Skip to content
image logo

Image

image

ImageAgent - Generate EVOLEA Brand Images with Reinforcement Learning

SKILL.md

Full skill instructions

EVOLEA Image Generation with Reinforcement Learning

Invoke with: /​image or use automatically when generating images for the website Last Updated: 2025-12-27

This skill provides an intelligent image generation system that learns from user feedback to produce increasingly better, brand-consistent images for EVOLEA.


MCP Server (Claude Desktop / Claude App)

The image generation system is available as an MCP server for use with Claude Desktop and Claude App.

Available Tools

ToolDescription
generate_imageGenerate a single image with EVOLEA branding
generate_ab_comparisonGenerate A/​B comparison for training (2 options + grid)
list_generated_imagesList recent generated images by category
get_brand_prompt_templateGet prompt templates for specific image types
publish_imagePublish image to GitHub for public URL access
get_training_guideGet the full training guide documentation

Setup (Claude Desktop)

Add to %APPDATA%\Claude\claude_desktop_config.json:

{
  "mcpServers": {
    "evolea-images": {
      "command": "python",
      "args": ["C:/​Users/​christoph/​evolea-website/​scripts/​mcp_image_server.py"],
      "env": {
        "GOOGLE_API_KEY": "your-gemini-api-key"
      }
    }
  }
}

Quick Commands in Claude Desktop

"Generate an image of children doing art together"
"Create an A/​B comparison for Mini Projekte hero"
"Show me recent training images"
"Publish the image to GitHub so I can see it on my phone"

How It Works

┌─────────────────────────────────────────────────────────────┐
│  1. LOAD LEARNINGS from LEARNINGS.md                        │
│     - Apply positive patterns (+3, +1 boosts)               │
│     - Include negative patterns in exclusions (-1, -3)      │
├─────────────────────────────────────────────────────────────┤
│  2. GENERATE OPTIONS (A and B)                              │
│     - Option A: Base prompt with current learnings          │
│     - Option B: Base + additional style modifiers           │
│     - Save to: public/​images/​generated/​training/            │
├─────────────────────────────────────────────────────────────┤
│  3. CREATE COMPARISON GRID                                  │
│     - Side-by-side A|B image for easy comparison            │
├─────────────────────────────────────────────────────────────┤
│  4. USER SELECTS                                            │
│     - A, B, or Neither with feedback                        │
├─────────────────────────────────────────────────────────────┤
│  5. UPDATE LEARNINGS                                        │
│     - Boost winning patterns                                │
│     - Record negative patterns from feedback                │
│     - Persist to LEARNINGS.md                               │
├─────────────────────────────────────────────────────────────┤
│  6. PUBLISH (optional)                                      │
│     - Push to GitHub → Cloudflare Pages                     │
│     - Get public URL for mobile access                      │
├─────────────────────────────────────────────────────────────┤
│  7. ITERATE until satisfied                                 │
└─────────────────────────────────────────────────────────────┘

Command Line Usage

Generate with Training Loop

# Start a training session for a specific image
python scripts/​generate_image.py "children ages 5-8 creating art" \
  --name mini-projekte-hero \
  --category programs \
  --training

Manual A/​B Generation

# Generate 2 options with comparison grid
python scripts/​generate_image.py "your prompt" \
  --name image-name \
  --count 2 \
  --comparison-grid

Fully Automated (with Claude selection)

python scripts/​generate_image.py "your prompt" \
  --name image-name \
  --auto-select

Key Files

FilePurpose
scripts/​mcp_image_server.pyMCP server for Claude Desktop
scripts/​generate_image.pyCore image generation script
LEARNINGS.mdPersistent style preferences with scores
TRAINING-GUIDE.mdUser guide for iterative training
training-log.jsonFull history of all training sessions
style-profiles/​*.mdPre-defined style templates

EVOLEA Style Guidelines

Central European Children (Ages 5-8)

  • Skin tones: Light/​fair with warm undertones
  • Hair colors: Blonde, light brown, auburn, brown
  • Features: Soft, rounded, friendly expressions
  • NO: Religious symbols, American cultural elements

Whimsical Style (User's Preferred)

  • Soft dreamy watercolor textures
  • Floating, layered clouds in spectrum colors
  • Ethereal, magical atmosphere
  • Delicate butterflies or unicorns
  • Pastel gradient backgrounds (lavender, mint, coral, cream)

Color Palette

ColorHexUsage
Magenta#DD48E0Primary accents
Mint#7BEDD5Fresh, nature
Lavender#CD87F8Calm, creative
Coral#FF7E5DEnergy, warmth
Cream#FFFBF7Backgrounds

NEVER Include

  • Puzzle piece symbols
  • Clinical/​medical settings
  • Religious symbols on children
  • American cultural elements (yellow school buses)
  • Photorealistic style
  • Dark or muted colors

Teacher Illustrations

Based on EVOLEA team members:

NameDescription
Gianna SpiessFemale, professional, warm, M.Sc. BCBA
Annemarie EliasFemale, friendly, approachable, M.Sc. BCBA
Christoph JennyMale, supportive, engaged
Alexandra AleksicFemale, young, energetic, B.Sc.

