Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add mindverse-ltd/macaron-artifacts --skill ui4agit clone --depth 1 https://github.com/mindverse-ltd/macaron-artifactsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/mindverse-ltd/macaron-artifacts/ui4a)<a href="https://agentmods.dev/skills/mindverse-ltd/macaron-artifacts/ui4a"><img src="https://agentmods.dev/badge/skills/mindverse-ltd/macaron-artifacts/ui4a/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/mindverse-ltd/macaron-artifacts/ui4a"><img src="https://agentmods.dev/badge/skills/mindverse-ltd/macaron-artifacts/ui4a.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00023 | $0.02304 |
| Opus 5 | $0.00012 | $0.01152 |
| Sonnet 5 | $0.00005 | $0.00461 |
| Haiku 4.5 | $0.00002 | $0.00230 |
Grade A, and why
ui4a scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 108 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Use ordinary prose for an answer that needs no interaction. Use inline UI for a focused choice, calculator, comparison, or small tracker. Use a canvas for a reusable workspace with several views or persistent files.
For inline UI, put one complete React module in a fenced block whose language is exactly ui4a/tsx. Export a default React component. It renders progressively as the block arrives; write a usable scaffold before long data or secondary details. Keep component and import names stable while editing. Close every block and include complete final syntax.
For a canvas, write .artifacts/canvases/<id>.ui4a.tsx; the UI discovers it and opens a side panel. Use a stable descriptive id. Supporting modules can live beside it and use relative imports with explicit .tsx, .ts, .jsx, .js, or .json extensions. Relative module imports are available for file canvases, not inline blocks. Keep the dependency graph acyclic. Mention the canvas briefly in the conversation instead of duplicating its source in a fence.
Use React and the small $ui4a/ui library below by default. Compose native HTML for simple layout; do not wrap every paragraph in a Card. Give controls labels, keyboard access, useful empty states, and real local behavior. Render the user's information immediately; do not replace it with a generic dashboard or mock data. Only send a chat message or write a file after an explicit user action. Surface asynchronous errors beside the relevant control.
For a simple equation in prose, use ( ... ) for inline math or [ ... ] for display math; display $$ on separate lines also works. Single dollar signs remain literal for prices. For mathematical explanations, prefer a compact ui4a/tsx block when a rendered formula or interaction makes the relationship clearer. Use $ui4a/ui/katex for formulas and $ui4a/ui/charts for charts. When parameters affect a function, place the formula, a restrained line chart, and labeled sliders together; derive the formula, plotted samples, and selected values from the same numeric state. Use native range inputs with explicit min/max/step and convert input with Number(event.target.value). Show variable names, units and current values. Keep a stable chart height and domain so changes are comparable, use theme series colors, and avoid decorative chart animation. Import chart primitives from the same built-in module. Put TeX in the value string prop, never raw JSX children; use String.raw or double the backslashes in ordinary JS strings.
Fit the provided container: use width:100%, min-width:0, wrapping text, and container-relative layouts. Avoid page headers, outer background fills, min-height:100vh, and viewport-width sizing. Inherit the host's theme. Semantic UnoCSS colors are surface, surface-2, surface-3, border, fg, muted, accent, accent-fg, danger, success, warn, and series-1 through series-6; use restrained spacing and hierarchy. Do not hard-code a light or dark page.
Use complete UnoCSS utilities in className, such as hover:bg-accent hover:text-accent-fg. The runtime extracts classes without build-time transformers: variant groups such as hover:(bg-accent text-accent-fg), Attributify, and CSS directives such as @apply are unavailable. Prefer literal class strings and mappings for conditional styles.
React, react-dom, react-dom/client, the React JSX runtimes, and the modules below are provided locally. Other npm imports can resolve through the CDN, but prefer built-in capabilities and avoid network dependencies for basic UI. Do not invent $ui4a/* modules; $ui4a/ai, command execution, and directory listing are unavailable. Persist only JSON-compatible state; do not store credentials. File read/write paths start with .artifacts/ and are relative to this session's workspace.
$ui4a/ui
Button: Native button props, variant?: "primary" | "secondary" | "ghost" | "danger", size?: "sm" | "md".Field: Native input props plus label: string and hint?: string. Includes an associated accessible label.Card: Native div props. A restrained filled surface for grouping; prefer spacing and surface hierarchy over adding borders.Badge: Native span props. Compact secondary status or metadata.Tabs: { items: { id: string, label: ReactNode, children: ReactNode }[], value?: string, onChange?: (id: string) => void }.Disclosure: { title: ReactNode, children: ReactNode, defaultOpen?: boolean }. Accessible expandable content; child state survives collapsing.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed · +56 lines 09894faab101
- 2d ago Changed 3c6afdb770d5
- 3d ago First seen · 52 lines · 23 tokens per session scan A 3ead61f3a05b
ui4a is a skill published in the GitHub repository mindverse-ltd/macaron-artifacts (10 stars, last pushed today), licensed Apache-2.0. It adds 23 tokens to every session and 2,304 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-09.
Other skills, from other repositories
openui-forge-zh
A guide to building generative user interfaces with OpenUI, an open format in which a language model describes interactive components that a React application renders as they arrive. It supports different language-model providers and backend languages.
openui-forge
Build generative UI with OpenUI — any LLM provider, any backend language. Scaffold, integrate, validate.
openui-forge-openai
OpenUI generative UI with OpenAI SDK backend. Streaming chat completions with gpt-5.5 (or any current OpenAI-compatible model).
openui-forge-anthropic
OpenUI generative UI with Anthropic Claude SDK backend. Stream conversion to OpenAI NDJSON format.
openui-forge-go
OpenUI generative UI with Go (net/http) backend. Direct OpenAI API streaming via HTTP.
react-patterns
React 18/19 patterns including hooks discipline, server/client component boundaries, Suspense + error boundaries, form actions, data fetching, state management decision trees, and accessibility-first composition. Use when writing or reviewing React components.