How do I trigger the agent creator skill?

You can trigger it using the /agent-creator command or by asking Claude to 'create a sub-agent' or 'make an agent for X'.

Do I need to restart Claude Code after creating a new agent?

No, new agents are recognized and available for use immediately after the file is created.

Where are the newly created sub-agents stored?

Agents can be stored at the project level in the .claude/agents/ directory or globally in the ~/.claude/agents/ directory.

What models can I assign to a sub-agent?

You can assign Claude 3.5 Sonnet, Opus, Haiku, or set it to 'inherit' to use the same model as the primary session.

Can I define specific tools for different agents?

Yes, you can configure sub-agents with specific toolsets such as Read, Edit, Bash, Grep, and WebSearch based on their purpose.

Sub-Agent Creator

Name: Sub-Agent Creator
Author: jo-minjun

byjo-minjun

•

Productivity & Workflow

Automates the generation and configuration of specialized sub-agents to extend Claude Code's capabilities for specific developer workflows.

About

The Sub-Agent Creator skill provides a structured framework for building custom sub-agents within the Claude Code environment. It streamlines the entire setup process—from gathering requirements like model selection and tool permissions to generating the necessary YAML frontmatter and system prompts. By allowing users to define specialized roles such as code reviewers, testers, or researchers, this skill enables a more modular and efficient AI-assisted development workflow, ensuring that Claude always has the right context and tools for the task at hand.

Key Features

Interactive multi-step requirement gathering for new agent definitions
Template-based system prompt construction for common development roles
Standardized YAML frontmatter generation including model and permission settings
2 GitHub stars
Real-time validation against official Claude Code tool specifications
Automated directory and file creation for project-level or global agents

Use Cases

Setting up a 'Security Auditor' agent focused on scanning codebases for vulnerabilities
Generating a dedicated 'Code Reviewer' agent with specific quality check guidelines
Creating a 'Documentation Specialist' to maintain project wikis using Read and Write tools

About

Key Features

Interactive multi-step requirement gathering for new agent definitions
Template-based system prompt construction for common development roles
Standardized YAML frontmatter generation including model and permission settings
2 GitHub stars
Real-time validation against official Claude Code tool specifications
Automated directory and file creation for project-level or global agents

Use Cases

Setting up a 'Security Auditor' agent focused on scanning codebases for vulnerabilities
Generating a dedicated 'Code Reviewer' agent with specific quality check guidelines
Creating a 'Documentation Specialist' to maintain project wikis using Read and Write tools