/image-generation

name: image-generation
description: Use this skill when the user requests to generate, create, imagine, or visualize images including characters, scenes, products, or any visual content. Supports structured prompts and reference images for guided generation.

Image Generation Skill

Overview

This skill generates high-quality images using structured prompts and a Python script. The workflow includes creating JSON-formatted prompts and executing image generation with optional reference images.

For **scientific illustrations** such as graphical abstracts, mechanism illustrations, concept explainers, or paper cover art, read `/mnt/skills/public/scientific-image-prompting/SKILL.md` first. That skill decides whether the request is allowed to use AIGC at all and defines the paper-ready prompt contract.

Core Capabilities

• Create structured JSON prompts for AIGC image generation
• Support multiple reference images for style/composition guidance
• Generate images through automated Python script execution
• Handle various image generation scenarios (character design, scenes, products, etc.)

Workflow

Step 1: Understand Requirements

When a user requests image generation, identify:

• Subject/content: What should be in the image
• Style preferences: Art style, mood, color palette
• Technical specs: Aspect ratio, composition, lighting
• Reference images: Any images to guide generation
• You don't need to check the folder under `/mnt/user-data`

Step 1.5: Reject Wrong Scientific Route

If the request is actually a real data figure, do **not** use this skill as the primary renderer.

Examples that must stay out of AIGC image generation:

• ROC / PR curves
• heatmaps
• volcano plots
• PCA / UMAP / t-SNE
• confusion matrices
• bar charts / line charts / histograms driven by real values

Route those requests to validated plotting workflows instead.

Step 2: Create Structured Prompt

Generate a structured JSON file in `/mnt/user-data/workspace/` with naming pattern: `{descriptive-name}.json`

Step 3: Execute Generation

Call the Python script:

python /mnt/skills/public/image-generation/scripts/generate.py \
  --prompt-file /mnt/user-data/workspace/prompt-file.json \
  --reference-images /path/to/ref1.jpg /path/to/ref2.png \
  --output-file /mnt/user-data/outputs/generated-image.jpg \
  --aspect-ratio 16:9

Parameters:

• `--prompt-file`: Absolute path to JSON prompt file (required)
• `--reference-images`: Absolute paths to reference images (optional, space-separated)
• `--output-file`: Absolute path to output image file (required)
• `--aspect-ratio`: Aspect ratio of the generated image (optional, default: 16:9)
• `--provider`: Optional image provider override (`google-ai-studio` or `openai-compatible`)
• `--model`: Optional image model override
• `--base-url`: Optional OpenAI-compatible image base URL override
• `--image-size`: Optional `1K`, `2K`, or `4K`
• `--output-mime-type`: Optional `image/png` or `image/jpeg`
• `--scientific-mode`: Enables scientific illustration guardrails and 4K/PNG defaults
• `--manifest-file`: Optional manifest output path for generation audit
• `--draft-mode`: Lower-cost mode when no explicit model is provided

[!NOTE] Do NOT read the python file, just call it with the parameters. If the user has already configured an active image provider in Settings, use that provider/model by default and do not hardcode Gemini unless the user explicitly asked for Google AI Studio.

Scientific Illustration Mode

When the request is a scientific illustration that passed the route check:

1. Read `/mnt/skills/public/scientific-image-prompting/SKILL.md` 2. Produce `prompt.json` using that contract 3. Generate the image in scientific mode 4. Keep the final output conceptual, not quantitative

Recommended command:

python /mnt/skills/public/image-generation/scripts/generate.py \
  --prompt-file /mnt/user-data/outputs/prompt.json \
  --output-file /mnt/user-data/outputs/scientific-illustration-4k.png \
  --manifest-file /mnt/user-data/outputs/generation_manifest.json \
  --aspect-ratio 16:9 \
  --scientific-mode \
  --image-size 4K \
  --output-mime-type image/png

Only add `--provider`, `--model`, or `--base-url` when the user explicitly wants to override the configured default provider.

Character Generation Example

User request: "Create a Tokyo street style woman character in 1990s"

Create prompt file: `/mnt/user-data/workspace/asian-woman.json`

{
  "characters": [{
    "gender": "female",
    "age": "mid-20s",
    "ethnicity": "Japanese",
    "body_type": "slender, elegant",
    "facial_features": "delicate features, expressive eyes, subtle makeup with emphasis on lips, long dark hair partially wet from rain",
    "clothing": "stylish trench coat, designer handbag, high heels, contemporary Tokyo street fashion",
    "accessories": "minimal jewelry, statement earrings, leather handbag",
    "era": "1990s"
  }],
  "negative_prompt": "blurry face, deformed, low quality, overly sharp digital look, oversaturated colors, artificial lighting, studio setting, posed, selfie angle",
  "style": "Leica M11 street photography aesthetic, film-like rendering, natural color palette with slight warmth, bokeh background blur, analog photography feel",
  "composition": "medium shot, rule of thirds, subject slightly off-center, environmental context of Tokyo street visible, shallow depth of field isolating subject",
  "lighting": "neon lights from signs and storefronts, wet pavement reflections, soft ambient city glow, natural street lighting, rim lighting from background neons",
  "color_palette": "muted naturalistic tones, warm skin tones, cool blue and magenta neon accents, desaturated compared to digital photography, film grain texture"
}

Execute generation:

python /mnt/skills/public/image-generation/scripts/generate.py \
  --prompt-file /mnt/user-data/workspace/cyberpunk-hacker.json \
  --output-file /mnt/user-data/ou