Soft Vector logo

Soft Vector

Organization
vm0-ai
soft-vector

Airy flat-vector spot illustration style — square 1:1 pale-blue canvas, delicate ~1.5-2px dark-navy linework with intentionally OPEN / unclosed outlines, flat color fills, one muted accent hue (mustard / lime-sage / warm-orange / violet), centered subject with generous breathing room, scattered thin-navy micro-marks (dots, plus signs, sparkles, or motion ticks). Trigger when the user says /soft-vector, asks for a "soft-vector illustration", a "friendly UI onboarding illustration", a "thin-line flat spot illustration", or briefs a scene metaphor + palette + complexity in this house style.

Overview

Publishervm0-ai
Repositoryvm0-skills
Skill namesoft-vector
Stars
76
Forks
18
Bundled files
3
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.

  • 3 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by vm0-ai on GitHub. Read the source before you install it.

Installation

Install the Soft Vector 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.
  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/vm0-ai/vm0-skills.git /tmp/vm0-skills
mkdir -p .claude/skills
cp -r /tmp/vm0-skills/illustration-template/soft-vector .claude/skills/soft-vector
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Soft Vector 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 Soft Vector 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 Soft Vector 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.

/soft-vector — locked airy flat-vector spot illustration style

Friendly, calm, "modern UI onboarding" spot illustrations. Every piece holds the same locked frame: a soft pale-blue square canvas, delicate dark-navy linework, flat color fills, and one muted accent hue. The signature lift comes from intentionally open / unclosed outlines — small visible gaps in rims, handles, ground lines, cloud silhouettes — that keep the artwork feeling hand-drawn rather than mechanical.

Think Storyset / Lottiefiles airy spot art with the airiness pushed harder by the open-line rule.

Prompt interpretation

The user will usually give a short brief — sometimes just a concept ("morning routine"), sometimes concept + palette ("empty inbox in violet"), sometimes concept + complexity ("focus mode, L3"). Your job is to turn that brief into a complete spot illustration spec without pausing for clarification:

  1. Pick a hero subject from the brief — a single theme-native object that visually carries the concept. One mug for morning, one mailbox for inbox, one lamp + book + plant for focus. The scene IS the metaphor — never default to generic filing-cabinet + megaphone.
  2. Pick a complexity level (L1 / L2 / L3) based on how rich the brief is.
  3. Pick a palette accent that matches the subject's mood (see palette options below).
  4. Pick an accent-mark motif — dots, plus signs, sparkles, or motion ticks — that fits the subject's emotional world.
  5. Pick 2–3 specific open-line callouts for the chosen subject. These vary by subject: mug rim gap, handle that doesn't close, ground line that ends in mid-air, cloud silhouette break, lamp arm fade, pot rim gap, etc.

Bias toward subjects that are concrete and iconic — things you can name in one or two words.

Locked style axes (NEVER vary)

Canvas

  • Square 1:1, recommended 1024×1024
  • A single flat pale-blue field fills the entire canvas — #EEF1F5
  • No gradient, no border, no horizon, no texture

Line quality

  • Thin dark-navy outlines, ~1.5–2px at 1024-wide canvas (delicate, NOT chunky)
  • One uniform stroke colour everywhere (a single deep navy, ~#1F2A44)
  • Uniform stroke weight on each shape — no calligraphic taper, no pencil scratch
  • Rounded line caps (no flat or square caps)

Open outlines (the signature)

  • Every piece must include several intentionally unclosed outlines. Examples by subject:
    • Mugs / cups → rim has a small visible gap; handle does not return to the body; steam wisps end as free-floating lines
    • Vehicles → ground line ends in mid-air on both sides; one body panel doesn't fully close
    • Buildings → roof eaves run past the wall; door jamb doesn't reach the floor
    • Plants → pot rim has a tiny gap; one leaf outline breaks at the tip
    • Clouds / tree blobs → one or two small visible breaks in the silhouette
    • Devices → speech-bubble lines are floating ticks, not closed pills
  • This open-line treatment is non-optional — it is what distinguishes Soft Vector from generic flat illustration. Always name the specific open-line spots when briefing a model.

