System Prompt Writer

概要

This skill provides expert guidance for developing high-performance system prompts for AI agents, focusing on the core philosophy of context engineering. It helps developers find the 'Goldilocks zone' of instruction depth—avoiding both over-rigidity and vagueness—while ensuring the minimum effective dose of information is delivered through high-signal tokens. The skill provides a standardized hybrid structure using Markdown and XML, and it includes critical, mandatory instructions for handling Python string formatting constraints, such as escaping curly braces to prevent KeyError exceptions during prompt rendering. By following these guidelines, developers can create clearer tool guidance, more effective few-shot examples, and robust prompts that maximize agent performance and reliability.

主な機能

• Minimum effective information principles for token efficiency
• Hybrid Markdown-XML structural patterns for clarity
• 304 GitHub stars
• Context engineering and 'Right Altitude' writing guidance
• Mandatory Python template variable escaping (double braces) rules
• Strategic tool selection logic and decision tree frameworks

ユースケース

• Designing new system prompts for complex autonomous agent workflows
• Debugging and preventing template rendering errors in Python-based prompt systems
• Refactoring existing prompts to reduce token usage while maintaining performance