Agentation

This skill should be used when the user asks to "create a plugin", "scaffold a plugin", "understand plugin structure", "organize plugin components", "set up plugin.json", "use ${CLAUDE_PLUGIN_ROOT}", "add commands/agents/skills/hooks", "configure auto-discovery", or needs guidance on plugin directory layout, manifest configuration, component organization, file naming conventions, or Claude Code plugin architecture best practices.

Overview and router for assistant-ui, the React library for building AI chat interfaces from composable primitives. Use for high-level, cross-cutting, or architecture-overview questions: choosing packages, picking a runtime, or understanding the layered model (RuntimeCore, Runtime, the aui client, primitives) and message model. Covers the `@assistant-ui/react` core plus `@assistant-ui/react-ai-sdk`, `@assistant-ui/react-langchain`, `@assistant-ui/react-langgraph`, `assistant-stream`, and `assistant-cloud`, and the platform bindings `@assistant-ui/react-native` and `@assistant-ui/react-ink`; `AssistantRuntimeProvider`; the primitives `ThreadPrimitive`, `MessagePrimitive`, `ComposerPrimitive`; the hooks `useAui`, `useAuiState`, `useAuiEvent`; and runtime selection across `useChatRuntime`, `useExternalStoreRuntime`, `useLangGraphRuntime`, `useLocalRuntime`. Current line is `@assistant-ui/react` 0.15.x on AI SDK v7. For a specific area route to a focused sibling instead: setup, runtime, pr

Vision-driven browser automation using Midscene. Operates from screenshots — no DOM or accessibility labels needed. Runs in headless Puppeteer — does NOT take over the user's mouse or keyboard. Also supports CDP mode and Bridge mode to connect to an existing Chrome. Use this skill when the user wants to: - Browse, navigate, or open web pages - Scrape, extract, or collect data from websites - Fill out forms, upload local files, click buttons, or interact with web elements - Verify, validate, test, or QA frontend UI behavior - Take screenshots of web pages - Automate multi-step web workflows - Test what was just built, see if it works in browser - Connect to Chrome via CDP, DevTools Protocol, or remote debugging - Connect to user's Chrome browser, control my browser, operate my Chrome Powered by Midscene.js (https://midscenejs.com)

Generate a design tokens file (CSS variables or Tailwind config) based on a chosen aesthetic philosophy, with light and dark mode palettes, spacing scale, type ramp, and component-level tokens. Use when starting a new project, establishing a visual system, setting up tokens, or mentions "tokens" or "design system".

Build or redesign polished browser-rendered visual artifacts with HTML/CSS/JavaScript/React: pages, dashboards, prototypes, slide decks, animations, UI mockups, and data visualizations. Use for visual front-end creation, design-system exploration, design critique, or explicit browser acceptance / QA of a web artifact. Not for back-end, CLI, non-visual coding, source-to-longform article conversion, or narration-driven click-through video presentations.

Registers LLM tools and renders custom tool-call UI in assistant-ui (@assistant-ui/react). Covers the current authoring model: a "use generative" file exporting defineToolkit({...}) that co-locates schema, execute, and render and is split across the client/server boundary by the build plugin (withAui from @assistant-ui/next, aui() from @assistant-ui/vite, withAui from @assistant-ui/metro), mounted with Tools({ toolkit }) through useAui and exposed to the model with AISDKToolkit. Covers the tool kinds and their execute markers: backend (plain execute), frontend ("use client" inside execute), human via humanTool()/hitl, provider via providerTool(), external via externalTool(), stubs via stubTool() plus useAuiToolOverrides, and MCP via defineMcpToolkit. Also covers the component-scoped path (makeAssistantTool / useAssistantTool / makeAssistantToolUI / useAssistantToolUI), ToolCallMessagePartProps (status.type running/complete/incomplete/requires-action, args, argsText, result, artifact, a