---

name: prompt-engineering description: LLM prompt optimization and design patterns. Use for crafting effective prompts, chain-of-thought, and AI integration.

🎯 Prompt Engineering Skill

Prompt Patterns

Zero-Shot

CopyTranslate the following text to Thai: "Hello, how are you?"

Few-Shot

CopyExamples:
Input: "Good morning" → Output: "สวัสดีตอนเช้า"
Input: "Thank you" → Output: "ขอบคุณ"
Input: "Goodbye" → Output: ?

Chain-of-Thought (CoT)

CopyLet's solve this step by step:
1. First, identify the problem
2. Break down the components
3. Apply relevant formula
4. Calculate the answer
5. Verify the result

---

Prompting Best Practices

Structure

Copy# Role
You are an expert Python developer.

# Context
We are building a REST API using FastAPI.

# Task
Create a user registration endpoint.

# Requirements
- Validate email format
- Hash password with bcrypt
- Return JWT token

# Output Format
Return complete, working code with comments.

Be Specific ✅

Copy❌ "Write code for a button"
✅ "Create a React button component that:
   - Has primary and secondary variants
   - Accepts onClick handler
   - Shows loading spinner when isLoading is true
   - Uses Tailwind CSS for styling"

---

Advanced Techniques

Self-Consistency

CopyGenerate 3 different solutions, then pick the best one.

Tree-of-Thought

CopyConsider multiple approaches:
Approach A: [describe]
Approach B: [describe]
Approach C: [describe]
Compare and select optimal solution.

RAG Pattern

CopyContext: [Retrieved documents]
Question: [User query]
Answer based ONLY on the context provided.

---

Temperature Guide

---

Checklist

• Define clear role/persona
• Provide sufficient context
• Be specific about output format
• Use examples when helpful
• Iterate and refine prompts