sub-agents

Configure and deploy specialized subagents with isolated contexts and tool permissions.

5|1|Updated Oct 30, 2025
One-click install
npx skills add https://github.com/reggiechan74/claude-plugins --skill sub-agents
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: sub-agents
Source: https://github.com/reggiechan74/claude-plugins/tree/main/plugins/claude-code-metaskill/skills/sub-agents
Command: npx skills add https://github.com/reggiechan74/claude-plugins --skill sub-agents

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Complex tasks often require different expertise or isolated contexts to prevent the main conversation from becoming polluted. Subagents allow Claude to delegate work to specialized AI assistants, improving focus and success rates.

Core Features & Use Cases

  • Task Delegation: Create specialized agents (e.g., code reviewer, debugger, data scientist) for focused, single-responsibility tasks.
  • Isolated Context: Each subagent operates in its own context window, preventing main conversation pollution and improving clarity.
  • Configurable Tools & Prompts: Define specific tool access and system prompts for each subagent, tailoring its behavior precisely.
  • Use Case: When a bug is reported, delegate the task to a "debugger" subagent, which will focus solely on root cause analysis and minimal fixes without affecting your main development context.

Quick Start

Show me how to create a new subagent for code review.

Frequently Asked Questions about sub-agents

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

FAQPage Schema
How do I delegate specialized tasks to separate AI assistants?

Task delegation through subagents lets you create focused AI assistants for specific roles—like code review or debugging—each with isolated context and tailored tool access, keeping your main conversation clean and improving task success rates.

What's the difference between using one AI assistant versus multiple subagents?

Multiple subagents operate in isolated context windows with customizable system prompts and granular tool permissions, preventing context pollution and enabling single-responsibility focus, whereas one assistant handles everything in a shared context.

How do I configure tool access and system prompts for individual subagents?

Define per-subagent configuration using YAML frontmatter in designated directories, specifying custom system prompts and tool permissions for each subagent to tailor behavior precisely to its specialized role.

Can subagents work in coding projects and data workflows?

Yes, subagents support both coding projects and data workflows with project-level or user-level scope, enabling task delegation across diverse development and analytical contexts with context isolation.

When should I use automatic delegation versus explicit subagent invocation?

Automatic delegation routes tasks to specialized subagents transparently, while explicit invocation gives you control over when to engage specific assistants—choose based on whether you want orchestration to happen implicitly or on demand.