More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Kimi Widget: Interactive Widgets Inside the Chat

    1

    Decide when an inline widget earns its place, then build it well: self-contained, glanceable, styled by the host's design tokens.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Kimi Widget: Interactive Widgets Inside the Chat

    Kimi Widget: Interactive Widgets Inside the Chat

    Example session with this skill installed

    Should I choose plan A or B? Plan A is $20/mo flat, plan B is $0.10 per unit. I use about 150 units a month, sometimes 400.

    • Read your context and instructions
    • Compiled the kimi widget

    Go decision: comparative plus numeric structure — exactly what a widget is for. Built a small calculator widget: a slider for monthly units, live cost comparison for both plans, and the break-even point highlighted at 200 units. Default state legible in two seconds: at 150 units Plan A costs $20 and Plan B costs $15, so Plan B wins today; at 400 units Plan A costs $20 and Plan B costs $40, so Plan A wins. The break-even line at 200 units is the visual anchor. Runtime contract honored: self-contained HTML/CSS/JS in the sandboxed iframe, no external scripts or fonts, colors from the host's design tokens so it reads correctly in dark mode, the break-even chart hand-built in SVG. Response text outside the widget: one sentence framing the trade-off — Plan B while usage stays under 200 units a month, Plan A once it crosses that line. No narration after the widget rendered; the visual carries it. Follow-up answered from the numbers, not from the widget's existence: if usage keeps spiking past 200, switch to A; the slider lets the user test their own months. The widget was checked in both light and dark mode before delivery, since hardcoded colors would have blinded half the audience. One job, one widget.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    A widget is a small visual surface rendered inside the chat — a calculator, a sortable comparison table, a timeline you can scrub, a state diagram you can step through. Used well, it replaces three paragraphs. Used badly, it is a distraction with buttons. This skill covers the go/no-go decision first: build for spatial, comparative, numeric, or interactive structure, and never for plain prose, code walkthroughs, file listings, or anything that is really a small app. Then it covers the runtime contract — self-contained HTML/CSS/JS in a sandboxed iframe, no external dependencies, styled through the host's design tokens, hand-built charts, graceful degradation — and the design rules that make widgets glanceable: one job, legible in two seconds, details on interaction.

    It is for small visual surfaces, not small apps — anything that blocks on user input to proceed or performs destructive actions is out of scope.
    Explanations, intros, and summaries live in the chat response, not inside the widget.
    If the host injects a sendPrompt bridge, pushed text must read as a genuine user message, never a robotic confirmation.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 3 days ago

    Permissions

    No special permissions declared or detected

    Listed3 days ago

    What's inside

    Frequently Asked Questions

    When the answer has structure prose handles poorly: spatial or sequential, comparative, numeric the user will play with, or interactive step-throughs. Plain prose answers do not need one.

    No. The runtime contract forbids external scripts, stylesheets, images, fonts, CDN libraries, fetch, and WebSocket. Charts are hand-built with SVG, Canvas, or DOM.

    The widget shows; you tell. Explanations, intros, and summaries live in your chat response. After it renders, stop narrating it — say only what the visual cannot.

    New in Digital Products