sub-agents

To create specialized AI agents with isolated contexts, you define new agents with unique names, descriptions, system prompts, and toolsets. This Skill supports interactive, manual, and CLI-based configuration for modular agent development.

You assign specific tools and Claude models to custom AI sub-agents during the configuration phase. You can select from Haiku, Sonnet, or Opus models and assign tools like Read, Write, or Bash to tailor the agent's functionality for its designated task.

The best way to set up a dedicated code reviewer agent is to create a custom agent with Read, Grep, and Bash tools, assign it the Sonnet model, and provide a system prompt focused on code quality and security to ensure accurate reviews.

Yes, you can manage AI agents interactively by using the /agents command. This opens an interactive menu that allows you to easily create, configure, and manage your custom sub-agents without needing to manually edit configuration files.

You should use modular sub-agents instead of a single AI agent when you need to tackle specific tasks efficiently within a larger system. This approach isolates contexts and allows you to assign specialized prompts, tools, and models to distinct functions.

Configuring custom AI sub-agents does not require any external dependencies or libraries. The Skill operates independently to facilitate agent creation, tool configuration, and model selection through its built-in interactive and CLI interfaces.