Delegate subtasks to specialized agents and collect results with wait_for_subagents.
npx skills add https://github.com/dtyq/magic --skill subagents-dtyq
Please help me install this Agent Skill. Skill: subagents Source: https://github.com/dtyq/magic/tree/main/backend/super-magic/agents/skills/subagents Command: npx skills add https://github.com/dtyq/magic --skill subagents-dtyq
This Skill enables orchestration by delegating sizable or multiple subtasks to specialized agents, keeping the main thread unburdened and avoiding cluttered conversations.
Dispatch a large task to a subagent and use wait_for_subagents to collect the results.
High-intent search queries and answers about installing and using this skill.
You can delegate AI agent tasks for parallel execution by dispatching independent subtasks to specialized agents via call_subagent, then aggregating the outcomes with wait_for_subagents to collect results efficiently.
Parallel task dispatch is applicable when your workload is decomposable into independent parts requiring different agent capabilities, keeping the main thread unburdened and avoiding cluttered conversations.
To prevent runaway processes during AI agent orchestration, enforce depth limits so sub-agents cannot recursively dispatch further subagents, maintaining strict flow control over the delegated tasks.
Yes, you can use dedicated agent types such as explore or shell for different subtasks during orchestration, which improves both execution efficiency and safety for the specific delegated operation.
The best way to orchestrate multiple independent subtasks is to dispatch them to specialized agents using call_subagent and aggregate the final outcomes with wait_for_subagents for efficient collection.
breverdbidderDispatch specialized agents to run independent debugging tasks concurrently.
empty-canCoordinate multiple specialist agents for parallel research and staged analysis workflows.
lucasfdigitalOrchestrate multiple specialized AI agents in the Orchard ecosystem.
skytiger6724Dispatch fresh subagents per task with spec and code-quality reviews.
