Skills
97,222 packaged capabilities and playbooks for agents. Install any of them with one command.
Development
Finds and finishes the last 20% of a built solution — the experiential layer agents skip: what the output actually reads like, which pages or screens should exist, first-run and empty states, opinionated defaults, the golden path. Use when planning any build so the magic gets specced as concretely as the plumbing, before calling any build "done", when a solution works but feels like a demo, or when the user says "last 20%", "finish this", "make it actually good", "is this actually done", or "why does this feel flat".
Choosing what form your data needs for MapLibre GL JS — GeoJSON read straight from a file versus vector or raster tiles, the feature-count and file-size thresholds where each stops working, and how a basemap and your own data combine as separate sources. Use when deciding which source type a dataset needs, not when hosting or configuring one
Apply when building catalog or SKU synchronization logic for VTEX marketplace seller connectors. Covers the changenotification endpoint, SKU suggestion lifecycle, product data mapping, price and inventory sync, and fulfillment simulation. Use for implementing seller-side catalog integration that pushes SKUs to VTEX marketplaces with proper notification handling and rate-limited batch synchronization.
Test-driven development skill for writing unit tests, generating test fixtures and mocks, analyzing coverage gaps, and guiding red-green-refactor workflows across Jest, Pytest, JUnit, Vitest, and Mocha. Use when the user asks to write tests, improve test coverage, practice TDD, generate mocks or stubs, or mentions testing frameworks like Jest, pytest, or JUnit.
Implement the RevenueCat purchase and restore flow. Use when the user asks to buy a package, purchase a subscription, fetch offerings, build the purchase logic behind a custom (hand-coded) paywall, handle purchase errors, detect user cancelled, or restore previous purchases on iOS, Android, Kotlin Multiplatform, Flutter, or React Native.
Integrate the Sumsub Mobile SDK (IdensicMobileSDK / SNSMobileSDK) into a native iOS or Android app—even when the user doesn't name the SDK (e.g. “add Sumsub/KYC to my iPhone app”, “add KYC to my Android app”, “show the verification screen”, “launch Sumsub from a view controller / activity”). Detects the platform from the project (iOS Xcode/Swift vs Android Gradle/Kotlin) and follows the matching track. Covers install, permissions, init, token refresh, and presenting / launching the flow. SKIP for web (`sumsub-integrate-websdk`), backend / API-only token signing (`sumsub-api-generic`), theming (`sumsub-theme-msdk`), and React Native / Flutter (not supported—native iOS & Android only)
UdonSharp scripting skill for VRChat SDK 3.10.5 (active and verified target). Use when writing, reviewing, debugging, or migrating UdonSharp C# and UdonBehaviour code. Positive triggers include UdonSharp, NetworkCallable, NetworkCalling, CallingPlayer, Udon network authorization, synced runtime state, a local public helper, public-method audit, C# to Udon conversion, and UdonSharp Assembly Version Defines. VRCQualitySettings and VRCTween calls, PhysBone/Contact callbacks, world VRCPhysBoneCollider runtime access, persistence, collection, web, and other component APIs trigger this skill when the request is about Udon, C#, or runtime API access. Excludes scene setup, component setup, Build Panel work, layers, optimization, and upload; route those requests to unity-vrc-world-sdk-3.
Searches and recovers Claude Code JSONL history across all active config homes and archives registered in ~/.claude/history-sources.json. Use --all-projects when the project is unknown, --codex to include Codex rollout search, and --kimi to include Kimi CLI (kimi-code) session search. Uses internal timestamps and searches messages, thinking, tool inputs/results, queues, attachments, summaries, titles, and file-history paths. Recovers exact captured bytes from Claude file-history snapshots, including post-Write edits and binary files; otherwise labels Write checkpoints as lower fidelity. Use for keyword/date-bounded history search, prior-conversation forensics, deleted-file recovery, vanished ~/.claude/jobs artifacts, tool/file-operation analysis, or requests mentioning session history, find in history, previous conversation, or .claude/projects. For a recent Claude+Codex+Kimi CLI inventory, use local-conversation-history instead
CodeStable is an AI coding workflow for serious software engineering that keeps humans in the loop. Rather than centering on Agent orchestration, it organizes requirements, architecture, features, issues, and historical decisions to make development driven by Codex and Claude controllable, traceable, and continuously evolvable.
