Skills
98,067 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
The html skill owns creating and revising durable, standalone HTML artifacts: a complete self-contained index.html saved in a requested or generated artifact directory, with inline dependencies, accessible structure, loopback serving, and exact-file verification. It triggers when a user asks to create, build, turn content or data into, save, revise, or update a “durable HTML artifact,” “one-file dashboard,” “single-page local artifact,” or named HTML artifact, or explicitly says “use /html” or “activate the HTML skill.” It does not trigger for integrating HTML into React or Next.js routing, reviewing or explaining HTML or accessibility, browser-based QA of an existing page, converting Figma designs into reusable components, or merely serving an existing dist directory.
Implement hybrid search combining vector and keyword retrieval for RAG systems. Use this skill when building RAG retrieval, combining semantic search with BM25, implementing reciprocal rank fusion (RRF), or optimizing retrieval accuracy. Activate when: vector search, keyword search, BM25, semantic search, hybrid RAG, retrieval optimization, search relevance, reranking.
Implements, configures, and debugs Unity In-App Purchases, including subscriptions, receipt validation, and direct-to-customer payments via Stripe or Coda. Use when asked about real-money purchases or microtransactions, or migrating from native store billing, RevenueCat, Adapty, Essential Kit, or UniPay.
Mirror an iOS Simulator into the Codex in-app browser and render SwiftUI previews from importable Swift packages in that simulator with hot reload. Use when a user wants to watch or interact with an iOS app in the browser, see a SwiftUI preview outside Xcode Canvas, iterate live on a preview, or capture browser-visible simulator proof.
Select and use companion engineering tools; install, update, reconcile, configure, verify, or remove tools, installed skills, host instructions, and native agent declarations. Use for non-obvious capability selection, source retrieval or measurement semantics, and environment readiness; exclude routine reads, skill authoring, engineering verdicts, and implementation.
Java coding standards for Spring Boot and Quarkus services: naming, immutability, Optional usage, streams, exceptions, generics, CDI, reactive patterns, and project layout. Automatically applies framework-specific conventions. Use when writing or reviewing Java in a Spring Boot or Quarkus service.
Full lifecycle management of JeecgBoot Online forms (cgform) — automatically create/edit database tables and form configurations via API, supporting single tables, master-detail tables, tree tables, 26 control types, plus JS/Java/SQL enhancements, permission configuration, data CRUD, and Jimu report integration. Use this skill whenever the user's intent involves "Online forms", including but not limited to: creating or configuring database tables ("create a leave request table", "create an online table", "make a table with a dropdown", "low-code form", "online form", "configure a table"); modifying existing Online table fields ("add a field", "change a field type", "add a subtable", "delete a field"); configuring form enhancements ("JS enhancement", "custom button", "form linkage", "Java enhancement", "SQL enhancement"); configuring permissions ("field permission", "button permission", "data permission", "grant to role"); managing form data ("insert data", "query records", "export CSV", "generate test data"); and linking Jimu reports ("add a report to this table", "integrated printing"). Even if the user only describes business requirements without saying "online" (e.g. "build an employee information management feature with name, department dropdown, and hire date"), as long as it involves metadata-driven form configuration, this skill should be triggered. Note: do not confuse with the "designer form" (desform) — desform is a drag-and-drop form designer handled by the jeecg-desform skill; do not confuse with "Online report" (cgreport) or "Online chart" (onlchart) either — they are SQL-driven read-only displays.
Create or migrate to a justfile (casey/just command runner) for any project. Use this skill when the user asks to add a justfile, replace a Makefile, set up project commands, create task runners, or mentions "just" in the context of build/dev workflows. Also trigger when you see a project with a Makefile that would benefit from just's simpler syntax, or when setting up a new project that needs common dev commands (build, test, lint, fmt). Covers Rust, Python, TypeScript/JavaScript, Go, and Ruby ecosystems. Do NOT use for CI pipeline configuration, Dockerfiles, or actual build system setup (cargo, webpack, etc.).
Review code following Kent Beck's refactoring philosophy and simple design principles. Detects code smells, guides refactoring techniques, and promotes incremental improvement (YAGNI, KISS). Use when reviewing code quality, refactoring, or improving code readability.
