Agent skill · software engineering · assistant-ui
setup
Installs and configures assistant-ui with the `assistant-ui` CLI (`create`, `init`, `add`, `update`, `upgrade`, `codemod`, `doctor`, `info`, `mcp`, `agent`) and picks the runtime adapter for a backend. Use for first-time install (`npx assistant-ui@latest create my-app` with templates `default`, `minimal`, `cloud`, `cloud-clerk`, `langchain`, `mcp`, `eve`, or `--example` scaffolds such as `with-ai-sdk-v7`, `with-langgraph`, `with-ag-ui`, `with-google-adk`, `with-eve`, `with-openui`, `with-expo`, `with-react-ink`), adding to an existing Next.js app (`init`), or keeping an install current (`update`, `upgrade`, `codemod`). Covers picking a runtime hook for a backend: `useChatRuntime` (AI SDK v7 via `@assistant-ui/ai-sdk`), `useLangGraphRuntime`, `useStreamRuntime` (LangChain), `useAdkRuntime` (Google ADK), `useA2ARuntime`, `useAgUiRuntime`, `useEveAgentRuntime`, `useOpenCodeRuntime`, `useLocalRuntime`, `useExternalStoreRuntime`, `useDataStreamRuntime`, and `useAssistantTransportRuntime`, plus wiring guides for Mastra, Cloudflare Agents, Claude Managed Agents, LLM gateways, a ChatGPT subscription, and Electron. Also covers the Expo (`--native`) and React Ink (`--ink`) platform scaffolds and the hosted MCP docs server. Route here for the CLI itself, choosing a runtime, or a platform target; once a runtime exists, use elements to install the styled Thread/composer, primitives for unstyled building blocks, runtime for the chosen runtime's hooks and adapters in depth, react-native or ink for platform specifics, and update for a deep migration.
Why this skill is useful
Adds executable CLI commands for setting up and configuring assistant-ui in various project types, which the AI cannot generate from existing knowledge.
What it needs
Requires node installed locally. About 8k tokens when loaded. Last updated 2026-08-02. 20 stars on the source repository.
What this skill does
assistant-ui Setup Always consult assistant-ui.com/llms.txt for the latest API. The assistant-ui CLI scaffolds projects, adds components, and keeps an install current. This skill covers the CLI end to end and the decision of which runtime adapter to wire up next; installing the styled UI itself is elements. References ./references/ai-sdk.md -- @assistant-ui/ai-sdk (v7): transport, frontend tools, multi-step, approval, quote context, token usage, history, useAISDKRuntime ./references/ai-sdk-legacy.md -- pinned @assistant-ui/react-ai-sdk (v6, v5) and @assistant-ui/react-data-stream (v4) setups ./references/langgraph.md -- @assistant-ui/react-langgraph: streaming, interrupts, message editing, threads, agent state ./references/langchain.md -- @assistant-ui/react-langchain: the useStream-based LangGraph adapter, and its comparison to react-langgraph ./references/google-adk.md -- @assistant-ui/react-google-adk: streaming, tool confirmations, auth flows, input requests, artifacts ./references/a2a.md -- @assistant-ui/react-a2a: A2AClient, task states, artifacts, multi-tenancy ./references/ag-ui.md -- @assistant-ui/react-ag-ui: agent state, interrupts, subagent nesting, custom events ./references/eve.md -- @assistant-ui/eve: sessions, connector authorization, withEve + withAui ./references/opencode.md -- @assistant-ui/react-opencode: permissions, questions, sub-agent tasks, session extras ./references/claude-managed-agents.md -- wiring Anthropic Managed Agents sessions through useExternalStoreRuntime ./references/custom-backend.md -- useLocalRuntime and useExternalStoreRuntime for a backend with no dedicated adapter ./references/tanstack.md -- Vite + TanStack Router/Start setup ./references/mastra.md -- Mastra full-stack and separate-server patterns over the AI SDK runtime ./references/cloudflare-agents.md -- Durable Object agents via @cloudflare/ai-chat and useAISDKRuntime ./references/providers.md -- LLM gateways, a ChatGPT/Codex subscription for local dev, and Electron ./ …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills assistant-ui/setup