Where are custom subagent files stored?

Custom agents are stored as markdown files in .claude/agents/ for project-specific use or ~/.claude/agents/ for global user-level access.

Which tools should I give my subagent?

It is best practice to follow the principle of least privilege, granting only the specific tools (like Read, Grep, or Bash) required for the agent's specific role.

What is the 'Task' tool in subagent configuration?

The Task tool allows a subagent to spawn its own subagents, enabling complex, hierarchical task delegation and recursive problem-solving.

Can subagents be triggered automatically?

Yes, by including 'PROACTIVELY' or specific trigger conditions in the agent's description, Claude can automatically invoke the agent when those conditions are met.

What is a Claude Code subagent?

A subagent is a specialized instance of Claude configured with a specific system prompt, model, and toolset to handle distinct tasks within your workflow.

Subagent Creator

Name: Subagent Creator
Author: BrendanShields

byBrendanShields

•

生产力与工作流

Guides the creation and configuration of specialized Claude Code subagents for delegated task management.

关于

This skill provides a structured framework for building custom Claude Code subagents, allowing developers to extend Claude's capabilities with domain-specific AI experts. It streamlines the entire setup process—from defining triggers and selecting optimal tool permissions to choosing the right LLM model and drafting effective system prompts. By automating the creation of .md agent files, this skill helps users implement sophisticated multi-agent workflows, such as specialized security auditors, codebase researchers, or automated code reviewers, ensuring each subagent is purpose-built and correctly scoped for its intended environment.

主要功能

1 GitHub stars
Step-by-step workflow for configuring subagent markdown files
Support for proactive invocation and trigger-phrase configuration
Templates for common agent types including reviewers and researchers
Model optimization advice comparing Opus, Sonnet, and Haiku performance
Tailored tool selection guidance for Read, Write, Edit, and Bash access

使用场景

Building a read-only research assistant to explore and index large legacy codebases.
Developing a domain expert agent with specific access to database migration tools.
Creating a project-specific security subagent to automatically audit all new code changes.

关于

主要功能

1 GitHub stars
Step-by-step workflow for configuring subagent markdown files
Support for proactive invocation and trigger-phrase configuration
Templates for common agent types including reviewers and researchers
Model optimization advice comparing Opus, Sonnet, and Haiku performance
Tailored tool selection guidance for Read, Write, Edit, and Bash access

使用场景

Building a read-only research assistant to explore and index large legacy codebases.
Developing a domain expert agent with specific access to database migration tools.
Creating a project-specific security subagent to automatically audit all new code changes.