Kaicalls Design logo

Kaicalls Design

Community
cgallic
kaicalls-design

Build KaiCalls-compliant UI components and pages using the KaiCalls design system. Use when creating new dashboard pages, components, forms, tables, cards, badges, or any UI within the KaiCalls admin panel. Ensures dark-first aesthetic, proper color tokens, typography classes, motion standards, and B2B SaaS best practices. Triggers on "build a page", "create a component", "add a dashboard", "design a form", or any frontend work in the adminpanelnew codebase.

Overview

Publishercgallic
Repositorykai-cmo-harness
Skill namekaicalls-design
Stars
61
Forks
8
Bundled files
Instructions only
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by cgallic on GitHub. Read the source before you install it.

Installation

Install the Kaicalls Design AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/cgallic/kai-cmo-harness/tree/main/harness/skills-v2/kaicalls-design
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/cgallic/kai-cmo-harness.git /tmp/kai-cmo-harness
mkdir -p .claude/skills
cp -r /tmp/kai-cmo-harness/harness/skills-v2/kaicalls-design .claude/skills/kaicalls-design
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Kaicalls Design in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Kaicalls Design on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Kaicalls Design is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

kaicalls-design — UI that belongs in the KaiCalls admin panel

Objective

UI that a KaiCalls user cannot tell was added later: dark-first, built from design-system tokens and shadcn/ui components, with every interactive element giving feedback the user can feel.

The product target is a Minimum Lovable Product — zero support required (so intuitive nobody opens docs), every click invokes emotion, micro-celebrations reward tedious work, conversational copy and helpful empty states make the software feel like a teammate, and the whole surface reads premium in the register of Linear, Notion, and Stripe.

Two rules pull against each other and both hold: every interactive element gives emotional feedback, and that feedback stays subtle. Our users include older professionals who spend long hours on screen; nothing flashes, strobes, or strains eyes.

Done when

Work type product-ui — floor E5/C4/O3 (harness/eco-floors.yaml).

  • E5 — the change is live in the product and a non-actor loaded the real screen at the real breakpoints. A green build and a deploy marked ready are E1 and E4; neither proves the screen renders.

  • C4 — the design field standard, verified before ship: hierarchy, WCAG AA contrast and focus order, token compliance over hardcoded values, responsive behavior at actual viewing sizes, and every loading, empty, and error state. Deviations need an explicit expiring waiver. The project's own tests plus this checklist are the C bar:

  • Every button has emotional feedback: hover, active, and success states

  • Feedback is subtle — no flashing, strobing, or harsh contrast

  • Design-system color tokens throughout, no hardcoded hex

  • Typography uses semantic classes (heading-*, text-body-*), not ad-hoc sizes

  • Loading states use skeletons, not spinners

  • Empty states carry helpful copy and a CTA

  • Success actions show a micro-celebration (toast, animation, or color change)

  • Destructive actions have confirmation plus reassurance afterward

  • Animations run at 60fps and stay under 300ms

  • prefers-reduced-motion respected via motion-safe / motion-reduce

  • Dark-mode compatible — no light-only colors

  • Mobile responsive down to 375px

  • shadcn/ui components used wherever one exists

  • O3 — name the user behavior the change was supposed to move (task completion rate, time to first action, support tickets on that surface, drop-off) with a baseline, threshold, owner, and a 30-day window, recorded before building. A UI change that shipped and moved nothing is SHIPPED, not CLOSED.

Constraints

  • Tokens, never hex. Every color comes from the CSS variables or kai-* Tailwind classes below. A hardcoded hex is a defect even when it matches.
  • Motion budget is hard. Micro-interactions (toggles, buttons) 120–160ms ease-out · small transitions 200–240ms ease-out · screen transitions 300–400ms cubic-bezier(0.4, 0, 0.2, 1). Tokens: --transition-fast: 150ms, --transition-normal: 200ms, --transition-slow: 300ms.
  • Forbidden feedback: bright flashes, strobing, high-contrast pops (white on dark), large jarring movement, anything over 300ms, anything that pulls attention off the task. Use soft glows, gentle scale (1.02–1.05x), muted color transitions, and spring physics with high damping.
  • Eye comfort: muted success green (#34D399 at ~60% opacity) rather than bright green · soft error red (#FF4B4B with transparency), never pure red · confetti and particles sparse and low-contrast · prefer opacity transitions to color flashes.
  • Never ship "No data found." An empty state carries an optional illustration, copy explaining what belongs there, and a CTA that populates it.
  • Errors speak plainly. "Something went wrong. We're looking into it. Try refreshing the page." — never a raw status code.
  • Destructive actions name their consequences in the confirm dialog (what is deleted, what goes with it, that it cannot be undone), and reassure immediately after.

Context

NeedLoad
Complete CSS variable and Tailwind class listreferences/full-color-tokens.md
Real-world dashboard, table, and form patternsreferences/component-examples.md

Core color tokens

css
/* Surfaces */          /* Brand */                    /* Text */
--background: #12121A   --kai-blue: #5C8CFF            --foreground: #F7F7FA
--card: #181B23         --kai-purple: #C68BF8          --muted-foreground: #A7A9BE
--secondary: #23263B    --kai-btn-primary: #3575F6     --kai-text-placeholder: #6B6D76
--surface-elevated: #2A2D3E

/* Status */
--kai-success-text: #34D399   --kai-error-text: #FF4B4B
--kai-warning-text: #FFC04B   --kai-info-text: #5C8CFF

Tailwind classes. Surfaces: bg-background (page), bg-card (cards), bg-secondary (inputs and secondary surfaces). Text: text-foreground (primary), text-muted-foreground (secondary), text-kai-blue, text-kai-purple. Borders: border-border, border-kai-table-border.

