporal data, sequential patterns, or time-indexed observations requiring specialized algorithms beyond standard ML approaches. Particularly suited for univariate and multivariate time series analysis with scikit-learn compatible APIs.

Quantum computing framework for building, simulating, optimizing, and executing quantum circuits. Use this skill when working with quantum algorithms, quantum circuit design, quantum simulation (noiseless or noisy), running on quantum hardware (Google, IonQ, AQT, Pasqal), circuit optimization and compilation, noise modeling and characterization, or quantum experiments and benchmarking (VQE, QAOA, QPE, randomized benchmarking).

Build complex AI systems with declarative programming, optimize prompts automatically, create modular RAG systems and agents with DSPy - Stanford NLP's framework for systematic LM programming

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.

Comprehensive spreadsheet creation, editing, and analysis with support for formulas, formatting, data analysis, and visualization. When Claude needs to work with spreadsheets (.xlsx, .xlsm, .csv, .tsv, etc) for: (1) Creating new spreadsheets with formulas and formatting, (2) Reading or analyzing data, (3) Modify existing spreadsheets while preserving formulas, (4) Data analysis and visualization in spreadsheets, or (5) Recalculating formulas

Open-source embedding database for AI applications. Store embeddings and metadata, perform vector and full-text search, filter by metadata. Simple 4-function API. Scales from notebooks to production clusters. Use for semantic search, RAG applications, or document retrieval. Best for local development and open-source projects.

Parse FCS (Flow Cytometry Standard) files v2.0-3.1. Extract events as NumPy arrays, read metadata/channels, convert to CSV/DataFrame, for flow cytometry data preprocessing.

Computational pathology toolkit for analyzing whole-slide images (WSI) and multiparametric imaging data. Use this skill when working with histopathology slides, H&E stained images, multiplex immunofluorescence (CODEX, Vectra), spatial proteomics, nucleus detection/segmentation, tissue graph construction, or training ML models on pathology data. Supports 160+ slide formats including Aperio SVS, NDPI, DICOM, OME-TIFF for digital pathology workflows.

Fast DataFrame library (Apache Arrow). Select, filter, group_by, joins, lazy evaluation, CSV/Parquet I/O, expression API, for high-performance data analysis workflows.

Fast structured generation and serving for LLMs with RadixAttention prefix caching. Use for JSON/regex outputs, constrained decoding, agentic workflows with tool calls, or when you need 5× faster inference than vLLM with prefix sharing. Powers 300,000+ GPUs at xAI, AMD, NVIDIA, and LinkedIn.

Comprehensive Python library for astronomy and astrophysics. This skill should be used when working with astronomical data including celestial coordinates, physical units, FITS files, cosmological calculations, time systems, tables, world coordinate systems (WCS), and astronomical data analysis. Use when tasks involve coordinate transformations, unit conversions, FITS file manipulation, cosmological distance calculations, time scale conversions, or astronomical data processing.

This skill should be used when working with LaminDB, an open-source data framework for biology that makes data queryable, traceable, reproducible, and FAIR. Use when managing biological datasets (scRNA-seq, spatial, flow cytometry, etc.), tracking computational workflows, curating and validating data with biological ontologies, building data lakehouses, or ensuring data lineage and reproducibility in biological research. Covers data management, annotation, ontologies (genes, cell types, diseases, tissues), schema validation, integrations with workflow managers (Nextflow, Snakemake) and MLOps platforms (W&B, MLflow), and deployment strategies.

Educational GPT implementation in ~300 lines. Reproduces GPT-2 (124M) on OpenWebText. Clean, hackable code for learning transformers. By Andrej Karpathy. Perfect for understanding GPT architecture from scratch. Train on Shakespeare (CPU) or OpenWebText (multi-GPU).

Generate spectrograms and feature-panel visualizations from audio with the songsee CLI.

Framework for computational fluid dynamics simulations using Python. Use when running fluid dynamics simulations including Navier-Stokes equations (2D/3D), shallow water equations, stratified flows, or when analyzing turbulence, vortex dynamics, or geophysical flows. Provides pseudospectral methods with FFT, HPC support, and comprehensive output analysis.

One iteration of an unattended development loop. Use when invoked by name.

Elite brainstorming specialist for facilitated ideation sessions. Use when the user asks to talk to Carson or requests the Brainstorming Specialist.

Trigger: PRD, product requirements, spec, problem framing, project scoping, roadmap. Scope: Creating and updating product requirement documents and framing problem scopes. Boundary: Excludes code implementation, design mocks, or CI/CD config.

