---
title: "Subagent"
description: "Subagent skill for Vellum — spawns autonomous background agents that work independently on tasks."
canonical_url: "https://www.vellum.ai/docs/skills-reference/subagent"
md_url: "https://www.vellum.ai/md/docs/skills-reference/subagent"
related:
  - "/docs/skills-reference"
  - "/docs/skills-reference/acp"
  - "/docs/skills-reference/amazon"
  - "/docs/skills-reference/app-builder"
  - "/docs/skills-reference/browser"
  - "/docs/skills-reference/chatgpt-import"
  - "/docs/skills-reference/computer-use"
  - "/docs/skills-reference/contacts"
---

# Subagent

## What it does

Spawns autonomous background agents that work independently on tasks. Your assistant can delegate work to parallel workers that run in the background.

## Setup required

None. Works immediately.

## Permissions

- Subagents inherit the same permission rules as your main assistant

## Common prompts

| You say...                                                       | What happens                      |
| ---------------------------------------------------------------- | --------------------------------- |
| “Spawn an agent to research AI startups in healthcare”           | Launches background research      |
| “Start a background agent to monitor my inbox for the next hour” | Long-running background task      |
| “Check on my research agent”                                     | Gets status of a running subagent |
| “Cancel the background agent”                                    | Aborts a running subagent         |

## Configuration

- Subagents can run in silent mode (no intermediate output) or verbose mode
- Status tracking: pending, running, completed, failed, aborted
- Results are delivered via follow-up message

## Tips & gotchas

- **Genuinely independent.** Subagents run in the background while you continue chatting.
- **No shared context.** They can't access your current conversation's context unless you explicitly pass it.
- **Great for long tasks.** Deep research, monitoring, and batch processing are ideal use cases.
- **Multiple agents.** You can spawn multiple agents and check on them later.