agent-builder

What problem does it solve?

Building effective Claude Code subagents requires careful design, prompt engineering, and tool configuration. This Skill provides expert guidance to create specialized AI assistants that excel at complex, multi-step workflows, ensuring they are efficient, reliable, and produce high-quality results.

Core Features & Use Cases

• Agent Design Principles: Learn to define clear agent scopes, single responsibilities, and measurable outcomes for optimal performance.
• System Prompt Engineering: Master the art of writing detailed system prompts that define agent roles, processes, output formats, and quality standards.
• Tool & Model Configuration: Understand how to grant minimal necessary tool access and select the appropriate Claude model (Opus, Sonnet, Haiku) for each agent's task.
• Use Case: A developer wants to create a specialized "code-reviewer" agent. This Skill guides them through defining its role, specifying review categories (quality, security, performance), configuring Read, Grep, and Bash(git *) tools, and structuring its output for actionable feedback.

Quick Start

Use the agent-builder skill to help me create a new Claude subagent. I want it to specialize in debugging TypeScript errors.