CLAUDE.md

This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.

First Run: Installation

On first use, ask the user where to install the skills and agents:

"Where should I install hormozi-skills?

• Project (default) — .claude/ in the current project folder. Only available in this project.
• User — ~/.claude/ in your home directory. Available in all projects.

Press Enter to use the default (project)."

Default: project-level (.claude/ in the current working directory).

Installation copies skills/ and agents/ into the chosen .claude/ directory:

# Project (default)
cp -r skills/ agents/ .claude/

# User (global)
cp -r skills/ agents/ ~/.claude/

Important: After installation, Claude Code must be restarted for the new skills and agents to be available.

What This Is

A skill library for coding agents to build Hormozi-inspired offer systems. No build pipeline — content only. Each skills/ folder is a standalone agent skill; agents/ contains an orchestrator and five subagents.

Architecture

Two-Layer System

Skills (skills/<name>/SKILL.md) — standalone, user-invocable. Each has a frontmatter name and description. Some have a references/ subfolder with supporting material the skill reads at runtime.

Agents (agents/<name>.md) — orchestrator + subagents. Frontmatter includes name, description, tools, model, and optionally color.

• hormozi-orchestrator is the entry point. It interviews the user, detects funnel stage (A–E), builds a structured brief, and delegates to subagents in dependency order.
• The five sub-* agents are internal — they receive briefs from the orchestrator only and never interview the user directly.

Subagent Dependency Order

sub-market → sub-offer → sub-value → sub-pricing → sub-sales

sub-value and sub-pricing can run in parallel when an offer already exists.

Output

All generated documents write to output/. The orchestrator produces a final output/SUMMARY.md after all subagents complete. The output/ folder ships empty (.gitkeep).

Skill Structure Convention

Every skill lives at skills/<name>/SKILL.md. If a skill needs reference material, it goes in skills/<name>/references/. No other files belong inside a skill folder.

Every agent lives at agents/<name>.md — flat, no subfolders.

Adding a New Skill

1. Create skills/<skill-name>/SKILL.md with frontmatter name and description.
2. If it needs reference docs, add them to skills/<skill-name>/references/.
3. Add it to the skills table in README.md.

Adding a New Subagent

1. Create agents/sub-<name>.md with frontmatter name, description, tools, model.
2. Mark it internal in the description: "Internal subagent. Called by hormozi-orchestrator only."
3. Define its output file(s) in output/.
4. Wire it into the orchestrator's Phase 3 stage map and Phase 4 delegation logic.
5. Add it to the agents table in README.md.

Key Design Rules

• Subagents receive a fully structured brief — they have no memory of the conversation.
• All output lands in output/ relative to the repo root.
• Skills are user-facing; subagents are internal execution units. Do not mix the two roles.
• _archive/ holds deprecated versions — do not delete, do not edit.