git clone https://github.com/guia-matthieu/clawfu-skills
T=$(mktemp -d) && git clone --depth=1 https://github.com/guia-matthieu/clawfu-skills "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/ai-design/minimalist-image-director" ~/.claude/skills/guia-matthieu-clawfu-skills-minimalist-image-director && rm -rf "$T"
skills/ai-design/minimalist-image-director/SKILL.mdMinimalist Image Director
Art direction framework for generating warm minimalist photography via AI image generators (Flux, Midjourney, DALL-E). Separates compositional minimalism from emotional minimalism to avoid the "beautiful but sad" trap.
When to Use This Skill
- Generating hero images, card images, or blog illustrations for a website
- Creating a cohesive visual identity across 10+ AI-generated images
- Briefing AI image generators (Replicate/Flux, Midjourney, DALL-E) with emotional precision
- When previous minimalist attempts came back "too cold" or "too sad"
- Building a visual style guide for a brand's AI-generated photography
Methodology Foundation
Sources:
- Editorial photography principles (Annie Leibovitz, minimal lifestyle photography trend 2024-2026)
- Emotional Design (Don Norman, 2004) — visceral, behavioral, reflective processing
- Color psychology research — warm tones (2700-3000K) activate approach behaviors, cool tones trigger avoidance
- Neuroscience of visual-thermal perception — 80% of experiments show visual environment manipulation affects thermal perception (red-orange = warmth, green-blue = cold)
- Black Forest Labs official prompting guides (Flux 1.1 Pro, Flux 2)
- Kodak Portra 400 color science — the gold standard for warm skin tones in AI photography
Core Principle: Minimalism is about what you KEEP, not what you REMOVE. The fewer elements in a frame, the more each one must carry emotional weight. Empty space amplifies — it amplifies warmth just as easily as coldness.
Why This Matters: AI image generators default to "aesthetic minimalism" which reads as cold, clinical, lonely. The skill teaches how to direct warmth INTO minimal compositions, getting the clean look without the emotional void.
The Neuroscience: Warm colors trigger approach behaviors and lower cognitive vigilance — the viewer feels safe. Cool colors trigger alertness and avoidance. This is not aesthetic preference; it's how photoreceptors and neural pathways process visual information.
What Claude Does vs What You Decide
"Claude handles the prompt engineering. You bring the emotional truth."
| Claude handles | You provide |
|---|---|
| Translating emotional intent into Flux/MJ prompt syntax | The emotion each image must convey |
| Applying the 4-layer prompt architecture consistently | Brand palette and visual identity |
| Flagging prompt anti-patterns that produce sad/cold images | Validation — does this FEEL right? |
| Generating batch-consistent style prefixes | Subject matter and context for each image |
| Optimizing aspect ratios and technical parameters | Final selection between generated options |
Remember: AI can generate technically perfect minimalist images that feel completely wrong. Your gut reaction to the emotion is the quality gate, not the composition.
What This Skill Does
- Emotional Calibration - Defines the target emotion BEFORE writing any prompt
- 4-Layer Prompt Architecture - Style + Subject + Emotion + Anti-patterns in every prompt
- Batch Consistency - Creates a shared style prefix for visual cohesion across sets
- Anti-Pattern Detection - Flags words/directions that trigger cold/sad/clinical outputs
- Brand Alignment - Maps brand voice to visual language (warm brand = warm photos)
How to Use
Generate images for website cards
I need 3 card images for a child development psychologist website. Brand palette: cream, coral, warm earth tones. Cards: Motor Development, Emotional Development, Cognitive Development. Target emotion: hopeful, warm, possibility. Generator: Replicate Flux 1.1 Pro, 3:4 aspect ratio.
Create a cohesive blog image set
Generate prompts for 13 blog articles about parenting and child psychology. All images must feel like they're from the same photo shoot. Brand: warm, approachable, Latin American families. Avoid: clinical, sad, isolated figures, stock photo poses.
Fix images that came back too cold
These minimalist images came back sad/cold. Here's the original prompt: [prompt]. Keep the minimalist composition but make it emotionally warm. The image should make a parent feel "I want to be that parent" not "that's beautiful but lonely."