Typography. Headings heading-xl (2xl bold white, page title) · heading-lg (xl semibold, section) · heading-md (lg semibold, card) · heading-sm (base semibold, subsection). Body text-body-primary (white) · text-body-secondary (slate-400) · text-body-tertiary (slate-500, timestamps). Labels label-uppercase (xs uppercase tracking-wide slate-400) · label-default (sm medium slate-400) · stat-label (xs uppercase slate-400). Values use data-value.

Components. All from @/components/ui/*.

  • Card — Card / CardHeader / CardTitle / CardDescription / CardContent. Base: rounded-lg border-slate-700/50 bg-slate-800/60 text-white shadow-sm.
  • Button — variants default (primary), secondary, outline, destructive, ghost, link; sizes sm, default, lg, icon.
  • Badges — temperature: badge-hot (bg #FF4B4B33, text #FF6B6B), badge-warm (#FFC04B33 / #FFD76B), badge-cold (#4B7BFF33 / #6B93FF). Status: status-success, status-error, status-warning, status-info. All badges: rounded-full px-2.5 py-0.5 font-semibold.
  • Tables — table-row hover:bg-kai-hover-row, table-row-selected, cells px-4 py-3, headers text-kai-text-table-heading (#B0B2C3).
  • Inputs — Input + Label wrapped in space-y-2. Base bg-secondary border-border text-foreground, focus ring-kai-blue.
  • Dialogs — DialogContent className="bg-modal-bg border-modal-border" with DialogHeader / DialogTitle / DialogDescription. Destructive flows use AlertDialog with a keep-it action and a consequence-naming confirm action.
  • Elevation — shadow-elevation-sm (subtle), -md (cards), -lg (dropdowns), -xl (modals). Z-index: z-dropdown 50, z-modal 100, z-toast 150.
  • Gradient — bg-gradient-kai and text-gradient-kai, both linear-gradient(90deg, #5C8CFF 0%, #C68BF8 100%).
  • Animation classes — skeleton-shimmer for loading, collapsible-content and collapsible-chevron for disclosure.

Micro-celebration triggers

ActionFeedbackImplementation
Task completeSoft checkmark fade-inIcon opacity 0→1 over 200ms
Form submitGentle green tint, quiet toastSubtle bg shift, no flash
Delete / archiveSmooth slide awaySlide-out 250ms with opacity fade
Toggle onSoft snap into placeSpring, damping 0.8
Save successSoft glow around buttonBox-shadow fade in/out over 400ms
ErrorGentle wobble + muted red2–3px horizontal shake, soft red

Hover and focus. Buttons hover:scale-[1.02] hover:brightness-105 transition-all duration-150. Cards hover:shadow-elevation-md hover:-translate-y-0.5 transition-all duration-200. Links hover:text-kai-purple/80 transition-colors duration-150. Reduced motion: motion-safe:hover:scale-[1.02] motion-reduce:hover:opacity-80, or check window.matchMedia('(prefers-reduced-motion: reduce)').matches in JS.

Button states. A success button flips to a Check icon with bg-kai-success-text scale-105 for ~1500ms, then reverts. A loading button is disabled and shows <Loader2 className="mr-2 h-4 w-4 animate-spin" /> with an anticipatory label ("Saving...").

Toasts come from sonner, with a title and a description that says what actually happened — toast.success("Agent created!", { description: "Your AI assistant is ready to take calls." }). Milestones may carry an emoji icon. After a destructive action, the toast reassures: "Agent deleted — don't worry, your leads are still safe."

Page layout. A page is space-y-6: a header row (flex items-center justify-between) holding heading-xl plus a text-body-secondary mt-1 description on the left and the primary action button on the right, followed by content in grid gap-6 md:grid-cols-2 lg:grid-cols-3.

Stat card. CardContent className="p-6" containing a flex items-center justify-between row with stat-label and a muted icon, then mt-2 with data-value and a delta in text-kai-success-text text-sm ml-2.

Loading. Skeletons match content shape — a card skeleton mirrors header and body block heights (h-6 w-48, h-4 w-32 mt-2, h-24 w-full); a table skeleton repeats rows of column-width bars.

Escalate when

  • The design calls for a color, spacing value, or component that no token or shadcn/ui component covers — the system extends deliberately, not inline.
  • Feedback strong enough to be noticed would breach the eye-comfort rules.
  • A pattern is needed that references/component-examples.md does not cover and would set precedent across the admin panel.
  • A flow requires a destructive action without a safe undo or a clear consequence statement.
  • Accessibility (contrast, focus order, reduced motion) and the requested aesthetic conflict.

Frequently asked questions

What does the Kaicalls Design AI skill do?

Build KaiCalls-compliant UI components and pages using the KaiCalls design system. Use when creating new dashboard pages, components, forms, tables, cards, badges, or any UI within the KaiCalls admin panel. Ensures dark-first aesthetic, proper color tokens, typography classes, motion standards, and B2B SaaS best practices. Triggers on "build a page", "create a component", "add a dashboard", "design a form", or any frontend work in the adminpanelnew codebase.

Why use Kaicalls Design on TypingMind?

Because you install it once and use it with any model. Kaicalls Design is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Kaicalls Design in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/cgallic/kai-cmo-harness/tree/main/harness/skills-v2/kaicalls-design. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Kaicalls Design?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Kaicalls Design?

As many as you like. As long as a model supports skills, you can use Kaicalls Design with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Kaicalls Design AI skill free?

It is published on GitHub by cgallic. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