Agentic Principles

These are the foundational patterns for working with AI coding agents, distilled into five actionable principles. The standout stat is real: breaking tasks into focused steps yields 39% better results than vague prompts. You'll use this when onboarding teammates or resetting your own approach after wrestling with a bloated context window. The 90% planning, 10% execution ratio might feel counterintuitive, but it's the difference between shipping validated code and debugging production fires. Think of it as the readme you send someone before they waste a week fighting the agent instead of collaborating with it. The bilingual presentation (Korean and English) suggests these patterns emerged from actual team experience rather than theory.

Install to Claude Code

npx -y skills add supercent-io/skills-template --skill agentic-principles --agent claude-code