Fill

  • Strictly flat colour — no gradients, no shading, no ambient occlusion, no inner highlights, no halftone
  • One lead accent hue (the chosen palette dial), plus always-available neutrals: soft grey-white #F4F5F1 and the dark-navy stroke #1F2A44
  • One optional secondary accent is permitted on L3 vignettes (e.g., warm-orange book cover inside a violet-led desk scene)

Subject placement

  • Centered, occupying ~60–75% of the canvas
  • Generous breathing room around the subject — pieces never fill edge-to-edge

Micro-marks

  • Scattered decorative marks drawn in the same thin navy stroke
  • Vocabulary: tiny dots, small plus signs, four-point sparkles, short motion ticks
  • 4–8 marks total — sparse but alive
  • Pick one motif per piece (a few dots can mix with anything)

What is NOT in the canvas

  • No text, no caption, no headline, no wordmark, no logo, no UI chrome
  • No full human figures, no faces
    • Humans appear at most as a silhouette behind glass or a partial hand
  • No second illustration, no border, no patterned background

Variable axes (the dials)

These are the only things that should change between pieces.

#DialWhat it controlsOptions
1Palette accentThe single muted hue leading the piecemustard #E8B53A · lime-sage #A8C95A · warm-orange #E87A3A · violet #7B6FE0
2Scene metaphorThe hero subject + any environment propscoffee mug for morning · bicycle for commute · lamp + book for focus · phone + bell for ping · mailbox for inbox · envelope + note for messages, etc. — always theme-native
3ComplexityHow many props share the canvasL1 single object hero · L2 object + 1–2 environment props · L3 vignette with 3–5 props composing a scene
4CastOptional human elementnone · silhouette-behind-glass · partial hand (never full figures, never faces)
5Accent marksThe micro-mark motifdots · plus signs · sparkles · motion ticks
6Action / moodEnergy in the scenestatic · gesture (wave, tilt) · in-motion (rolling wheel, ringing bell)

Palette mood pairing hints

  • Mustard → warm, welcoming, morning, food / drink, mail
  • Lime-sage → outdoor, commute, nature, growth
  • Warm-orange → energetic, billing / cards, alerts, activity
  • Violet → focus, evening, reading, devices

Brief template

When generating, expand the user's input into this internal brief before describing the image:

Palette accent:    <one of mustard / lime-sage / warm-orange / violet, with hex>
Scene metaphor:    <one theme-native subject + environment, in one sentence>
Complexity:        <L1 / L2 / L3, with prop count>
Cast:              <none / silhouette-behind-glass / partial hand>
Accent marks:      <one motif, with rough count>
Action / mood:     <static / gesture / in-motion>
Open-line spots:   <2-3 specific outlines that will be unclosed, named explicitly>

Generation guidance

Models with strong flat-vector handling and crisp uniform stroke output (e.g. gpt-image-1.5, nano-banana-2, or gpt-image-2) produce the most reliable output here. When briefing the model:

  • State explicitly that the canvas is square 1:1 and the background is a single flat pale-blue #EEF1F5 field with no gradient, no border, no horizon.
  • State explicitly the stroke style — thin (~1.5–2px), uniform, dark navy #1F2A44, rounded caps, NOT chunky.
  • State explicitly that several outlines are intentionally OPEN / unclosed, and name the 2–3 specific spots for this scene (e.g. "the mug rim has a small visible gap, the handle does not return to the body, the steam wisps end as free-floating lines"). Without naming the open-line spots, most models default to fully closed mechanical outlines.
  • State explicitly that fills are flat — no gradients, no shading, no inner highlights.
  • Name the accent-mark vocabulary explicitly — models otherwise default to generic stars / sparkles everywhere.
  • State explicitly "no text, no logo" — models often slip in extra text on UI subjects (phones, mailboxes).

