How should I structure the system prompt for an agent?

A well-structured prompt should include the agent's role, core responsibilities, a step-by-step analysis process, quality standards, and a clear definition of the output format.

What are the rules for naming an agent identifier?

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

What is the difference between an agent and a command in Claude Code?

Agents are designed for autonomous, multi-step work that handles complex tasks independently, while commands are typically for direct, user-initiated actions.

How do I trigger a specific agent in my plugin?

Agents are triggered via the 'description' field in the markdown frontmatter. This field must include specific triggering conditions and multiple examples with context, user requests, and commentary.

Which model should I use for my subagent?

Using 'inherit' is the recommended practice as it allows the agent to use the same model as the parent. However, you can specify 'sonnet', 'opus', or 'haiku' for specific cost or capability needs.

Agent Development for Claude Code

Name: Agent Development for Claude Code
Author: Activer007

byActiver007

開発者ツール

Provides comprehensive guidance and templates for building, structuring, and optimizing autonomous agents within Claude Code plugins.

概要

This skill serves as a master guide for creating autonomous agents in the Claude Code ecosystem, focusing on the specialized markdown file structure and YAML frontmatter required for high-performance subagents. It equips developers with best practices for designing system prompts, defining precise triggering conditions using example-based descriptions, and selecting the appropriate tools and model inheritance patterns to handle complex, multi-step tasks independently. Whether building a code reviewer or a security analyzer, this skill ensures agents are reliable, well-structured, and correctly namespaced.

主な機能

Automated validation rules for identifiers and description formats
Example-driven triggering logic to optimize agent activation
Granular tool restriction guidelines for the principle of least privilege
Standardized YAML frontmatter templates for agent configuration
System prompt design patterns for expert persona development
0 GitHub stars

ユースケース

Building a specialized code reviewer agent that triggers on PR-style requests
Designing a documentation writer that generates API specs from source code
Creating an autonomous security analyzer to proactively scan for vulnerabilities

概要

主な機能

Automated validation rules for identifiers and description formats
Example-driven triggering logic to optimize agent activation
Granular tool restriction guidelines for the principle of least privilege
Standardized YAML frontmatter templates for agent configuration
System prompt design patterns for expert persona development
0 GitHub stars

ユースケース

Building a specialized code reviewer agent that triggers on PR-style requests
Designing a documentation writer that generates API specs from source code
Creating an autonomous security analyzer to proactively scan for vulnerabilities