More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Icon Generator: Consistent UI Icon Sets

    1

    Produce a coherent icon set, not a pile of singles: lock the grid, stroke, and corner system first, build a seed set, expand against a scored rubric, and script the exports.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    Icon Generator: Consistent UI Icon Sets

    Icon Generator: Consistent UI Icon Sets

    Example session with this skill installed

    We are building a music practice app and need about 30 icons: playback, metronome, tuner, practice stats, settings, the usual navigation. Our designer made a few nice ones but they do not match each other. Can you set up the system and show how the first icons would be produced?

    • Read your context and instructions
    • Compiled the icon generator

    SYSTEM (locked before drawing): 24px grid, 2px inset safe area, stroke treatment at 2px with round caps and joins, corner radius 3px, adjusted small-size versions below 20px, personality: "precise and friendly, like good sheet music." Stroke beat fill in the decision table because the smallest common size in the app is 20px and the screens are airy.

    SEED SET BUILT FIRST (9 anchors): home, search, play, pause, settings, back, plus, check, menu. Each tested at 16/24/48px on the day it was drawn, not at the end.

    WORKED ICON — metronome. First draft copied a realistic metronome: pyramid body, swinging arm, tempo window. At 16px it read as a triangle with a stick — rubric clarity 0, rejected. Second metaphor: simplified pyramid with the arm at a strong diagonal and the window dropped. Squint test: readable. Lineup against the seed set: same stroke, same radius, weight matched. Scores: system fit 2, clarity 2, metaphor 2, weight 2, distinctness 2 = 10/10, admitted.

    COLLISION CAUGHT: "practice stats" (bar chart) and "tuner" (a needle gauge) converged at small size. The gauge silhouette changed to a tuning-fork arc — the standard tuner metaphor — rather than shrinking the chart.

    EXPORT: SVG sources named ic_<name>_24.svg; one script emits 16/24/32/48 PNG, web SVG, iOS PDF, Android vector drawables, and validates that every source is on-grid at exactly 2px stroke with no embedded rasters. Report: 9/9 seed icons pass; the designer's original unmatched icons were rebuilt on the system rather than imported, so the set the remaining 21 icons grow into is coherent from the first row.

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

    About this skill

    Icons fail as individuals and succeed as families. This skill is the production method behind a set that still looks like one set at icon 200: five system decisions (grid, stroke-vs-fill, corner radius, small-size rule, personality) made once and written down, a seed set of 8–12 anchors built and approved before any expansion, and every later icon admitted only after passing the squint test, the lineup test, and a 10-point coherence rubric. It covers choosing metaphors users already recognize, resolving silhouette collisions, using AI image generation correctly (concepts to rebuild on the grid, never finished icons pasted into a set), and automating the whole export matrix — PNG sizes, web SVG, iOS PDF, Android vectors — behind one script that validates grid and stroke compliance and fails loudly on violations. Dark-mode weight, brand redesigns, and deprecated icons are handled as system changes, not one-off edits.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for UX & UI Design.

    • It defines and enforces a system; the drawing itself still happens in your vector tool of choice.
    • Judging metaphors uses small tester checks described in the method — it cannot run live user research for you.
    • Sets with strong existing brand iconography may need the forked-spec approach described for platform divergence.

    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 buying a skill.

    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

    The system scales down: for a handful of icons it is five decisions on one page, and it is exactly what stops those few from clashing with each other and with your interface. The full pipeline pays off as the set grows.

    Yes, as concepts. The method uses generated images for metaphor exploration, then rebuilds the chosen idea on your grid with your stroke system. Raw generated icons pasted into a set drift visibly and fail the coherence rubric.

    They become the reference anchors: the system is reverse-derived from them (grid, stroke, radius measured, not guessed), the seed set is validated against them, and new icons are admitted against that standard.

    New in UX & UI Design