More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Expo Design System: Tokens and Components

    1

    One visual source of truth for your Expo app: detect the system you already have, derive token scales, shape component contracts, promote views at the right moment, and audit drift.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Expo Design System: Tokens and Components

    Expo Design System: Tokens and Components

    Example session with this skill installed

    My Expo app has 14 screens built over a year by me and two contractors. Buttons look different on almost every screen, spacing is inconsistent, there's no theme file, and dark mode is half broken. I don't want a rewrite — where do I start?

    • Read your context and instructions
    • Compiled the expo design system

    DETECTION FIRST — no theme file and no styling library in package.json, so this is a de facto system: tokens get derived from what the app actually uses, not imposed from a template.

    Census of the 14 screens: 11 distinct button background hexes, 4 of them within a few points of each other (one real accent, four typos of it); spacing literals cluster on 8/12/16/24 with outliers at 10, 14, and 18; fontSize appears 60+ times outside any shared component. Most-repeated values win: the accent becomes the modal hex, spacing snaps to the 4-point scale (outliers migrate to the nearest step), and the type ramp is built from the six modal sizes.

    Deliverable one: a single theme.ts holding spacing, radius, type, and motion, plus semantic colors resolved through Platform.select with the brand accent as an explicit light/dark pair. That is also the dark-mode fix: the half-broken mode comes from per-screen conditionals, so colors move to resolution at the token layer instead of being patched screen by screen.

    Component calls: Button appears on 13 screens in 11 costumes, so it promotes now — primary/secondary/destructive variants, three sizes mapped to tokens, mandatory pressed state, style merged last. Its variant set covers every observed costume except one screen's green confirm button, which is a missing variant, not a new component. The order-row layout repeats on two screens, but its props would re-expose every internal style, so it stays colocated until its API shrinks.

    Drift order after tokenizing: score-3 items first (the four near-miss accent hexes, the green confirm routed through Button), then the missing-token repeats. Screens change one at a time as they're touched, each verified by screenshot against the four audit questions — no rewrite, no big-bang restyle.

    FIRST — no theme file and no styling library in package.json

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

    About this skill

    Apps don't drift on purpose — they drift one pasted hex at a time. This skill builds, or repairs, the single visual source of truth inside an Expo app, and it starts with detection rather than construction: if a styling system already exists (NativeWind, Tamagui, Restyle, Unistyles, or a hand-rolled theme), it is the authority and gets extended in its own idiom; if only de facto values exist, tokens are derived from the most frequent ones and snapped to a 4-point grid. It defines the token scales — spacing, platform semantic colors with an explicit brand pair, a named type ramp consumed through one text component, radius, shadows, motion — and the four-part contract every shared component must declare: variants, sizes, states (pressed feedback mandatory), and a style prop merged last. A three-test rule governs when a repeated view earns promotion to a shared component, and a grep-driven drift audit with a 1–3 scoring rubric fixes what already shipped, in score order, without a rewrite and never by installing a second system beside a working one.

    • The default scales (4-point spacing, six-step type ramp, three elevations) are starting points for apps without a system; an existing system's own scales always override them.
    • It governs values and component contracts, not brand or product design — it won't invent your visual identity, only make the one you have consistent.
    • Third-party SDK screens (auth, payments, maps) can only be matched at their boundaries; their internals are excluded from the system and the audit.

    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 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    Yes, and that's the common case. The skill detects the existing system, treats it as the authority, extends it in its own idiom, and audits drift against it. It never installs a second system beside a working one.

    No — that's the failure mode this skill avoids. Tokens land first, then screens migrate one at a time as they're touched, in drift-score order, with screenshots re-verified after each change.

    When all three tests pass: it appears in two or more screens, it has a nameable role, and its API is smaller than its implementation. Until then it stays colocated with its screen — a second copy is cheaper than a primitive with a bad API.

    New in Mobile Development