creativity
About
This skill provides AI-powered creative direction for marketing campaigns, offering structured recommendations across visual styles, platforms, voiceovers, and music. Developers can use it to generate briefs, plan content, and ensure platform-specific creative alignment. It functions as a searchable database of 55+ styles and 30+ campaign categories for ads, videos, and social content.
Quick Install
Claude Code
Recommended/plugin add https://github.com/majiayu000/claude-skill-registrygit clone https://github.com/majiayu000/claude-skill-registry.git ~/.claude/skills/creativityCopy and paste this command in Claude Code to install this skill
Documentation
Creativity Pro Max - Creative Direction Intelligence
Comprehensive creative direction guide for digital marketing campaigns. Contains 55 styles, 18 platforms, 12 voiceover types, 17 music genres, and 30 campaign category recommendations. Searchable database with priority-based recommendations.
When to Apply
Reference these guidelines when:
- Planning creative campaigns for any channel
- Selecting visual styles for ads/content
- Choosing voiceover/music direction
- Creating platform-specific content
- Developing creative briefs
Core Principle
92% prefer authentic content over polished ads. Prioritize real, relatable content over perfection.
Quick Style Selection
| Audience/Goal | Recommended Styles |
|---|---|
| Premium/Luxury | Minimalist Clean, Quiet Luxury, Dark Moody Cinematic |
| Tech/SaaS | Futuristic Sci-Fi, Gradient Smooth, Minimalist Tech |
| Youth/Gen Z | Raw Unpolished, Humor Comedic, Nostalgic 90s, Neon Cyber |
| Wellness/Organic | Earthy Organic, Emotional Storytelling, Pastel Soft |
| B2B/Finance | Authoritative Expert, Data Visualization, Corporate Modern |
| Community | Authentic UGC, Conversational Casual, Behind The Scenes |
How to Use This Skill
When user requests creative direction (campaign, ad, video, content), follow this workflow:
Step 1: Analyze User Requirements
Extract key information from user request:
- Campaign type: Product launch, brand awareness, lead gen, etc.
- Audience: Gen Z, Millennials, B2B, etc.
- Industry: SaaS, e-commerce, healthcare, etc.
- Platform: TikTok, Instagram, YouTube, etc.
Step 2: Generate Creative Brief (REQUIRED)
Always start with --creative-brief to get comprehensive recommendations:
python3 .claude/skills/creativity/scripts/search.py "<campaign_type> <audience> <industry>" --creative-brief [-c "Campaign Name"]
This command:
- Searches 5 domains in parallel (style, platform, voiceover, music, reasoning)
- Applies reasoning rules from
creative-reasoning.csvto select best matches - Returns complete brief: style, platform specs, voiceover, music, effects
- Includes anti-patterns to avoid
Example:
python3 .claude/skills/creativity/scripts/search.py "SaaS product launch gen z" --creative-brief -c "App Launch Campaign"
Step 3: Supplement with Detailed Searches (as needed)
After getting the creative brief, use domain searches for additional details:
python3 .claude/skills/creativity/scripts/search.py "<keyword>" --domain <domain> [-n <max_results>]
When to use detailed searches:
| Need | Domain | Example |
|---|---|---|
| More style options | style | --domain style "minimalist luxury" |
| Platform specs | platform | --domain platform "tiktok reels" |
| Voiceover direction | voiceover | --domain voiceover "authoritative expert" |
| Music recommendations | music | --domain music "ambient electronic" |
| Campaign guidance | reasoning | --domain reasoning "e-commerce conversion" |
Step 4: Load Reference Details (as needed)
For detailed implementation, load relevant references:
| Reference | Content | Use When |
|---|---|---|
| creative-styles-part1.md | Styles 1-25 with keywords, colors, effects | Selecting visual direction |
| creative-styles-part2.md | Styles 26-55 with keywords, colors, effects | Selecting visual direction |
| color-psychology.md | Color meanings, combinations, applications | Choosing color palette |
| voiceover-styles.md | 5 VO categories with delivery specs | Video/audio content |
| audio-music.md | Music trends, sonic branding | Video content, ads |
| visual-trends.md | 2025-26 design trends | Modern visual direction |
Search Reference
Available Domains
| Domain | Use For | Example Keywords |
|---|---|---|
style | Visual aesthetics | minimalist, ugc, cinematic, retro, luxury |
platform | Channel specs | tiktok, instagram, youtube, linkedin |
voiceover | VO direction | conversational, authoritative, playful |
music | Audio direction | ambient, upbeat, orchestral, trending |
reasoning | Campaign strategy | product launch, brand awareness, gen z |
Example Workflow
User request: "Create a TikTok campaign for Gen Z SaaS product launch"
Step 1: Analyze Requirements
- Campaign type: Product launch
- Audience: Gen Z
- Industry: SaaS
- Platform: TikTok
Step 2: Generate Creative Brief (REQUIRED)
python3 .claude/skills/creativity/scripts/search.py "gen z saas product tiktok" --creative-brief -c "SaaS Launch"
Step 3: Supplement with Detailed Searches (as needed)
# Get TikTok-specific guidelines
python3 .claude/skills/creativity/scripts/search.py "tiktok" --domain platform
# Get Gen Z style options
python3 .claude/skills/creativity/scripts/search.py "gen z authentic raw" --domain style
Output Formats
# ASCII box (default) - best for terminal display
python3 .claude/skills/creativity/scripts/search.py "e-commerce launch" --creative-brief
# Markdown - best for documentation
python3 .claude/skills/creativity/scripts/search.py "e-commerce launch" --creative-brief -f markdown
Key Stats
- Humor increases recall 33% (only 33% of ads use it)
- Color drives 85% purchase decisions
- Brand recognition +80% with consistent colors
- UGC: 28% higher engagement, 4x CTR, 50% lower CPA
Platform Quick Specs
| Platform | Ratio | Length | Hook | Style Note |
|---|---|---|---|---|
| TikTok | 9:16 | 15-60s | 3s | Raw, trending sounds |
| Instagram Reels | 9:16 | 15-90s | 3s | Polished-authentic |
| YouTube | 16:9 | 30s-10m | 30s | High production |
| 1:1, 4:5 | 30s-2m | 5s | Professional, value-first | |
| Stories | 9:16 | 15s | 1s | Immediate hook, swipe CTA |
Pre-Production Checklist
- Brand guidelines reviewed
- Audience persona confirmed
- Creative style approved from search results
- Platform specs verified
- Success metrics defined
- Hook in first 3 seconds
- Sound-off optimized (captions)
- Mobile-first designed
- A/B test variables identified
GitHub Repository
Related Skills
content-collections
MetaThis skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.
creating-opencode-plugins
MetaThis skill provides the structure and API specifications for creating OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It offers implementation patterns for JavaScript/TypeScript modules that intercept and extend the AI assistant's lifecycle. Use it when you need to build event-driven plugins for monitoring, custom handling, or extending OpenCode's capabilities.
sglang
MetaSGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.
evaluating-llms-harness
TestingThis Claude Skill runs the lm-evaluation-harness to benchmark LLMs across 60+ standardized academic tasks like MMLU and GSM8K. It's designed for developers to compare model quality, track training progress, or report academic results. The tool supports various backends including HuggingFace and vLLM models.
