Does delegating tasks save on token usage?

Yes. Subagents run in separate contexts, meaning their specific token consumption does not impact your primary Claude Code context window.

How does the Maestro Delegation Advisor improve performance?

It analyzes the specific requirements of a task—such as the need for web access or high-speed coding—and routes it to the agent best suited for that specific workload, ensuring higher accuracy and faster completion.

When should I handle a task directly instead of delegating?

You should handle tasks directly when they involve complex system architecture, deep codebase analysis, or high-stakes security audits, as these are primary strengths of the core Claude model.

Is the delegation process manual?

No. When running in Claude Code, delegation happens automatically. The system recognizes when a task matches a subagent's profile and invokes it without requiring special syntax.

Can this skill handle web research?

Yes, it automatically delegates research and automation tasks to the Gemini subagent, which is specialized in web-based data gathering and UI tasks.

Maestro Delegation Advisor

Name: Maestro Delegation Advisor
Author: NehoraiHadad

byNehoraiHadad

0•

Productividad y Flujo de Trabajo

Optimizes multi-agent workflows by intelligently delegating specialized tasks to the most effective AI models within the AgentMaestro ecosystem.

The Maestro Delegation Advisor is a specialized expert system designed for the AgentMaestro orchestration suite. It provides a structured decision framework that allows Claude to analyze task requirements—such as security sensitivity, execution speed, or research needs—and automatically route them to the optimal subagent. By leveraging the specific strengths of Claude for architecture, Codex for rapid code generation, and Gemini for web-based research, this skill ensures maximum performance, cost-efficiency, and token conservation in complex development environments.

Características Principales

01Support for sequential multi-agent task execution and parallelization

02Pre-defined decision rules for security audits and system architecture

03Context-aware delegation that preserves main-thread token usage

04Automated task analysis and agent matching based on performance metrics

050 GitHub stars

06Seamless subagent invocation for Codex and Gemini within Claude Code

Casos de Uso

01Offloading high-volume code generation to Codex to maintain speed and reduce costs

02Delegating web-intensive data gathering and browser automation to Gemini

03Orchestrating complex software projects that require simultaneous research, design, and implementation phases

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add nehoraihadad/agent-maestro-cli maestro-delegation-advisor

For use in Claude.ai and ChatGPT

Download Skill