What is a Claude Code subagent?

A subagent is a specialized, isolated Claude instance with a focused role and specific tool access designed to perform autonomous tasks without requiring user interaction during its execution.

How do I invoke a specific subagent manually?

You can explicitly invoke an agent by typing a command like 'Use the code-reviewer subagent to check my recent changes' in the Claude Code terminal.

Can subagents ask the user questions?

No, subagents operate as black boxes and cannot use the AskUserQuestion tool. They must receive all necessary requirements from the main chat before execution.

Why use XML tags in subagent system prompts?

Semantic XML tags are more token-efficient and provide clearer instruction boundaries for Claude than standard markdown, leading to better task compliance and reasoning.

Claude Code Subagent Creator

Name: Claude Code Subagent Creator
Author: Eric-Ness

byEric-Ness

생산성 및 워크플로

Empowers developers to build, configure, and orchestrate specialized Claude Code subagents for autonomous multi-agent workflows.

소개

This skill provides comprehensive guidance on creating specialized Claude subagents that operate in isolated contexts to perform focused tasks without user interaction. It teaches developers how to write effective system prompts using XML structure, configure tool access with least-privilege principles, and design sophisticated multi-agent workflows using the Task tool. By delegating complex operations like research, security reviews, or code generation to autonomous subagents, users can significantly enhance their productivity and handle complex software engineering challenges with high precision within the Claude Code environment.

주요 기능

Interactive /agents command interface for rapid subagent creation and management
Granular tool configuration including Read, Write, Edit, Bash, and Grep
Scope management for project-level or global user-level agent availability
Optimized XML-based system prompt engineering for peak agent performance
Multi-model orchestration supporting Sonnet, Haiku, and Opus models
0 GitHub stars

사용 사례

Automating security and quality audits through a dedicated code-reviewer subagent
Delegating API research and documentation gathering to autonomous background agents
Orchestrating multi-step generation tasks where agents write, test, and refine code

소개

주요 기능

Interactive /agents command interface for rapid subagent creation and management
Granular tool configuration including Read, Write, Edit, Bash, and Grep
Scope management for project-level or global user-level agent availability
Optimized XML-based system prompt engineering for peak agent performance
Multi-model orchestration supporting Sonnet, Haiku, and Opus models
0 GitHub stars

사용 사례

Automating security and quality audits through a dedicated code-reviewer subagent
Delegating API research and documentation gathering to autonomous background agents
Orchestrating multi-step generation tasks where agents write, test, and refine code