System Prompt Writer

About

This skill provides comprehensive guidance for engineering effective AI agent system prompts by focusing on the 'minimum effective dose' of information. It helps developers find the 'Goldilocks zone' of prompt altitude—ensuring instructions are neither too rigid nor too vague—while providing a structured hybrid framework of Markdown and XML for maximum clarity. Crucially, the skill includes mandatory instructions for Python template escaping to prevent initialization errors and offers advanced strategies for defining tool selection logic and diverse few-shot examples to maximize agent performance.

Key Features

• 16 GitHub stars
• Context engineering principles for high-signal token density
• Unambiguous tool selection guidance and decision trees
• Mandatory Python template variable escaping rules for code samples
• Best practices for diverse and canonical few-shot prompting
• Structured hybrid Markdown and XML formatting patterns

Use Cases

• Refining existing prompts to improve response reliability and tool usage
• Architecting complex AI agent behaviors and system instructions
• Preventing template substitution errors in Python-based agent frameworks