templates

📝 Prompt Templates

Overview

This directory contains reusable, optimized templates for creating various types of prompts and AI interactions. Unlike task-specific prompts, these templates provide structured frameworks that can be adapted to multiple use cases.

Philosophy

"The constraint is the feature" - All templates follow ADHD optimization principles, treating cognitive and token constraints as design advantages rather than limitations.

Available Templates

🎯 Universal Agent Card (universal-agent-card.md)

Purpose: Standardized template for creating AI agent personas and capabilities

Key Features:

• ADHD-optimized with strict word limits
• 5-second comprehension test
• Token optimization (<300 tokens)
• Visual hierarchy using emojis
• Built-in validation checklist

When to Use:

• Designing new AI agents or assistants
• Standardizing agent documentation
• Creating consistent agent libraries
• Rapid prototyping of agent capabilities

Example Output:

name: code-reviewer
description: Code quality guardian. Use for: reviews, refactoring, standards. 
            Examples: 'check PR' → detailed analysis.

---

🗺️ Visual Prompt Patterns (visual-prompt-patterns.md)

Purpose: Transform visual thinking structures (mind maps, diagrams) into text-based prompts

Patterns Included:

1. Radial/Mind Map structures
2. Hierarchical trees
3. Flow diagrams
4. Matrix structures
5. Layered context (onion model)
6. Network graphs
7. Quadrant analysis
8. Nested context boxes

When to Use:

• Complex system design requiring spatial organization
• Multi-faceted problems needing visual hierarchy
• Converting whiteboard sessions to prompts
• Preserving relationships between concepts

Example Pattern:

🎯 CENTER: API Design
├── 🔵 ENDPOINTS: CRUD operations
├── 🟢 VALIDATION: Input checking
├── 🟡 SECURITY: Auth & rate limiting
└── 🔴 ERRORS: Handling strategies

---

🎙️ ElevenLabs Conversational AI (elevenlabs-conversational-ai.md)

Purpose: Design natural, human-like conversational AI agents for voice interactions

Key Features:

• Emotional state detection patterns
• Natural language phrase banks
• Progressive information gathering
• TTS (Text-to-Speech) optimization
• LLM selection matrix
• Pre-deployment testing checklist

When to Use:

• Building voice-enabled AI assistants
• Customer service chatbots
• Interactive voice response (IVR) systems
• Any conversational AI requiring natural dialogue

Core Principle: "Conversation over script" - Agents should feel like talking to a knowledgeable colleague, not navigating a phone tree.

---

Template Selection Guide

Need	Recommended Template	Why
Create an AI assistant	Universal Agent Card	Structured, comprehensive agent design
Design complex system	Visual Prompt Patterns	Preserves spatial relationships
Build voice interface	ElevenLabs Conversational	Natural dialogue optimization
Optimize token usage	Universal Agent Card	Built-in token constraints
Convert diagrams to text	Visual Prompt Patterns	Multiple pattern options
Handle emotional users	ElevenLabs Conversational	Emotion detection built-in

How to Use Templates

1. Choose Your Template

Review the available templates and select based on your specific need.

2. Copy the Structure

Each template provides copy-paste ready structures. Start with the base template.

3. Fill in Your Content

Follow the guidelines and examples provided. Pay attention to:

• Word limits (where specified)
• Required vs optional sections
• Format specifications

4. Validate

Use built-in checklists to ensure quality:

• 5-second comprehension test
• Token count verification
• Completeness check

5. Iterate

Templates are starting points. Customize based on:

• Specific domain requirements
• User feedback
• Performance metrics

Best Practices

DO:

• ✅ Start with templates for consistency
• ✅ Respect word/token limits
• ✅ Use visual anchors (emojis) consistently
• ✅ Test with real scenarios
• ✅ Combine templates when needed

DON'T:

• ❌ Override core structure without reason
• ❌ Ignore validation checklists
• ❌ Add complexity unnecessarily
• ❌ Mix incompatible patterns
• ❌ Forget the 5-second rule

Creating New Templates

When adding new templates to this directory:

1. Follow the Format:

   # 🎯 Template Name
   *One-line description*
   
   ## Purpose
   ## Key Features
   ## When to Use
   ## Template Structure
   ## Examples
   ## Best Practices

2. Include Metadata:

   • YAML frontmatter
   • Version number
   • Creation date
   • Tags

3. Provide Examples:

   • Good vs bad comparisons
   • Real-world applications
   • Common variations

4. Add Validation:

   • Checklists
   • Success criteria
   • Common pitfalls

Integration with Frameworks

Templates work best when combined with frameworks:

Template	Pairs Well With	Result
Universal Agent Card	Vibecoding Archetypes	Philosophically-grounded agents
Visual Patterns	ADHD Framework	Maximum clarity and scanning
ElevenLabs	Context Engineering	Token-optimized conversations

Token Optimization Tips

All templates emphasize token efficiency:

1. Front-load critical info - Most important first
2. Use structure over prose - Lists beat paragraphs
3. Employ visual markers - Emojis = instant recognition
4. Progressive disclosure - Details only when needed
5. Explicit > Implicit - State everything clearly

Community Contributions

We welcome new templates! Consider submitting templates for:

• Specific industries (healthcare, finance, education)
• Specialized tasks (data analysis, creative writing)
• New AI platforms (OpenAI, Anthropic, Google)
• Novel interaction patterns

Resources

• ADHD Prompting Framework
• Context Engineering
• Vibecoding System
• Search Tools

---

Remember: Great templates make complex simple and simple powerful.