What is a Claude Code subagent?

A subagent is a specialized configuration of Claude Code that runs in an isolated context with its own specific tools and instructions to perform dedicated tasks efficiently.

How do subagents trigger automatically?

By adding 'Use proactively when...' or similar trigger phrases to the subagent's description, Claude Code can automatically invoke the subagent when those conditions are met.

Where are these subagent configurations saved?

Project-specific configurations are saved in the .claude/agents/ directory of your repository, while global agents are stored in your home directory's .claude/agents/ folder.

Why use isolated contexts for subagents?

Isolated contexts prevent context bloat by ensuring the model only processes information relevant to its specific role, which increases accuracy and reduces token consumption.

Can I restrict which tools a subagent can use?

Yes, you can define a minimal CSV list of tools in the configuration, ensuring the subagent only has access to the specific utilities required for its role.

Subagent Creator

Name: Subagent Creator
Author: wayne930242

bywayne930242

•

Productivity & Workflow

Generates specialized Claude Code subagent configurations with isolated contexts to streamline complex development workflows.

About

Subagent Creator is a powerful utility for implementing Agentic Context Engineering (ACE) within Claude Code. It allows developers to define and deploy specialized subagents—such as code reviewers, test runners, or domain experts—each operating within its own isolated context window. By automating the creation of these configurations in the .claude/agents/ directory, the skill ensures that tasks are delegated to focused agents with minimal, necessary toolsets, preventing context bloat and improving the overall precision and reliability of AI-driven development processes.

Key Features

Automated generation of .claude/agents/ configuration files
Isolated context management to improve model reasoning and focus
Support for model-specific selection including Sonnet, Haiku, and Opus
7 GitHub stars
Granular tool permissioning for enhanced security and task specificity
Configurable proactive trigger patterns for automatic subagent invocation

Use Cases

Setting up a dedicated code reviewer agent that automatically triggers during the development cycle
Creating an isolated test runner agent to debug and execute test suites without polluting the main chat context
Developing domain-specific expert agents for architectural planning or security auditing within a project

About

Key Features

Automated generation of .claude/agents/ configuration files
Isolated context management to improve model reasoning and focus
Support for model-specific selection including Sonnet, Haiku, and Opus
7 GitHub stars
Granular tool permissioning for enhanced security and task specificity
Configurable proactive trigger patterns for automatic subagent invocation

Use Cases

Setting up a dedicated code reviewer agent that automatically triggers during the development cycle
Creating an isolated test runner agent to debug and execute test suites without polluting the main chat context
Developing domain-specific expert agents for architectural planning or security auditing within a project