Agent Development Pro FAQs

Question 1

How do I validate my agent configuration?

Accepted Answer

Agents must follow specific validation rules: identifiers must be 3-50 lowercase characters with hyphens, and system prompts should typically be between 500 and 3,000 characters for optimal performance.

Question 2

Why is the description field critical for agents?

Accepted Answer

The description field defines the triggering conditions and includes example scenarios that Claude uses to determine exactly when to activate the agent during a user session.

Question 3

What is the recommended model for a subagent?

Accepted Answer

Using the 'inherit' setting is generally recommended so the agent uses the same model as the parent session, though you can specify 'sonnet', 'opus', or 'haiku' for specific performance needs.

Question 4

What is a Claude Code agent?

Accepted Answer

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.

Question 5

Can I limit which tools an agent can access?

Accepted Answer

Yes, the 'tools' field in the YAML frontmatter allows you to specify exactly which capabilities (like Read, Write, or Bash) the agent is permitted to use, following the principle of least privilege.

Agent Development Pro

概要

主な機能

ユースケース

Agent Development Pro

概要

主な機能

ユースケース