Subagents

subagents() exposes one tool per configured child Agent Definition. Use it when the main Agent should delegate bounded work without giving the model a generic agent runner.

Installation

Import the Capability factory from @vite-hub/agent/capabilities and add it to defineAgent({ capabilities }). Name each subagent with a lowercase stable identifier and give it a concrete description.

What it adds

The Capability adds model-facing tools such as run_researcher or an explicit toolName. Each tool starts the configured child Agent with a message, optional structured context, optional call options, and an inherited invoker.

Configuration

import { defineAgent } from '@vite-hub/agent'
import { subagents } from '@vite-hub/agent/capabilities'
import researcher from './researcher'

export default defineAgent({
  driver: { model },
  capabilities: [
    subagents({
      agents: {
        researcher: {
          agent: researcher,
          description: 'Research one narrow question and return sourced notes.',
        },
      },
    }),
  ],
})

Runtime behavior

subagents() validates the configured names and tool names before the Agent runs. At invocation time, each subagent tool calls runAgent() with the parent runtime context and a child run id derived from the parent run id when available.

Requirements

Each subagent entry requires an agent and a non-empty description. Subagent keys must be lowercase stable identifiers, and generated tool names use underscores instead of dashes.

Driver support

Options

Inspect and verify

Open Agent DevTools and confirm each subagent appears as a tool with the expected name and description. Run one delegated task and verify the child invocation id and inherited invoker metadata.

Reference

• Agent definitions
• Agent invocations
• Source: packages/agent/src/capabilities/subagents.ts