Instructions
When generating minimalist image prompts, follow this methodology precisely:
Step 1: Define the Emotional Target
Before writing ANY prompt, answer:
## Emotional Brief **This image should make the viewer feel:** ________________ **The viewer should want to:** ________________ **This is NOT about:** ________________ **Emotional quadrant:** WARM | ACTIVE --+-- CALM | COLD Target: [e.g., Warm + Calm = nurturing serenity]
Key principle: If you can't name the emotion in 2 words, the image will be vague.
Emotional vocabulary for warm minimalism:
| Warm + Active | Warm + Calm |
|---|---|
| Delight, play, discovery | Serenity, connection, trust |
| Courage, determination, pride | Presence, intimacy, safety |
| Freedom, possibility, wonder | Patience, tenderness, focus |
| Cold + Active (AVOID) | Cold + Calm (AVOID) |
|---|---|
| Anxiety, urgency, pressure | Loneliness, melancholy, void |
| Frustration, anger, defeat | Isolation, clinical, sterile |
Color psychology for emotional targeting:
| Color range | Emotional effect | Use when... |
|---|---|---|
| Cream/ivory (#FAF8F5) | Soft, approachable, comfortable base | Every warm minimalist image (background) |
| Terracotta (#C2704F) | Earthy warmth, trustworthiness, permanence | Brands in family, wellness, coaching |
| Warm pink (#FFC0CB) | Nurturing, gentleness, calming | Child development, early childhood |
| Golden/yellow (2700K) | Happiness, energy, sunlight, cozy | Golden hour shots, living room scenes |
| Orange tones | Friendly, fights depression, inviting | Social/community-oriented images |
| Sage/olive (muted green) | Natural, grounded, trustworthy | Earthy brand palettes alongside terracotta |
Step 2: Build the 4-Layer Prompt
Every prompt has exactly 4 layers:
## Prompt Architecture [LAYER 1: STYLE] Technical photography direction [LAYER 2: SUBJECT] Who/what is in the frame [LAYER 3: EMOTION] Specific emotional cues [LAYER 4: ANTI-PATTERNS] What to explicitly exclude
Layer 1 — Style Prefix (reuse across batch):
Warm minimalist photography. Soft natural light, shallow depth of field, [BRAND PALETTE TONES]. Candid moment, not posed. [DEMOGRAPHIC]. Shot on 85mm f/1.8 lens, Kodak Portra 400 film look, natural skin texture. No text, no logos, no watermarks. Warm color temperature.
Film stock trick: Adding "Kodak Portra 400" or "Kodak Portra 800" instantly introduces organic warmth, fine grain, and natural skin tones. This single phrase fights AI's default plastic/clinical rendering better than any other modifier.
HEX color precision (Flux 2+): Associate HEX codes with specific objects —
"The wall is #FAF8F5 cream" works better than "use #FAF8F5 in the image". Always pair HEX with a color name.
Key style levers:
| Lever | Warm direction | Cold direction (avoid) |
|---|---|---|
| Light | Soft natural, golden hour, window light | Studio flash, overhead fluorescent |
| Background | Cream, warm wood, sunlit room | White void, concrete, gray |
| Depth of field | Shallow (f/1.8) — intimacy | Deep (f/11) — documentary |
| Color temp | Warm (2700-3000K golden, 3200-4500K daylight) | Cool (6500K+) |
| Framing | Close, eye-level, inclusive | Wide, above, distant |
| Film stock | Kodak Portra 400, Fujifilm Pro 400H | No film reference (digital default) |
| Texture | "natural skin texture, pores, freckles" | "smooth skin, flawless" (= plastic) |
Layer 2 — Subject:
A [age] [demographic] child [action verb + specific detail]. [Body language cue]. [One environmental detail].
Rules:
- One action verb, one detail (not a paragraph)
- Body language > facial expression for Flux
- One environmental detail grounds the scene (wooden floor, sunlit garden)
- "Mid-action" > "posing" (hands placing a block > holding a block)
- Always specify demographics — Flux has training biases and will default if unspecified
Body language science — warm vs cold signals:
| Warm signals (USE) | Cold signals (AVOID) |
|---|---|
| Duchenne smile (eyes squeezing + mouth) | Fake smile (mouth only, no eye engagement) |
| Direct eye contact, maintained gaze | Eyes turned to side or downward |
| Open posture, arms uncrossed | Arms crossed over chest (barrier) |
| Relaxed, self-assured stance | Rigid posture, head tilted back |
| Physical proximity or gentle touch | Distance between subjects |
| Leaning in, at eye level | Leaning away, looking from above |
Layer 3 — Emotion Injection:
[Mood word]. [Light descriptor that reinforces mood].
