Subagent

A second AI agent spun up to handle one focused task in its own context, then report back. Used to search, review, or work in parallel without cluttering the main session.

What it is

A subagent is a separate agent that the main agent launches to handle one focused job. It works in its own context window, does the task (search across files, review a diff, draft one section), and returns just the result. The main session stays clean because the heavy lifting happened somewhere else.

Simple analogy

It is delegating to an assistant. You do not need to watch them read every file. You need the answer they come back with. The pile of paper stays on their desk, not yours.

Why this matters for designers

Subagents are how you keep a session fast and focused, and how you run work in parallel. One subagent can review a layout for accessibility while another checks the copy, each reporting back without flooding your main conversation. It also saves usage: a big file search happens in the subagent’s context, not yours.

How it works in practice

  1. The main agent identifies a focused task worth offloading.
  2. It launches a subagent with just that task and the context it needs.
  3. The subagent returns a result; the main session continues lean.