How does this help with token optimization?

The guidelines recommend target sizes of 300-400 lines per agent and provide 'DRY' strategies to externalize shared knowledge into skills.

What architecture patterns does this skill support?

It provides detailed documentation for Single Agent, Orchestrator-Worker, Pipeline, and Skill-Augmented patterns based on your plugin's complexity.

When should I use the Orchestrator-Worker pattern?

This pattern is ideal when you need parallel execution of independent subtasks or when specialized workers require different tool sets.

Does this skill provide naming conventions?

Yes, it enforces strict kebab-case naming for plugins, agents, and skills to ensure consistency and prevent routing errors.

Plugin Creation Guidelines

Name: Plugin Creation Guidelines
Author: L-Sypniewski

byL-Sypniewski

Learning & Documentation

Architects optimized Claude Code plugins using standardized patterns, token-efficient structures, and multi-agent orchestration best practices.

About

This skill provides a comprehensive framework for building high-performance Claude Code plugins by distilling official best practices into actionable architectural patterns. It guides developers through selecting the right structure—whether it's a simple Single Agent, a parallel Orchestrator-Worker model, or a multi-stage Pipeline—to ensure plugins remain token-efficient and maintainable. By following these guidelines, developers can implement standardized naming conventions and integration protocols that prevent routing errors and maximize the reliability of custom AI workflows.

Key Features

0 GitHub stars
Token optimization strategies with specific line-count targets for agents
Description engineering patterns to improve AI routing accuracy
Standardized naming and directory structure conventions for consistency
Multi-pattern architecture decision trees (Orchestrator, Pipeline, Single Agent)
Agent-to-Skill delegation protocols for shared domain knowledge

Use Cases

Optimizing existing plugins to reduce token consumption and improve speed
Standardizing custom plugin development for engineering teams
Designing complex, multi-agent plugins for specialized development workflows

About

Key Features

0 GitHub stars
Token optimization strategies with specific line-count targets for agents
Description engineering patterns to improve AI routing accuracy
Standardized naming and directory structure conventions for consistency
Multi-pattern architecture decision trees (Orchestrator, Pipeline, Single Agent)
Agent-to-Skill delegation protocols for shared domain knowledge

Use Cases

Optimizing existing plugins to reduce token consumption and improve speed
Standardizing custom plugin development for engineering teams
Designing complex, multi-agent plugins for specialized development workflows