Can I restrict the tools available to a specific subagent?

Yes, you can specify exactly which tools an agent can access using the 'tools' field in the agent's YAML frontmatter, which is ideal for safety and focus.

What is the difference between a Claude Skill and a Claude Agent?

Skills are capability packages triggered automatically by context, while Agents are specialized subagents stored in an agents directory and invoked via the Task tool for focused expertise.

Where should I store my Claude subagents?

Agents can be stored in the project-specific 'agents/' directory, your personal '~/.claude/agents/' directory, or bundled within a plugin.

What model should I use for my subagent?

The 'inherit' setting is recommended to adapt to the parent context, but you can also specify 'haiku' for speed, 'sonnet' for balance, or 'opus' for complex reasoning.

How do I trigger a custom agent in Claude Code?

Agents are triggered by specific keywords or phrases defined in their frontmatter description, such as 'create agent' or specific domain terms like 'security review'.

Claude Agent Architect

Name: Claude Agent Architect
Author: outfitter-dev

byoutfitter-dev

•

Developer Tools

Standardizes the creation and management of specialized subagents within Claude Code to handle domain-specific tasks with focused expertise.

About

Claude Agent Architect provides a comprehensive framework for developing specialized subagents that extend Claude Code's capabilities with focused domain expertise. It offers standardized templates, frontmatter schemas, and validation workflows to ensure agents are correctly configured with optimal tools, models, and trigger conditions for specific tasks. By guiding developers through the discovery, design, and implementation stages, this skill enables the creation of predictable, efficient, and highly specialized AI-driven workflows tailored to unique project requirements.

Key Features

Specialized agent archetypes for analysis, implementation, and research
Standardized YAML frontmatter schema for precise agent configuration
Comprehensive validation checklists for quality assurance and naming conventions
Optimized tool and model selection guidance for subagent performance
Automated scaffolding scripts for rapid subagent creation
18 GitHub stars

Use Cases

Orchestrating multi-agent workflows by defining clear operational boundaries and responsibilities
Building a custom security auditor subagent with restricted tool access and specialized prompts
Creating a documentation-focused subagent that automatically triggers during project reviews

About

Key Features

Specialized agent archetypes for analysis, implementation, and research
Standardized YAML frontmatter schema for precise agent configuration
Comprehensive validation checklists for quality assurance and naming conventions
Optimized tool and model selection guidance for subagent performance
Automated scaffolding scripts for rapid subagent creation
18 GitHub stars

Use Cases

Orchestrating multi-agent workflows by defining clear operational boundaries and responsibilities
Building a custom security auditor subagent with restricted tool access and specialized prompts
Creating a documentation-focused subagent that automatically triggers during project reviews