A professional knowledge base management system designed to address the curse of knowledge and cognitive bias. It builds a structured Markdown knowledge base by making tacit knowledge explicit, scanning code to extract domain concepts, and integrating industry best practices.
Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies
Migrate KMP projects from CocoaPods (kotlin("native.cocoapods")) to Swift Package Manager (swiftPMDependencies DSL)—replaces pod() with swiftPackage(), transforms cocoapods.* imports to swiftPMImport.*, and reconfigures the Xcode project.
Framework for building LLM-powered applications with agents, chains, and RAG. Supports multiple providers (OpenAI, Anthropic, Google), 500+ integrations, ReAct agents, tool calling, memory management, and vector store retrieval. Use for building chatbots, question-answering systems, autonomous agents, or RAG applications. Best for rapid prototyping and production deployments.
Invoke this skill when you need human-in-the-loop approval, custom middleware, or structured output. Covers HumanInTheLoopMiddleware for human approval of dangerous tool calls, creating custom middleware with hooks, Command resume patterns, and structured output with Pydantic/Zod.
Build and configure Laravel 10+ applications, including creating Eloquent models and relationships, implementing Sanctum authentication, configuring Horizon queues, designing RESTful APIs with API resources, and building reactive interfaces with Livewire. Use when creating Laravel models, setting up queue workers, implementing Sanctum auth flows, building Livewire components, optimising Eloquent queries, or writing Pest/PHPUnit tests for Laravel features.
Integrates the LevelPlay ad mediation SDK via the Ads Mediation package. Use when the user asks about adding rewarded, interstitial, or banner ads, mediation or ad privacy setup, Android or iOS ad dependency build failures, or migrating from Unity Ads or IronSource APIs.
Build and maintain software architecture as code with LikeC4 DSL. Use when tasks mention architecture diagrams, C4/context/container/component views, system landscape, dependency/integration maps, or when the user wants architecture generated/evolved from code. Apply for creating new `.c4`/`.likec4` models, updating existing models, validating with LikeC4 CLI, and preparing preview/build/export outputs.
Links a PostHog error occurrence or session's errors to OpenTelemetry traces, logs with the same trace ID, and Session Replay. Use when the user asks for trace or span IDs on an error, same-trace logs, or backend traces for errors in a session. Starts from an occurrence UUID, session ID, or error tracking issue URL. For a general issue diagnosis or replay summary, use investigating-error-issue or investigating-replay.
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.
Use when the user wants to query Lore git trailers from commit history — surface constraints, rejected alternatives, directives, test gaps, and other decision context. Triggers on "최근 제약조건 확인", "리젝된 대안 보여줘", "check recent constraints", "show directives", or explicit skill invocation (/lore-query in Claude Code, $lore-query in Codex CLI, /skill:lore-query in Kimi Code, lore-query in Qwen Code).
Explains how to use Mamba selective state-space models (state-spaces/mamba package, Mamba-1 with d_state=16 and Mamba-2 with multi-head structure and d_state=128) for linear-time sequence modeling. Covers installation, the Mamba block, building a language model with MambaLMHeadModel, loading pretrained state-spaces checkpoints (130M to 2.8B) from HuggingFace, and benchmarking against Transformers. Use when implementing or loading Mamba models, processing very long sequences without a KV cache, building streaming applications, choosing between Mamba-1 and Mamba-2, or fixing install and CUDA memory problems. Requires Linux and an NVIDIA GPU. Do not use for standard Transformer models, or for RWKV, RetNet, or Hyena architectures.
Use when designing DETERMINISTIC algorithms, analyzing time/space complexity, selecting numerical methods (ODEs, optimization, linear algebra), or verifying mathematical correctness. NOT for statistical/probabilistic method selection (use statistician) or physics-based engineering feasibility (use calculator).
Deploy AI models to embedded hardware using MathWorks tools (MATLAB, Simulink, Embedded Coder). Covers two workflow patterns: (1) MathWorks-native or imported models rebuilt as dlnetwork for lean hardware, (2) direct C/C++ code generation from PyTorch and LiteRT models. Both patterns support all targets (Cortex-M/A/R, x86, GPU). Trigger when: user wants to deploy AI to embedded targets; generate C/CUDA from neural networks; compress AI models for MCU; integrate AI in Simulink for system-level simulation; import PyTorch/ONNX/TensorFlow models for embedded deployment; optimize AI for resource-constrained hardware; or use loadPyTorchExportedProgram, loadLiteRTModel, importNetworkFromPyTorch, importNetworkFromONNX, importNetworkFromTensorFlow, importNetworkFromKeras, dlquantizer, exportNetworkToSimulink, or Embedded Coder with AI models.
