anthropic-prompt-engineer

# Anthropic Prompt Engineer

Master the art and science of prompt engineering with Anthropic's proven techniques. Generate new prompts from scratch or improve existing ones using best practices for Claude AI models (Claude 4.x, Sonnet, Opus, Haiku).

## What This Skill Does

Helps you create and optimize prompts for Claude AI using Anthropic's official techniques:
- **Generate new prompts** - Build effective prompts from requirements
- **Improve existing prompts** - Optimize prompts for better results
- **Apply best practices** - Use proven techniques from Anthropic
- **Avoid common mistakes** - Prevent hallucinations and unclear outputs
- **Optimize for Claude 4.x** - Leverage latest model capabilities
- **Structure complex prompts** - Build multi-step, production-ready prompts

## Why Prompt Engineering Matters

**Without proper prompting:**
- Inconsistent or incorrect outputs
- Hallucinations and made-up information
- Unclear or verbose responses
- Wasted tokens and API calls
- Poor performance on complex tasks
- Difficulty reproducing results

**With engineered prompts:**
- Precise, reliable outputs
- Factual, grounded responses
- Clear, formatted results
- Efficient token usage
- Excellent complex task performance
- Reproducible, production-ready results

## Quick Start

### Generate a New Prompt

```
Using the anthropic-prompt-engineer skill, create a prompt that:
- Extracts structured data from customer emails
- Returns JSON format
- Handles missing information gracefully
- Includes 2 examples
```

### Improve an Existing Prompt

```
Using the anthropic-prompt-engineer skill, improve this prompt:

"Analyze this code and tell me if there are bugs"

Make it more effective using Anthropic's best practices.
```

## Core Techniques Summary

### 1. Be Clear and Direct
Provide explicit, unambiguous instructions. Claude 4.x excels with precise direction.

### 2. Use XML Tags for Structure
Organize prompts with semantic tags like `<instructions>`, `<example>`, `<conte