When should I use the Skill tool instead of frontmatter loading?

Use the Skill tool for dynamic discovery at runtime when the specific skills needed depend on the user's request; use frontmatter for static, known skills required at start time.

What is context isolation in AI subagents?

Context isolation ensures that intermediate processing and tool outputs stay within a subagent's specific context, preventing the main conversation from becoming cluttered and exceeding token limits.

What defines an Enhanced Agent pattern?

An Enhanced Agent utilizes codebase exploration tools like Serena Gateway and context persistence tools like claude-mem to automatically discover skills and maintain state across different sessions.

How do I choose between a Single-Skill and Multi-Skill orchestrator?

Choose Single-Skill for narrow, focused tasks to maximize speed and clarity. Choose Multi-Skill for complex workflows involving multiple domains where coordination is required.

Orchestration Patterns

Name: Orchestration Patterns
Author: chkim-su

bychkim-su

Productivity & Workflow

Defines and implements standardized subagent architectures for managing complex AI workflows and skill coordination.

About

Orchestration Patterns provides a framework for designing specialized Claude Code subagents, focusing on context isolation and efficient task delegation. It guides developers through three primary architectural patterns—Single-Skill, Multi-Skill Orchestrator, and Enhanced Agent—to ensure AI agents remain focused, maintain context windows, and leverage dynamic skill loading. This skill is essential for building scalable agentic systems that require codebase awareness, cross-session memory, and intelligent tool selection.

Key Features

Best practices for static vs. dynamic skill activation
Standardized subagent architectures for context isolation
Dynamic skill loading patterns via the Skill tool
0 GitHub stars
Decision matrices for choosing between single and multi-domain agents
Integration patterns for codebase exploration and memory persistence

Use Cases

Building a specialized SQL agent with a focused, single-domain context
Creating a full-stack orchestrator to coordinate UI, API, and database tasks
Developing enhanced agents that use automated discovery for large codebases

About

Key Features

Best practices for static vs. dynamic skill activation
Standardized subagent architectures for context isolation
Dynamic skill loading patterns via the Skill tool
0 GitHub stars
Decision matrices for choosing between single and multi-domain agents
Integration patterns for codebase exploration and memory persistence

Use Cases

Building a specialized SQL agent with a focused, single-domain context
Creating a full-stack orchestrator to coordinate UI, API, and database tasks
Developing enhanced agents that use automated discovery for large codebases