Craft effective prompts for AI models (Claude, GPT, Gemini, Midjourney, DALL-E, Stable Diffusion, Imagen, Flux) using proven patterns.

What This Skill Does

Challenge: Getting high-quality outputs from AI models requires understanding model-specific syntax, prompt structure, and iterative refinement techniques. Generic prompts produce generic results.

Solution: The AI Artist skill provides prompt engineering patterns for LLMs (Claude, GPT, Gemini), image generators (Midjourney, DALL-E, Stable Diffusion, Imagen, Flux), and video generators (Veo, Runway). Includes style keywords, negative prompts, few-shot examples, and domain-specific patterns.

Activation

Implicit: Activates when agents generate prompts for AI generation tasks.

Explicit: Use natural language to activate: “Activate ai-artist skill for prompt optimization”

Capabilities

1. LLM Prompt Structure

Consistent format for text generation models.

Universal Template:

[Role] You are a {expert type} specializing in {domain}.
[Context] {Background information and constraints}
[Task] {Specific action to perform}
[Format] {Output structure - JSON, markdown, list}
[Examples] {1-3 few-shot examples if needed}

Example - Marketing Copy:

[Role] You are a conversion copywriter specializing in SaaS landing pages.
[Context] Target audience: B2B project managers frustrated with email-based tracking.
[Task] Write above-the-fold hero section (headline + subheadline + CTA).
[Format] Markdown with labeled sections.

Guide: references/llm-prompting.md

2. Image Generation Prompts

Model-specific syntax for image generation.

Universal Structure:

[Subject] {main subject with details}
[Style] {art style, medium, artist references}
[Composition] {framing, angle, lighting}
[Quality] {resolution, quality modifiers}
[Negative] {what to avoid - if supported}

Example - Product Mockup:

Modern SaaS dashboard interface, clean minimalist design, blue and white color scheme, soft shadows, centered composition, professional photography, 4k resolution --ar 16:9 --style raw

Model Syntax Guide: references/image-prompting.md

3. Model-Specific Optimization

Leverage each model’s strengths and syntax.

Model Comparison:

Syntax guide: references/image-prompting.md

Prerequisites

• Access to target AI model (API keys or platform accounts)
• Clear creative brief or brand guidelines

Configuration

No configuration needed. Skill provides prompt templates and examples.

Best Practices

1. Clarity Beats Cleverness “A blue button” still beats “A rectangular azure interactive element”. Be specific and direct.

2. Iterate In Small Steps Change one variable at a time (subject, style, composition). Big rewrites make it hard to identify what works.

3. Use References When Available “In the style of Apple product photography” is clearer than manually describing the style.

Common Use Cases

Use Case 1: Generate Product Hero Image

Scenario: Create hero image for SaaS landing page.

Workflow:

1. Define core elements: Dashboard UI interface, professional setting, clean aesthetic
2. Draft prompt:

   Modern SaaS dashboard interface displayed on MacBook Pro,
   sitting on white desk with coffee cup, soft natural light from window,
   minimalist office background, professional photography,
   depth of field, 8k resolution --ar 16:9

3. Generate with Imagen 4 or Midjourney
4. Refine if needed (adjust lighting, composition)

Iteration: Typically needs 2-3 generations to find winner.

Use Case 2: Write Email Campaign Copy

Scenario: Write re-engagement email for inactive trial users.

Workflow:

1. Structure prompt:

   [Role] Conversion copywriter for B2B SaaS
   [Context] Inactive trial users 14+ days, tried project management tool
   [Task] Write re-engagement email using PAS formula
   [Format] Subject line + 3-paragraph body + CTA
   [Examples]
   Subject: We miss you [Name]
   Body: Problem → Agitate → Solution structure

2. Generate with Claude or GPT
3. A/B test 3 variations

Outcome: High-converting email copy following proven formula.

Troubleshooting

Issue: Generated images don’t match brand Solution: Add specific brand colors, style keywords, and reference existing assets. Load brand-guidelines skill for brand context.

Issue: LLM output too verbose or generic Solution: Add format constraints (“Max 50 words”, “Use bullet points”). Include few-shot examples of desired output.

Issue: Image generation keeps failing safety filters Solution: Review model’s content policy. Avoid ambiguous terms that trigger filters. Use alternative phrasing.

Related Skills

• AI Multimodal - Image/video generation execution
• Copywriting - Copy formulas for LLM prompts
• Creativity - Creative direction and style selection
• Brand Guidelines - Brand-aligned prompts

Related Commands

• /design/good - Generate images with optimized prompts
• /content/good - Generate high-quality copy
• /ask - Get prompt optimization advice