Can I analyze a specific plugin instead of the whole repo?

Yes, you can use the --focus flag with a specific plugin name to perform a targeted deep-dive analysis.

Does this skill automatically implement the changes?

It generates a detailed recommendation report first. Once you review the priorities, you can instruct Claude to implement the specific agent files it suggested.

What is the primary benefit of sub-agent isolation?

It prevents 'context pollution' by moving verbose outputs like build logs or test results into a separate agent, keeping your main conversation focused on high-level logic.

Why does this skill recommend different models like Opus or Haiku?

Efficiency and cost: Opus is ideal for complex reasoning and architecture, while Haiku is faster and more cost-effective for structured, mechanical tasks like file formatting.

How does it help with security?

It identifies opportunities to create read-only agents or agents with restricted network/filesystem access, ensuring sensitive tasks are isolated.

Sub-Agent Analysis & Optimization

Name: Sub-Agent Analysis & Optimization
Author: laurigates

bylaurigates

•

开发者工具

Analyzes plugin collections to identify opportunities for sub-agent specialization and improved context management.

The Agents Analyze skill provides a systematic framework for auditing Claude Code plugin repositories to enhance performance and security through specialization. It scans existing skills to identify tasks that produce verbose output—such as build logs or test results—and recommends isolating them into dedicated sub-agents. By evaluating model selection (matching Opus to reasoning and Haiku to mechanical tasks) and enforcing the principle of least privilege for tools, it provides a comprehensive roadmap for scaling and optimizing your AI development environment.

主要功能

01Comprehensive inventory of plugins and skill coverage

02Model selection optimization between Opus and Haiku

0310 GitHub stars

04Security-focused tool constraint analysis

05Automated gap analysis for delegation workflows

06Identification of verbose tasks for context isolation

使用场景

01Refining security by restricting tool access to specialized agents

02Optimizing large development environments to prevent context window clutter

03Architecting modular AI workflows for complex enterprise repositories

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add laurigates/claude-plugins agents-analyze

For use in Claude.ai and ChatGPT

Download Skill