Proven emotion-to-prompt mappings:
| Target emotion | Prompt language |
|---|---|
| Joy/delight | "pure delight", "laughing", "arms wide" |
| Connection | "eye contact", "faces close", "at eye level" |
| Curiosity | "deeply focused", "hands mid-action", "slight smile" |
| Safety | "gentle touch", "both at ease", "calm conversation" |
| Pride | "standing tall", "determination", "just accomplished" |
| Possibility | "looking up/ahead", "about to", "the moment before" |
Layer 4 — Anti-Pattern Blockers:
Words that trigger cold/sad in AI generators:
| NEVER use | Use instead |
|---|---|
, , | |
, | |
, | , |
, , | , , |
, | , |
, , | , , |
, (alone) | , , |
| |
, , | , , |
, | , , |
, , | , , |
Negative prompt suffix (append to every prompt for Flux):
--no plastic skin, glossy surfaces, artificial lighting, airbrushed, sterile, clinical, 3D render, CGI, harsh shadows, cool tones
Step 3: Validate Before Generating
Before sending to the API, run this checklist:
## Pre-Generation Checklist - [ ] Can I name the target emotion in 2 words? - [ ] Does the subject have an ACTION (not just a state)? - [ ] Is there at least one warmth signal (light, touch, smile, color)? - [ ] Are there zero isolation signals (alone, empty, quiet)? - [ ] Is the demographic consistent with the brand? - [ ] Does the style prefix match the batch?
Step 4: Evaluate Generated Images
Rate each generated image:
## Image Evaluation **Emotional hit?** [Yes/No] — Does it trigger the target emotion within 2 seconds? **Warmth level:** [1-5] — 1=clinical, 3=neutral, 5=cozy **Brand fit:** [Yes/No] — Does it feel like it belongs on the brand's site? **Minimalism quality:** [Clean/Busy] — Is the composition uncluttered? **Stock photo test:** [Pass/Fail] — Would you mistake this for generic stock? If emotional hit = No → rewrite Layer 3 (emotion) first If warmth < 3 → add warm lighting/color cues to Layer 1 If stock photo test = Fail → make Layer 2 more specific (exact age, exact action)
Step 5: Iterate on Failures
Common failure patterns and fixes:
| Problem | Root cause | Fix |
|---|---|---|
| Image is beautiful but sad | Isolation signals in prompt | Add connection (person+person or person+activity) |
| Image is warm but generic | Subject too vague | Add one hyper-specific detail ("wooden blocks" not "toys") |
| Image feels like stock | "Looking at camera" or "smiling" | Switch to candid mid-action |
| Inconsistent batch style | Style prefix varies | Copy-paste exact same Layer 1 |
| Wrong age/demographic | Generator defaults | Be explicit: "4-year-old", "Latin American" |
Platform-Specific Guide: Flux 1.1 Pro
Flux is the primary recommended generator for warm minimalist photography. These rules are Flux-specific.
Syntax Rules
- Write like you're talking to a photographer, not typing keywords into a search engine
- Flux uses natural language — full sentences, NOT comma-separated keyword lists
- 30-80 words is the sweet spot — under 200 tokens for optimal quality
- NO prompt weights —
is ignored. Use "with emphasis on X" instead(text:1.5) - Specify demographics explicitly — Flux has training biases and will default to Western/white/young if unspecified
Recommended Structure
[Subject + specific details] performing [action] in [environment], [lighting: direction + quality + time of day], shot on [camera/lens], [film stock reference], [composition note], [emotional tone], natural skin texture, candid moment --no [anti-patterns]
Camera Filename Trick (Raw Mode)
For hyper-realistic results, prefix the prompt with a camera filename:
— iPhone snapshot realismIMG_2847.HEIC
— Canon DSLR raw aestheticIMG_5847.CR2
— Sony camera lookDSC03291.ARW
Example:
IMG_2847.HEIC A mother and child laughing together at a café table, soft window light, slightly grainy, Kodak Portra 400 warmth, authentic moment
Raw Mode
Flux Raw Mode adds natural imperfections: skin texture, realistic lighting, natural asymmetry. Best for portraits and lifestyle. Use minimal prompts — let the realism engine do its work.