Apply when implementing order integration hooks, feeds, or webhook handlers for VTEX marketplace connectors. Covers Feed v3 (pull) vs Hook (push), filter types (FromWorkflow and FromOrders), order status lifecycle, payload validation, and idempotent processing. Use for building order integrations between VTEX marketplaces and external systems such as ERPs, WMS, or fulfillment services.
CodeStable is an AI coding workflow for serious software engineering that keeps humans in the loop. Rather than centering on Agent orchestration, it organizes requirements, architecture, features, issues, and historical decisions to make development driven by Codex and Claude controllable, traceable, and continuously evolvable.
Build RAG (Retrieval Augmented Generation) pipelines with web search and LLMs. Tools: Tavily Search, Exa Search, Exa Answer, Claude, GPT-4, Gemini via OpenRouter. Capabilities: research, fact-checking, grounded responses, knowledge retrieval. Use for: AI agents, research assistants, fact-checkers, knowledge bases. Triggers: rag, retrieval augmented generation, grounded ai, search and answer, research agent, fact checking, knowledge retrieval, ai research, search + llm, web grounded, perplexity alternative, ai with sources, citation, research pipeline
Website replication/cloning methodology. USE WHEN the user says to replicate or clone a website, clone website, copy a site, imitate a site, make one based on this site, reproduce a site, recreate a webpage effect, download this site and adapt it, or replicate an interaction/WebGL/Canvas/Three.js effect. Provides a portable decision tree for “get the actual source first → determine the path → reverse-engineer and break it down → set up the project → replace the content,” covering three branches: static sites, React/Vue/Next content sites, and WebGL-Canvas-heavy frontend sites. Requires checking any executable code in secondhand AI analysis.
Autonomous multi-page site builder using a baton-passing loop. Each iteration reads a task from .design/next-prompt.md, generates a page in HTML/Tailwind, integrates it into the site, verifies visually, then writes the next task to keep the loop alive. Use whenever the user asks to build an entire site autonomously, build all pages of a site, generate multiple pages in sequence, or run a 'design loop' / 'baton loop' / 'autonomous site build' — even if they say 'just keep going' or 'build the next page' or 'next page' mid-flow.
Use when building, refactoring, or documenting Graft apps and proxies, including when asked to create a tool server, API server, dual-protocol server, or MCP-HTTP bridge. Graft's core thesis: define tools once and serve them as both HTTP REST endpoints and MCP tools from the same server, with discovery, docs, and OpenAPI generated automatically. Covers concrete actions such as defining tools and handlers, configuring authentication middleware, setting up HTTP and stdio transports, generating OpenAPI documentation, wrapping existing APIs via proxy mode, and wiring up the full CLI workflow
Add AgentMail tools to agent frameworks with the TypeScript or Python AgentMail Toolkit. Use for Vercel AI SDK, LangChain, OpenAI Agents SDK, LiveKit Agents, or MCP adapters; do not use for direct mailbox operations, raw SDK implementation, CLI usage, or MCP client setup.
Debug LLM applications using the Phoenix CLI. Fetch traces, analyze errors, structure trace review with open coding and axial coding, inspect datasets, review experiments, query annotation configs, and use the GraphQL API. Use whenever the user is analyzing traces or spans, investigating LLM/agent failures, deciding what to do after instrumenting an app, building failure taxonomies, choosing what evals to write, or asking "what's going wrong", "what kinds of mistakes", or "where do I focus" — even without naming a technique.
