new-agent

Create Claude Code agent scaffolds by delegating design to the agent-author subagent.

Updated Apr 1, 2026
One-click install
npx skills add https://github.com/julianjab/ia-tools --skill new-agent-julianjab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: new-agent
Source: https://github.com/julianjab/ia-tools/tree/main/plugins/scaffold/skills/new-agent
Command: npx skills add https://github.com/julianjab/ia-tools --skill new-agent-julianjab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates the initial creation of Claude Code agent definitions by gathering a brief, delegating design to an agent-author subagent, generating the agents/<name>.md artifact, and validating it with an audit.

Core Features & Use Cases

  • Delegates agent design to the agent-author subagent to ensure consistent, high-quality definitions.
  • Writes the target agents/<name>.md file and performs an integrity audit with /audit-agent.
  • Interactive flow that asks clarifying questions before writing to ensure complete, usable artifacts.

Quick Start

Run /new-agent <name> to scaffold an agent, then answer prompts for missing fields.

Frequently Asked Questions about new-agent

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scaffold a Claude Code agent definition file?

To scaffold a Claude Code agent, run /new-agent <name> to trigger an interactive flow that delegates design to an agent-author subagent, writes the agents/<name>.md file, and audits the artifact for quality.

What is the process for creating an AI agent in Claude Code?

Creating an AI agent involves an interactive flow that asks clarifying questions about the agent's purpose, mode, and destination, then delegates the design to an agent-author subagent to generate a complete definition file.

Can I bypass the interactive prompts when generating a Claude Code agent?

Yes, you can bypass the interactive prompts during agent creation by supplying optional flags, allowing the scaffolding process to proceed automatically without asking for missing fields like purpose or mode.

Does the agent scaffolding process validate the generated markdown file?

Yes, the scaffolding process validates the generated markdown file by running an integrity audit with /audit-agent and validating the output path to ensure the artifact passes quality checks before use.

What are the limitations of automating Claude Code agent creation?

A limitation of automating agent creation is that the interactive flow depends on gathering a brief and clarifying fields; if essential inputs are missing and prompts are bypassed, the resulting agents/<name>.md artifact may lack complete design specifications.