delegation

Distributes tasks to subagents via synchronous forks or asynchronous background processes.

Updated Jun 22, 2026
One-click install
npx skills add https://github.com/alangeb/tau --skill delegation-alangeb
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: delegation
Source: https://github.com/alangeb/tau/tree/main/src/skills/delegation
Command: npx skills add https://github.com/alangeb/tau --skill delegation-alangeb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill addresses context bloat and performance degradation by providing a structured framework for offloading tasks to subagents, preventing the primary agent from exceeding token limits or becoming overwhelmed.

Core Features & Use Cases

  • Strategic Delegation: Choose between subagent (isolated), fork (context-aware), or background (asynchronous) execution based on task requirements.
  • Context Management: Automatically offload work when context usage exceeds 30% to maintain optimal performance.
  • Use Case: When tasked with a complex project, use this skill to spawn a subagent for isolated code review while you maintain the high-level orchestrator role, ensuring efficient memory usage.

Quick Start

Use the delegation skill to fork the current session and assign a subagent to perform a code review on the latest changes.

Frequently Asked Questions about delegation

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

FAQPage Schema
How do I prevent context bloat when running a complex multi-agent project?

To prevent context bloat in a multi-agent project, you can delegate tasks to subagents. This offloads work automatically when context usage exceeds 30%, maintaining optimal agent performance and responsiveness.

When should I use asynchronous background execution instead of a synchronous fork?

Use asynchronous background execution for independent tasks to keep the primary agent responsive, and use synchronous forks when the delegated subagent requires shared, context-aware execution within the current session.

What is the best way to run parallel tasks without exceeding token limits?

The best way to run parallel tasks without exceeding token limits is to distribute the workload to isolated subagents. This strategic delegation prevents the primary agent from becoming overwhelmed and avoids token overflow.

How do I delegate a code review while keeping my main orchestrator agent active?

You can delegate a code review by spawning an isolated subagent for the review process. This allows the primary agent to maintain its high-level orchestrator role and manage memory usage efficiently.

Does task delegation automatically trigger when context usage gets high?

Yes, task delegation can automatically trigger when context usage exceeds 30%. This automated context management offloads work to subagents to maintain optimal system performance and prevent token overflow.