subagent

Spawn parallel background agents for decomposed tasks with status and result retrieval.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/CerisonAutomation/clud --skill subagent-cerisonautomation
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagent
Source: https://github.com/CerisonAutomation/clud/tree/main/assistant/src/config/bundled-skills/subagent
Command: npx skills add https://github.com/CerisonAutomation/clud --skill subagent-cerisonautomation

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Subagent helps you complete multi-part requests faster by delegating independent tasks to autonomous workers running in parallel, while keeping coordination with the main conversation.

Core Features & Use Cases

  • Parallel task delegation: Spawn background agents that work independently on subtasks like research, planning, or implementation.
  • Role-based tool access: Use specialized roles (researcher, coder, planner) to scope permissions and reduce risk.
  • Parent coordination & notifications: Send notify_parent updates while the subagent runs, and retrieve results with subagent_read only after terminal status.

Quick Start

Ask an AI to spawn a researcher subagent labeled "research-task" with the objective "Gather sources and summarize key findings on X", then read its output after it completes.

Frequently Asked Questions about subagent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run parallel background agents for complex research and coding tasks?

Parallel background agents execute multiple independent workstreams concurrently by spawning subagents that handle tasks like web research or codebase exploration while the main conversation continues.

What is the best way to decompose complex tasks into independently progressing workstreams?

Decomposing complex tasks uses subagent lifecycle management via spawn, status, and read primitives to delegate subtasks like planning and implementation to autonomous workers running in parallel.

Can I scope tool access permissions for background agents based on their role?

Role-based tool access constraints allow you to assign specialized roles like researcher, coder, or planner to scope permissions and reduce risk for each delegated background agent.

How do I retrieve results from a background subagent after it finishes running?

Retrieve results with the subagent_read primitive only after the subagent reaches terminal status, ensuring the background work is complete before collecting the final output.

Do background agents notify the parent conversation while they are still running?

Send notify_parent updates while the subagent runs to maintain coordination with the main conversation, providing intermediate status updates before the task reaches terminal status.

When should I avoid using parallel subagent orchestration for task delegation?

Avoid parallel subagent orchestration for simple, sequential tasks that lack independent components, as the lifecycle management overhead outweighs the benefits of concurrent execution.