Aspect Ratios
- 3:4 — Portrait cards, service pages (recommended for website cards)
- 4:5 — Instagram, vertical lifestyle
- 16:9 — Blog headers, hero images
- 1:1 — Social media, profile images
Examples
Example 1: Child Development Website Cards
Context: 3 service area cards for a child psychologist's website. Brand is warm, approachable, cream/coral palette. Target: parents who want to understand their child's development.
Input:
Generate 3 card images for Desarrollo Motor, Emocional, Cognitivo. Warm minimalist, Flux 1.1 Pro, 3:4 ratio.
Output:
Style prefix (shared):
Warm minimalist photography. Soft natural light, shallow depth of field, warm cream and beige tones. Candid moment, not posed. Latin American family. Shot on 85mm f/1.8 lens. No text, no logos, no watermarks. Warm color temperature, hopeful joyful mood.
Card 1 — Desarrollo Motor (Emotion: joyful possibility)
[STYLE PREFIX] A 4-year-old child mid-jump on warm wooden floor, arms wide open, barefoot. Pure delight and concentration on face. Soft motion blur on hands. Single subject, clean warm background. The moment right before landing — suspended possibility.
Card 2 — Desarrollo Emocional (Emotion: intimate connection)
[STYLE PREFIX] A mother crouching at eye level with her 5-year-old daughter, faces close, gentle smiles, eye contact. Golden hour light. Intimate connection between adult and child. Minimal warm background, soft cream tones.
Card 3 — Desarrollo Cognitivo (Emotion: serene curiosity)
[STYLE PREFIX] A 6-year-old child sitting cross-legged on the floor, deeply focused building a tall wooden block tower. Hands mid-action placing a block. Calm concentration on face, slight smile. Soft overhead natural light, earth tones.
Why this works: Each image has one clear emotion, one specific action, warm lighting, and no isolation signals. The shared style prefix ensures visual cohesion across the set.
Example 2: Fixing a "Beautiful but Sad" Image
Context: A minimalist illustration of a woman crouching alone was generated for a "limits without yelling" blog post. The art direction was "minimalist" but the result felt melancholy.
Input:
The line art minimalist image came back too sad. It's a woman crouching alone in black and white. Fix it.
Analysis of failure:
→ isolation signalalone
→ removes warmthblack and white
with no context → reads as defeatedcrouching- No other person or activity → loneliness
Fixed prompt:
Warm minimalist photography. Soft natural light, shallow depth of field, warm cream and beige tones. Candid moment, not posed. Shot on 85mm f/1.8 lens. No text, no logos, no watermarks. Warm color temperature. A mother and 4-year-old child sitting face to face on a couch, mother holding both of child's hands gently, calm conversation. Both at ease. Warm living room light filtering through curtains.
What changed:
- Solo → pair (connection defeats loneliness)
- B&W → warm tones (color = life)
- Crouching → sitting face to face (equals, not defeated)
- Added environmental warmth (couch, living room light)
Skill Boundaries (Frontier Recognition)
This skill excels for:
- Generating cohesive sets of 3-20+ images with consistent style
- Warm/approachable brands (family, wellness, education, coaching)
- Photorealistic AI generators (Flux, Midjourney v6+, DALL-E 3)
This skill is NOT ideal for:
- Brands that WANT cold/clinical aesthetics (tech, luxury, medical) → Adjust Layer 1 accordingly
- Abstract/conceptual images (infographics, diagrams) → Use
skill insteaddata-visualizer - Product photography → Requires different prompt architecture
- Illustration styles (watercolor, vector, line art) → Adapt Layer 1 for illustration-specific generators
Quality Checkpoints
Before accepting the output, verify:
- 2-second gut check: does the image make you feel the target emotion?
- Warmth score >= 4 out of 5
- No accidental isolation signals in the composition
- Consistent with the rest of the batch (same light, same tones)
- Would NOT be mistaken for generic stock photography
Iteration Guide
"The first output is a starting point, not a destination."
Recommended Iteration Pattern
| Pass | Focus | Questions to Ask |
|---|---|---|
| 1st | Emotion | "Does this FEEL right within 2 seconds?" |
| 2nd | Specificity | "Is this too generic? What one detail would make it unique?" |
| 3rd | Consistency | "Does this match the rest of the set?" |
| 4th | Brand | "Would the client recognize this as THEIR brand?" |
Useful Follow-up Prompts
- "The image is warm but feels generic. Add one hyper-specific detail to the subject."
