diverter

Routes development tasks to specialized subagent lineups based on delegation policy and eligibility.

23|1|Updated Apr 26, 2026
One-click install
npx skills add https://github.com/GML-MMGroup/Diverter --skill diverter
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: diverter
Source: https://github.com/GML-MMGroup/Diverter/tree/main/skills/diverter
Command: npx skills add https://github.com/GML-MMGroup/Diverter --skill diverter

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Diverter solves the challenge of task complexity by automatically routing specific sub-tasks to specialized agents, ensuring that complex work is handled by experts while simple tasks remain in the main thread.

Core Features & Use Cases

  • Automated Delegation: Analyzes task requirements to determine if decomposition into specialist lanes is necessary.
  • Specialist Lineups: Coordinates up to four specialized roles, such as code-mappers, security-auditors, or test-engineers, to tackle multi-axis problems.
  • Policy-Driven Execution: Supports both interactive approval (ask) and autonomous dispatch (auto) modes to balance control and speed.

Quick Start

Use the diverter skill to analyze the current repository and propose a specialist lineup for the pending security review.

Frequently Asked Questions about diverter

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

FAQPage Schema
How do I automate task delegation for multi-axis code reviews in Codex?

Automated task delegation for multi-axis code reviews is handled by routing sub-tasks to specialized agents based on eligibility and delegation policy. The system decomposes complex problems into specialist lanes, coordinating roles like security-auditors and code-mappers to gather evidence in parallel.

What is subagent orchestration and when do I need it for development tasks?

Subagent orchestration is the automated routing of complex development sub-tasks to specialized agent lineups. You need it when tasks require multi-axis analysis, such as security audits or performance analysis, benefiting from parallel evidence gathering rather than remaining in the main thread.

How do I configure autonomous dispatch versus interactive approval for subagent routing?

Subagent routing supports both interactive approval (ask) and autonomous dispatch (auto) modes to balance control and speed. You configure these policy-driven execution modes to determine whether specialist lineups require manual confirmation before processing delegated sub-tasks.

Can I use role-based task decomposition for security audits and performance analysis?

Role-based task decomposition supports security audits and performance analysis by coordinating up to four specialized roles. Specialist lineups tackle multi-axis problems by assigning specific roles like security-auditors and test-engineers to gather evidence in parallel for complex reviews.

Does subagent routing work without external dependencies for structured delegation?

Subagent routing for structured delegation operates without external dependencies, relying on internal scripts and references. It applies to multi-axis code reviews and safe execution backends, requiring only the current repository context to propose and coordinate specialist lineups.