Designs, reviews, and migrates MATLAB R2026a and GNU Octave numerical code, covering functions with arguments blocks, arrays and indexing, tables and timetables, matlab.unittest tests, MATLAB Projects, exportgraphics figures, MAT file inventory, and MATLAB-Python interoperability. Includes static helper scripts that scan .m files, plan argv commands, and hash artifacts without executing anything. Use when writing or reviewing .m code, migrating code between MATLAB releases or to Octave, setting up reproducible tests and projects, inspecting an untrusted MAT or .mlx file safely, or checking Python Engine compatibility. Do not use for pure Python numerical work with NumPy or SciPy.
.NET MAUI app lifecycle guidance covering the four app states (not running, running, deactivated, stopped), cross-platform Window lifecycle events, backgrounding and resume behaviour, platform-specific lifecycle mapping for Android and iOS/Mac Catalyst, and state-preservation patterns. USE FOR: "app lifecycle", "OnStart", "OnSleep", "OnResume", "backgrounding", "app state", "window lifecycle", "save state on background", "resume app", "deactivated event", "lifecycle events". DO NOT USE FOR: navigation events (use maui-shell-navigation), dependency injection setup (use maui-dependency-injection), or platform APIs (use maui-platform-invoke).
Check Module Federation build configuration: verify correct MF plugin for the bundler, async entry configuration, exposes key format, and exposes path existence. Use when modules cannot be correctly exposed, expected modules are missing from the build output, or runtime initialization errors occur.
Diagnose Module Federation type issues across three categories: (1) producer type file generation failures (TYPE-001), (2) consumer failing to pull remote types, (3) tsconfig not configured to consume remote types. Use when the user encounters TS type errors, missing @mf-types folder, or cannot consume remote module types.
Migrates Next.js projects to vinext (Vite-based Next.js reimplementation). Load when asked to migrate, convert, or switch from Next.js to vinext. Handles compatibility scanning, package replacement, Vite config generation, ESM conversion, and deployment setup (Cloudflare Workers natively, other platforms via Nitro).
Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.
Authoritative guidance for modern React 19+ (Actions, use, Compiler, View Transitions, Fragment refs, Activity, browser, useEffectEvent). Use when writing, reviewing, refactoring, or migrating React components, forms, data fetching, concurrent UI, or upgrading to React 19+. Triggers on React, React 19, useActionState, useOptimistic, forwardRef, useEffect data fetch, React Compiler, ViewTransition, Suspense patterns, or codemods. Always prefer latest official patterns over training data.
General performance optimization techniques for Motoko. Reducing allocations, efficient Text building, fixed-width arithmetic, block processing, async patterns, and more. Load when you need to improve hot paths or reduce overhead without changing behavior.
Expert MVVM decisions for iOS/tvOS: choosing between ViewModel patterns (state enum vs published properties vs Combine), service layer boundaries, dependency injection strategies, and testing approaches. Use when designing ViewModel architecture, debugging data flow issues, or deciding where business logic belongs. Trigger keywords: MVVM, ViewModel, ObservableObject, @StateObject, service layer, dependency injection, unit test, mock, architecture
Provides nanoGPT, Karpathy's minimal PyTorch GPT implementation (model.py and train.py), with workflows for training character-level Shakespeare on CPU, reproducing GPT-2 124M on OpenWebText with multi-GPU torchrun, fine-tuning pretrained GPT-2 checkpoints, and training on custom text. Use when learning how GPT and transformer blocks work from scratch. Use when running a small training experiment on CPU or a single GPU. Use when modifying a transformer variant in plain PyTorch. Use when preparing character-level or BPE binary datasets. Use when troubleshooting out-of-memory or slow training in nanoGPT. Not for production or large-scale distributed training; use HuggingFace Transformers or Megatron-LM instead.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.