mp-agent-create

Create Claude Code custom agents with compliant frontmatter and workflow templates.

1|Updated Feb 7, 2026
One-click install
npx skills add https://github.com/MartinoPolo/mpx-claude-code --skill mp-agent-create
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mp-agent-create
Source: https://github.com/MartinoPolo/mpx-claude-code/tree/main/skills/mp-agent-create
Command: npx skills add https://github.com/MartinoPolo/mpx-claude-code --skill mp-agent-create

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It helps you create new Claude Code custom agents that follow your project’s conventions and Claude Code’s latest agent/sub-agent authoring guidelines, without manual trial-and-error.

Core Features & Use Cases

  • Guideline-first agent drafting: Fetches the latest authoring rules from a guide agent, then builds your agent to match required frontmatter fields, structure, and conventions.
  • Requirements gathering in one pass: Collects purpose, tool allowlist, model choice, read-only vs read-write, and status color in a single round-trip.
  • Validation and iteration with you: Compares the draft against fetched guidelines, fixes mismatches, and reviews the final agent for approval.

Quick Start

Create a new agent with the prompt: "Create an agent named 'mp-foo' that drafts PR descriptions, uses only Read and Write tools, and should run with sonnet."

Frequently Asked Questions about mp-agent-create

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

FAQPage Schema
How do I create a Claude Code custom agent with compliant frontmatter?

To create a Claude Code custom agent, you define its purpose, tool allowlist, and model choice. The agent is generated as an agents/<agent-name>.md file with validated frontmatter and a structured workflow template matching the latest authoring guidelines.

What is the best way to define a tool allowlist for a Claude Code sub-agent?

Defining a tool allowlist for a Claude Code sub-agent involves specifying permitted tools during the requirements gathering phase. This restricts the agent's capabilities to necessary actions, ensuring it operates within defined boundaries using approved tools only.

Can I validate a Claude Code agent draft against the latest authoring guidelines?

Yes, you can validate a Claude Code agent draft against the latest authoring guidelines. The validation process fetches current rules, compares the draft frontmatter and structure, fixes mismatches, and reviews the final agent for compliance before approval.

Do I need to manually format the workflow template when generating a Claude Code agent?

No, you do not need to manually format the workflow template when generating a Claude Code agent. The generation process automatically applies a structured workflow and output template while ensuring frontmatter validation against fetched guidelines.

What information is required to generate a Claude Code sub-agent in one pass?

Generating a Claude Code sub-agent requires the agent's purpose, tool allowlist, model selection, read-only or read-write permissions, and status color. These parameters are collected in a single round-trip to draft the agent.