- "The emotion is too [intense/subtle]. Dial it [down/up] by adjusting the body language."
- "The background is too busy. Simplify to [one element] and increase the bokeh."
- "This looks like stock. Make the child's action more specific — what exactly are their hands doing?"
Checklists & Templates
Batch Brief Template
## Image Batch Brief **Brand:** ________________ **Palette:** ________________ **Demographic:** ________________ **Generator:** Flux 1.1 Pro / Midjourney v6 / DALL-E 3 **Aspect ratio:** ________________ **Number of images:** ________________ ### Style Prefix (copy-paste for ALL prompts) [Write once, use everywhere] ### Per-Image Briefs | # | Subject | Target emotion (2 words) | Specific action | |---|---------|--------------------------|-----------------| | 1 | | | | | 2 | | | | | 3 | | | |
Red Flags Checklist
## Warning Signs in Your Prompts - [ ] Any word from the "NEVER use" list (alone, empty, dark, moody, studio) - [ ] Subject has no action verb (just standing/sitting with no activity) - [ ] No warmth signal (no mention of light quality, color temperature, or human connection) - [ ] Demographic not specified (generator will default to its biases) - [ ] More than 3 adjectives in a row (over-direction = generic output) - [ ] Prompt longer than 80 words (Flux sweet spot is 30-80 words, degrades past 200 tokens)
References
Core Methodology
- Norman, Don. "Emotional Design" (2004) - Three levels of design processing (visceral, behavioral, reflective)
- Annie Leibovitz. Masterclass on Portrait Photography - Light as emotion
- Kittl x Savee. "2026 Design Trends Report" - Warm minimalism as dominant trend
Flux & AI Image Generation
- Black Forest Labs Prompting Guide - Official Flux prompt best practices
- Flux 2 Prompting Guide (fal.ai) - JSON/HEX color structured prompts
- Flux Raw Mode Guide (Segmind) - Natural imperfections
- Official BFL Skills Repo - Prompting patterns per AgentSkills spec
- Kodak Portra 400 Midjourney Style (Midlibrary) - Film stock reference
Color Psychology & Neuroscience
- Color Psychology in Photography (Skylum) - Warm/cold tones and emotional response
- Visual Environment & Thermal Perception (ScienceDirect) - 80% of experiments show visual → thermal link
- Cold Temperatures in Photos Increase Cognitive Control (ScienceDaily) - Warm → relaxed, cool → alert
Photography Technique
- Photographer's Essential Guide to Body Language (SLR Lounge) - Warm/cold posture cues
- Photography Composition Definitive Guide (Anton Gorlin) - Frame-within-frame for intimacy
- Fixing Plastic AI Skin (Rezience) - Negative prompts for realistic texture
- 120+ Stable Diffusion Negative Prompts (ClickUp) - Anti-pattern word lists
Warm Minimalism Trend
- Warm Minimalism Trend 2026 (Good Housekeeping) - "Less but better"
- Earthy Color Palette Ideas (Rose Benedict Design) - Brand application of earth tones
Art Direction Methodology
- How to Write a Photoshoot Brief (Milanote) - Emotional objectives in briefs
- Creative Briefs for Photographers (VSCO) - SMART emotional criteria
Related Skills
- design-trends-2026 - Current visual trends to align with
- brand-strategy - Brand foundation before visual direction
- image-batch - Post-processing (resize, compress, WebP)
Skill Metadata
name: minimalist-image-director category: ai-design subcategory: art-direction version: 2.0 author: GUIA source_expert: Editorial Photography + Don Norman (Emotional Design) + Color Psychology + Neuroscience of Visual Perception + Black Forest Labs (Flux) source_work: null difficulty: intermediate mode: centaur estimated_value: Art director day rate (~500-800 EUR/day) tags: [image-generation, art-direction, minimalism, flux, replicate, midjourney, brand-photography, emotional-design, color-psychology, warm-minimalism, kodak-portra] created: 2026-02-12 updated: 2026-02-12
This skill is part of the GUIA Premium Marketing Skills Library — the 201 layer that bridges AI basics and technical implementation.