WayToAI Image Inspirer

Image Inspirer.skill

See it. Learn it. Adapt it.

1,538 cases across 13 categories.

Every case ships with a finished image. See it, then make it yours.

Install View GitHub

13 categories · 1538 cases · GPT-Image2 · Apache-2.0

The problem

The prompt is written. The image still misses.

  1. The prompt is flying before the need is clear.

  2. You can find pictures. Not cases with finished images.

  3. Vague type, subject, or style, and the render fails.

1,538 cases and a pitfall method, ready inside your agent.

Capabilities

Search, guide, and rewrite

  1. Search cases and reuse them

    Route by type. Pick prompts that already have a finished image.

  2. Ask first, then draw

    Confirm type, subject, use, style, and ratio.

  3. Rewrite from a case

    Change subject, headline, and constraints. Do not guess from a blank page.

  4. Pitfalls become constraints

    Readable text, locked style, complete parameters.

Categories

13 categories, 1,538 cases

Each type has prompts and finished images. Search by use.

  1. 344

    Posters & Layout

    Campaigns, movie posters, covers

  2. 258

    Photography & Realism

    Portraits, product, fashion

  3. 206

    Illustration & Art

    Anime, ink, watercolor, sci-fi

  4. 161

    UI & Interface

    Apps, pages, components, shots

  5. 149

    Infographics & Data Viz

    Infographics, flows, exploded views

  6. 103

    E-commerce & Products

    Hero shots, detail pages, ads

  7. 94

    Other Applications

    Composites and mashups

  8. 65

    Characters & Roles

    Character sheets, cards, action

  9. 51

    Branding & Logos

    Logos, brand systems, icons

  10. 37

    Scenes & Narrative

    Cinematic frames and storyboards

  11. 36

    Architecture & Space

    Interiors, renders, cities

  12. 20

    Documents & Publications

    Magazines, menus, notes

  13. 14

    Historical & Ancient Styles

    Dynasties, figures, Chinese chic

Total 1538

How to use

Three paths

  1. You already know the job

    Search by type, reuse the closest case, then rewrite.

  2. You only want “something nice”

    Confirm type, use, and style, then pick a direction.

  3. You need a custom image

    Lock subject, headline, ratio, and constraints.

Install

Add it to the agent you already use

terminal 
npx skills add wukongnotnull/image-inspirer
View GitHub

Or send the repository URL to your agent and ask it to install.

After install, you can say

  • Find infographic visualization examples
  • Generate a Chinese-style poster prompt
  • Give me a nice image

Pitfalls

Lock these five things first

  • Clear typePoster, UI, photography, or illustration—name the category.
  • Specific subjectName the person, object, or scene. “Nice” is not enough.
  • Locked-in stylePhotoreal, cinematic, minimal, or Chinese aesthetic—say it.
  • Complete parametersRatio, resolution, and composition belong in the prompt.
  • Readable textIf the image contains words, require them to be readable.

FAQ

Install and use

What is Image Inspirer?

A skill for agents. Search 1,538 cases by type, then rewrite from them.

How do I install it?

Run npx skills add wukongnotnull/image-inspirer. Or paste the GitHub URL to your agent.

How is this different from a prompt template?

Every case ships with a finished image, plus type-specific constraints.

Do I need an account or a network?

The skill and cases live locally. Rendering still depends on your image model.

Next

Start with one picture

See an image that works, then rewrite it into yours.

Install Image Inspirer View GitHub

See it. Learn it. Adapt it.