If the model produces chunky 3–4px outlines, regenerate with stronger emphasis on "thin, delicate, NOT chunky". If outlines are all closed, regenerate with the specific open-line spots called out more aggressively (e.g. "the rim MUST have a visible 4–6px gap on the left").

Worked examples

Three reference pieces below — one per complexity level — each holds the locked frame and varies the dials.

Example 1 — coffee mug · mustard · L1 (single object hero)

coffee mug · mustard · L1

  • Palette accent: mustard #E8B53A
  • Scene metaphor: ceramic coffee mug with two short steam wisps (morning)
  • Complexity: L1 — single object hero
  • Cast: none
  • Accent marks: plus signs + scattered dots
  • Action: static
  • Open-line spots: mug rim has a small gap on the left, handle does not fully close at the bottom, steam wisps end as free-floating lines

Example 2 — bicycle scene · lime-sage · L2 (object + environment)

bicycle scene · lime-sage · L2

  • Palette accent: lime-sage #A8C95A
  • Scene metaphor: vintage bicycle parked on a short ground line, one rounded sage tree behind, one bumpy white cloud above (commute)
  • Complexity: L2 — object + 1–2 environment props
  • Cast: none
  • Accent marks: short motion ticks behind the rear wheel + 2 tiny sky dots
  • Action: static (parked)
  • Open-line spots: ground line ends in mid-air on both sides, tree blob silhouette has small breaks at the top, cloud outline has a tiny gap on the right

Example 3 — desk vignette · violet · L3 (full scene)

desk vignette · violet · L3

  • Palette accent: violet #7B6FE0 (lead); warm-orange #E87A3A (secondary, on book cover + pot)
  • Scene metaphor: violet desk lamp arched over the desk, open book in the center, small potted plant with two tall leaves on the right, two loose papers tucked under the book corner (focus mode)
  • Complexity: L3 — vignette with 4 props
  • Cast: none
  • Accent marks: small four-point sparkles near the lamp head + a few tiny dots near the plant
  • Action: static
  • Open-line spots: desk surface line ends in mid-air on both sides, lamp arm fades into a free-floating line near the joint, pot rim has a tiny gap

Anti-patterns

  • Chunky 3–4px outlines — too heavy, kills the airy feel.
  • Fully closed mechanical outlines everywhere — looks like Material Icons, not Soft Vector.
  • Saturated neon palette — accents stay muted; lime sits well below pure neon green.
  • Gradients, drop shadows, inner highlights — strictly flat.
  • Subject filling the canvas edge-to-edge — needs breathing room.
  • Full human figures or faces — humans are suggestive at most.
  • Generic filing-cabinet + megaphone metaphors — pick a theme-native scene.
  • Adding text labels onto UI subjects (phone screens, mailbox flags) — keep the canvas text-free.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Soft Vector AI skill do?

Airy flat-vector spot illustration style — square 1:1 pale-blue canvas, delicate ~1.5-2px dark-navy linework with intentionally OPEN / unclosed outlines, flat color fills, one muted accent hue (mustard / lime-sage / warm-orange / violet), centered subject with generous breathing room, scattered thin-navy micro-marks (dots, plus signs, sparkles, or motion ticks). Trigger when the user says /soft-vector, asks for a "soft-vector illustration", a "friendly UI onboarding illustration", a "thin-line flat spot illustration", or briefs a scene metaphor + palette + complexity in this house style.

Why use Soft Vector on TypingMind?

Because you install it once and use it with any model. Soft Vector 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 Soft Vector in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/vm0-ai/vm0-skills/tree/main/illustration-template/soft-vector. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Soft Vector?

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 Soft Vector?

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

Is the Soft Vector AI skill free?

It is published on GitHub by vm0-ai. 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 👇