You are a code education expert specializing in explaining complex code through clear narratives, visual diagrams, and step-by-step breakdowns. Transform difficult concepts into understandable explanations for developers at all levels
Apply when implementing fulfillment, invoice, or tracking logic for VTEX marketplace seller connectors. Covers the External Seller fulfillment protocol: fulfillment simulation (checkout and indexation), order placement with reservation id, order dispatch (authorize fulfillment), OMS invoice and tracking APIs, and partial invoicing. Use for seller-side services that must answer within the simulation SLA and integrate with VTEX marketplace order management.
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.
World-class prompt engineering skill for LLM optimization, prompt patterns, structured outputs, and AI product development. Expertise in Claude, GPT-4, prompt design patterns, few-shot learning, chain-of-thought, and AI evaluation. Includes RAG optimization, agent design, and LLM system architecture. Use when building AI products, optimizing LLM performance, designing agentic systems, or implementing advanced prompting techniques.
Diagnoses and fixes slow Kotlin/Native compilation and linking in Kotlin Multiplatform projects that target iOS. Use when the user reports slow iOS or shared-framework builds, long linkDebug*/linkRelease* or XCFramework tasks, cold CI builds that re-download the Kotlin/Native toolchain, KSP or other generated code on the native path, transitiveExport usage, or asks for a local-development versus CI build performance plan.
Implement Terraform provider configuration and authentication with the Plugin Framework: provider schema for credentials (Optional + Sensitive attributes), environment variable fallbacks, credential provider chains (static config, then environment variables, shared credentials file, and platform identity), unknown-value guards in Configure(), secret redaction, configure-time credential validation, and diagnostics that name every source tried. Use when implementing or reviewing a provider's Configure method or provider schema, adding authentication options (API keys, tokens, profiles, credentials files, assume-role), deciding how a provider should resolve credentials, debugging "no valid credential sources" or missing-credentials errors, or unit testing credential resolution.
Export Webflow Designer components to React/Next.js code for external projects. Configure devlink settings in webflow.json, sync design updates with devlink sync, validate generated code, show diffs, and provide integration examples. Use when building with Webflow designs in React/Next.js.
Single entry point for all AEM as a Cloud Service Workflow skills. Covers workflow model design, custom process step and participant chooser development, launcher configuration, workflow triggering, and production support including debugging stuck or failed workflows, triaging incidents with Cloud Manager logs, thread pool analysis, and Sling Job diagnostics for the Granite Workflow Engine.
Looks up implementation details in the latest Cloudinary docs via the relevant llms.txt file. Use when building code or answering questions about image or video uploads, management, SDKs, APIs, webhooks, or integrations. Prefer a specialized Cloudinary skill for topics it covers; use this skill alongside it when the full use case requires capabilities beyond its scope
Consolidates redundant documentation while preserving all valuable content; keeps a doc set truthful after things change. Modes: (1) POST-CHANGE GOVERNANCE — when code, scripts, config, env vars, ports, paths, deployment, auth, tests, or a documented procedure changed; also "which docs are now wrong", "check docs for inconsistencies", "过时的命令/路径", "文档同步". Use it even when the user mentions only the change and never the docs — updating them is part of the change, not optional cleanup. (2) CONSOLIDATION — "clean up docs", "documentation bloat", "reduce documentation sprawl", "consolidate documentation", "too many doc files", "merge redundant docs", "merge these docs", "文档太多/太乱", multiple files covering the same topic, or documentation exceeding 500 lines across multiple files on similar topics. Finds stale copies of a fact changed in only one place, values that should never have been persisted (counts, totals, "N items", restated summaries), rival definitions that drift apart, and orphaned cross-references.
The HyperFrames composition contract — build one renderable project. Use for composition structure, the `data-*` timing attributes, `class="clip"`, tracks, sub-compositions, variables, framework-owned media playback, deterministic-render rules, and validation. Read before writing composition HTML.