Trigger: project task list, task decomposition, phase planning, master orchestration, delivery governance. Scope: Project sequencing, task checklist creation, operating governance. Boundary: Excludes tactical code debugging.

Trigger: a11y, accessibility, ARIA, screen reader, contrast, keyboard, tap targets, focus. Scope: Audit web interface accessibility compliance (WCAG AA). Boundary: Do not use for general backend logic or CSS styling aesthetics.

Design, implement, audit, and plan interface animation and motion systems. Use for animation reviews, motion audits, easing and duration decisions, gesture physics, transform origins, reduced-motion behavior, animation performance, SvelteKit or Motion integration, perceptual feel checks, or self-contained animation improvement plans.

Trigger: shopping cart, checkout conversion, checkout flow, payment trust, pricing layout. Scope: Commerce UX optimization, trust badges, checkout pathways. Boundary: Excludes payment gateway API integrations or backend authentication.

Trigger: design review, visual audit, layout inspection, QA checks, screenshot analysis. Scope: Performing design system compliance audits and interface QA. Boundary: Excludes functional unit testing or database indexing.

Audit a repository, vet and prioritize evidence-backed findings, write drift-aware implementation plans, delegate bounded work, and independently verify results. Use when asked to improve, audit, modernize, harden, review, plan fixes for, or systematically reduce risk in a codebase. Use native isolated agents for implementation; use OpenRouter/free only for narrow advisory audit slices under its safety limits.

Trigger: web security, XSS mitigation, secure cookies, CSRF, CSP headers, CORS policies. Scope: Frontend and browser-level security boundaries. Boundary: Excludes operating system file storage permissions.

Trigger: OAuth 2.1, JWT session, Secure cookies, Keychain storage, auth flow, user authentication. Scope: User authentication, session storage, authorization boundaries. Boundary: Excludes generic data caching.

Trigger: web performance, slow load, bundle size, lazy loading, LCP, CWV optimization, fetch priority. Scope: Frontend optimization, loading speeds, asset minimization. Boundary: Excludes backend database indexing or visual design systems.

Trigger: tech stack, preferred tools, SvelteKit, Bun, Drizzle, Zero Sync, Better Auth, Vite. Scope: Project tooling preferences, framework selections. Boundary: Excludes application logic or visual styling guidelines.

Trigger: database schema, indexing, cache asidePattern, Redis caching, ORM models, Drizzle schema. Scope: Database query tuning, indexing, and data cache designs. Boundary: Excludes user identity validation.

Design and implement clear, modular, resilient software and create executor-grade implementation plans with explicit scope, drift checks, verification gates, tests, STOP conditions, dependencies, and delegated-work review. Use for development, refactoring, architecture, implementation planning, safe migrations, handoffs to another engineer or agent, or verifying delegated code changes.

Trigger: TDD, test first, unit testing, integration tests, mock assertions, test-driven dev. Scope: Writing test assertions before coding features, testing paradigms. Boundary: Excludes UX aesthetics or continuous deployment settings.

Research the open web for enhancement opportunities (Layer 3 — Opus subagent with WebSearch + WebFetch). Topic-driven. Returns structured findings with hype tags + source URLs. Does NOT auto-implement.

Turn an English description (or mermaid source) into a diagram triplet: the source, an editable .excalidraw file you can open (gstack)

Auto-review pipeline — reads the full CEO, design, eng, and DX review skills from disk and runs them sequentially with auto-decisions using 6 decision principles. (gstack)

CEO/Dispatcher. Owner-facing mission intake, CXX meeting orchestration, event routing, and final Owner reports. Trigger with /goal, /submission, /hot-fix, /ceo.

Holistic product leader who owns the full product lifecycle — from discovery and strategy through roadmap, stakeholder alignment, go-to-market, and outcome measurement. Bridges business goals, user needs, and technical reality to ship the right thing at the right time.

Strategic thought-leadership book collaborator for founders, experts, and operators turning voice notes, fragments, and positioning into structured first-person chapters.

Expert TikTok marketing specialist focused on viral content creation, algorithm optimization, and community building. Masters TikTok's unique culture and features for brand growth.

Expert civil and structural engineer with global standards coverage — Eurocode, DIN, ACI, AISC, ASCE, AS/NZS, CSA, GB, IS, AIJ, and more. Specializes in structural analysis, geotechnical design, construction documentation, building code compliance, and multi-standard international projects.

Writes a structured standup update as a session.standup LogEntry — capturing what was done, what is in progress, what comes next, and any blockers. Use when the user says "write a standup", "standup update", "what did I do today", "write a status update for the team", or at the end of a session on projects that run daily standups.

Post-process AI-generated text through the unslop CLI to strip AI writing patterns before publishing