What is a Claude Code Agent?

An agent is an autonomous subprocess within a Claude Code plugin designed to handle complex, multi-step tasks independently using specific tools and specialized system prompts.

Which model should my Claude Code agent use?

Using 'inherit' is the recommended default as it matches the parent model. However, you can specify 'sonnet' for balance, 'opus' for high complexity, or 'haiku' for speed and cost efficiency.

What are the rules for agent identifiers?

Agent names must be 3-50 characters long, use only lowercase letters, numbers, and hyphens, and must start and end with an alphanumeric character.

How can I restrict an agent's tool access?

You can use the 'tools' field in the agent's frontmatter to provide an array of specific allowed tool names, ensuring the agent only has access to the utilities necessary for its task.

How do I trigger an agent in Claude Code?

Agents are triggered based on the 'description' field in their YAML frontmatter. This field must include triggering conditions and multiple blocks showing the context and expected assistant response.

Agent Development for Claude Code

Name: Agent Development for Claude Code
Author: himmelreich-it

byhimmelreich-it

0•

개발자 도구

Streamlines the creation of autonomous agents for Claude Code plugins through standardized structures and best practices.

This skill provides a comprehensive framework for building, configuring, and optimizing autonomous agents within the Claude Code ecosystem. It offers detailed guidance on defining agent frontmatter, crafting precise triggering conditions using example-based commentary, and designing robust system prompts that guide complex, multi-step tasks. By implementing these standardized patterns, developers can create specialized subagents—such as code reviewers or security analyzers—that activate reliably and execute with high precision while following the principle of least privilege for tool access.

주요 기능

01Validation rules for agent identifiers and documentation

02Tool restriction configurations for secure agent execution

03Example-driven triggering logic with specialized commentary tags

04Standardized YAML frontmatter templates for agent metadata

050 GitHub stars

06Structured system prompt patterns for autonomous personas

사용 사례

01Developing specialized subagents for automated code reviews and testing

02Configuring precise triggering conditions to prevent agent overlap

03Optimizing agent personas for specific domain expertise like security or API design

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add himmelreich-it/agent-skill-converter agent-development

For use in Claude.ai and ChatGPT

Download Skill