Builds terminal chat UIs with @assistant-ui/react-ink and ANSI markdown with @assistant-ui/react-ink-markdown. Use when scaffolding `create --ink`, mounting `AssistantRuntimeProvider` around `useChatRuntime`, composing `ThreadPrimitive`, `ComposerPrimitive`, `MessagePrimitive`, `LoadingPrimitive`, `TextInput`, thread history, attachments, notifications, and terminal keyboard navigation, or when a chat needs grapheme safe editing, multiline display columns, an absolute backend URL, local file storage, or a custom `RemoteThreadListAdapter`. Route terminal markdown, tool output, message rendering, and Ink focus problems here. For browser application setup, web elements, and the standard web runtime, use [setup](../setup/SKILL.md).
Skyline custom routing and page transition skill. Covers custom route animations (routeBuilder), preset routes (7 wx:// types), page back gestures, container transition animations (open-container), and the Router API. For implementing transitions such as half-screen dialogs, page zoom, bottom pop-ups, and card expansion. Trigger keywords: custom routing, custom-route, routeBuilder, navigateTo, page transitions, half-screen, preset routes, back gestures, open-container
Guides development of Fusion portal shells — scaffolding, module configuration, app loading, routing, header/context integration, analytics, and deployment using the Fusion Framework CLI portal commands. USE FOR: create portal, scaffold portal, configure portal modules, portal app loading, portal routing, portal header, context selector, portal analytics, portal telemetry, portal manifest, ffc portal dev, portal deployment, embed apps in portal. DO NOT USE FOR: app-level feature development (use fusion-app-react-dev), backend service changes, Fusion Help Center integration, skill authoring.
VRChat World SDK 3 guide for scene and Inspector setup, component placement, optimization, and upload. Use for VRChat world scene configuration, VRC SDK components, layers, baked lighting, Quest/Android performance, Dynamics for Worlds, Build Panel warning triage, validation, and upload. Covers WorldQualitySettings, VRC_SceneDescriptor, VRC_Pickup, VRC_Station, VRC_Mirror, VRC_ObjectSync, VRC_CameraDolly, spawn points, collision matrices, PhysBone and Contact component placement, Box Contacts, Global Avatar PhysBone Colliders, and VRCPhysBoneCollider component setup. Active support / last verified: SDK 3.10.5. Triggers on: VRChat world scene, VRC SDK, scene setup, component placement, optimization, Quest support, light baking, upload, SDK validation, Build Panel warning, Auto Fix, red warning, yellow warning, or white warning. Do not use for UdonSharp C# or VRCTween calls; use unity-vrc-udon-sharp for runtime scripting. UdonSharp package, asmdef, and Version Defines work also belongs to that skill.
Remediation only — repair web accessibility (a11y) violations against WCAG 2.2 with a baseline, edit, and verify loop. Takes a target (URL, files, directory) or a findings worklist from `accessibility-scan`/`accessibility-inspect`/`accessibility-audit`, applies mechanical fixes as given, leaves TODOs for visual or contextual judgment, and verifies by re-running the baseline check. It only fixes. To find issues use `accessibility-scan` (one page, automated), `accessibility-inspect` (one page, manual), or `accessibility-audit` (whole site, WCAG-EM); to check for regressions use `accessibility-diff`. Use it for 'fix the a11y issues in X', 'make this accessible', 'add missing alt text and labels', 'apply these accessibility fixes', 'remediate these violations'.
Dodo Payments checkout via hosted Checkout Sessions (client.checkoutSessions.create), static payment links, and overlay or inline checkout for one-time and subscription products. Use when building a pay button, checkout page, trial signup, custom fields, discount codes at checkout, or return_url redirects; use subscription-integration for post-checkout lifecycle.
Use when creating a new OpenCLI adapter from scratch, adding support for a new website or platform, exploring a site's API endpoints via browser DevTools, or when a user asks to automatically generate a CLI for a website (e.g. "帮我生成 xxx.com 的 cli"). Covers automated generation, API discovery workflow, authentication strategy selection, TS adapter writing, and testing.
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.