subagents

Distribute complex computational tasks across parallel subagents with permission hooks.

2|1|Updated Mar 13, 2026
One-click install
npx skills add https://github.com/bongrealty/skillcraft --skill subagents-bongrealty
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: subagents
Source: https://github.com/bongrealty/skillcraft/tree/main/subagents
Command: npx skills add https://github.com/bongrealty/skillcraft --skill subagents-bongrealty

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill allows you to tackle complex, exploratory, or parallelizable tasks by distributing the workload across multiple subagents, keeping the main context clean and focused.

Core Features & Use Cases

  • Parallelization: Run independent tasks simultaneously across multiple agents.
  • Context Isolation: Offload heavy research or analysis to subagents to maintain focus in the main agent's context.
  • Permission Routing: Automatically approve safe permissions via hooks, flagging ambiguous ones for review.
  • Use Case: When refactoring a large system, use subagents to explore different components in parallel, then have the main agent synthesize the findings and implement the changes.

Quick Start

Use 5 subagents to explore the codebase for entry points, React component structure, tools implementation, state management, and testing infrastructure.

Frequently Asked Questions about subagents

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

FAQPage Schema
How do I scale compute for complex codebase exploration and large refactors?

To scale compute for large refactors, you can distribute independent tasks across parallel subagents to explore different codebase components simultaneously, then synthesize the findings in the main context.

What is the best way to keep the main agent context clean during heavy research?

The best way to keep the main agent context clean is offloading heavy research and analysis to subagents, isolating the exploratory workload while the main agent maintains focus and processes the summarized results.

Can I run parallel analysis and task distribution without external libraries?

Yes, you can run parallel analysis and task distribution without external libraries by relying on agentic capabilities and hooks to automatically route safe permissions, flagging only ambiguous ones for review.

How do I explore a codebase for entry points and state management in parallel?

You can explore a codebase for entry points, React component structure, state management, and testing infrastructure by assigning these independent tasks to multiple subagents running simultaneously.

Does context management with subagents work for writing tests and large system refactoring?

Yes, context management with subagents works for large system refactoring and test writing by isolating compute-heavy tasks and automatically approving safe permissions via hooks to streamline the process.