Prompt Templates

Program Hero

Children aged [AGE] [ACTIVITY] in [SETTING].
Swiss/​Central European children with light skin and varied natural hair colors.
[SPECIFIC DETAILS].
Mood: [EMOTION]. Dominant colors: [COLORS].
Soft watercolor children's book illustration style with delicate butterflies.

Abstract/​Decorative

Soft dreamy [SUBJECT] in whimsical watercolor style.
Layered colorful clouds in lavender, mint, coral, and cream.
Ethereal atmosphere with gentle sparkles.
Delicate butterflies floating softly.
Dominant colors: [COLORS]. Mood: magical, warm, inviting.

Backend Configuration

The image generation always uses Gemini 3 Pro:

ModelID
Gemini 3 Progemini-3-pro-image-preview

This is configured in scripts/​generate_image.py via CONFIG.gemini_model.

Replicate (Flux) is available as a fallback if Gemini is unavailable in your region.


Related Files

  • TRAINING-GUIDE.md - Detailed training workflow guide
  • LEARNINGS.md - Current learned preferences
  • style-profiles/ - Style profile templates
  • .claude/​skills/​Design skills/​illustrations.md - Illustration guidelines

More skills from majiayu000

xiaohongshu logo
majiayu000/claude-arsenal

xiaohongshu

xiaohongshu

286 148
View
agent-task-conductor logo
majiayu000/claude-skill-registry

agent-task-conductor

Conduct multi-agent task orchestration and workflow coordination.

663 1
View
conductor-setup logo
majiayu000/claude-skill-registry

conductor-setup

Initialize project with Conductor artifacts (product definition,

663 1
View
animation-designer logo
majiayu000/claude-skill-registry

animation-designer

Expert in web animations, transitions, and motion design using Framer Motion and CSS

663 1
View
diagramming logo
majiayu000/claude-skill-registry

diagramming

Creates Mermaid and ASCII diagrams for flowcharts, architecture, ERDs, state machines, mindmaps, and more. Use when user mentions diagram, flowchart, mermaid, ASCII diagram, text diagram, terminal diagram, visualize, C4, mindmap, architecture diagram, sequence diagram, ERD, or needs visual docume...

663 1
View
h3-pg logo
majiayu000/claude-skill-registry-data

h3-pg

PostgreSQL bindings for H3 hexagonal grid system. Use when working with H3 cells in Postgres, including spatial indexing, geometry/geography integration, and raster analysis.

23 1
View
conductor-development logo
majiayu000/claude-skill-registry

conductor-development

Context-Driven Development skill for projects using Conductor. Use this skill when you detect a `conductor/` directory in the project, when working on tasks defined in a `plan.md` file, or when the user asks about tracks, specs, or plans. Automatically applies TDD workflow, tracks task completion...

663 1
View
conductor-status logo
majiayu000/claude-skill-registry

conductor-status

Display project status, active tracks, and next actions

663 1
View
dockerization logo
majiayu000/claude-skill-registry

dockerization

Official Stakpak application containerization standard operating procedure, a step-by-step guidline to properly dockerize applications. This is a rule book curated by the Stakpak Team.

663 1
View

Popular AI tools

Kaiber logo
Video

Kaiber

Generate, edit, and beat-sync AI video with leading models in one workspace.

Paid
View
Vimcal logo
Productivity

Vimcal

The world's fastest calendar for remote work

Free
View

Transform Your Design with AI Designer by ImgCreator.ai

Freemium
View
Akool AI logo
Content & writing

Akool AI

Revolutionizing Video Production with AI-Powered Creativity

Paid
View

Extend an image past the frame and let AI fill the new aspect ratio.

Freemium
View
StarByFace logo
Security

StarByFace

Discover your celebrity doppelgänger with StarByFace!

Free
View
C

ChainClarity explains 700+ crypto whitepapers in plain English, with layered summaries, comparisons, research tools, alerts, and a $4.99 Pro plan.

Freemium
View
Opus Clip logo
Coding & apps

Opus Clip

Opus.ai: Revolutionize Your Web Experience

Free
View