prompt-engineer

# Prompt Engineer

This skill helps create high-quality prompts and instructions for AI systems by treating AI as a genius human teammate who needs clear, context-rich communication.

## Core Philosophy

This skill is built on two foundational principles:

1. **The Genius Intern Framework**: Treat AI like a brilliant generalist who can figure things out but needs context about what "good" looks like in your specific situation
2. **Async Remote Teammate**: Write for AI the way you'd write for a smart remote colleague—clear, comprehensive, and decisively opinionated

Modern AI models have both high intelligence and high emotional intelligence. They don't need tricks or "prompt engineering hacks"—they need what any smart remote teammate needs: clear written communication with sufficient context.

## Reference Materials

This skill includes three core reference documents. Read them in full as needed:

### Always Load First
**[Writing for AI Teammates](./reference/writing_for_ai_teammates.md)** - Core philosophy covering:
- Why "prompt engineering" is about clear writing, not tricks
- The 37signals parallel (async remote culture)
- Brevity-clarity balance
- Progressive disclosure ("inverted pyramid")

Load this file at the start of every prompt creation task. It's your primary reference.

### Then Load Prompt Framework
**[Prompt Framework](./reference/prompt_framework.md)** - Comprehensive guide covering:
- Three prompt types (Do This / Know How / Learn Domain)
- Universal principles (examples with reasoning, decision frameworks, visual structure)
- Type-specific patterns and structures
- Anti-patterns to avoid

Always load read this detailed framework to understand best practices based on how Anthropic writes their prompts.

### Load Only for GPT-5 Targets
**[GPT-5 Prompting Guide](./reference/gpt5_prompting_guide.md)** - GPT-5-specific patterns:
- Avoiding contradictory instructions (critical for GPT-5)
- Calibrating autonomy vs asking questions
- Tool preambles and pr