Skip to content

JonEliRey/pam-framework

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Professional Agent Model (PAM)

A framework for building AI agent systems as professionals, not tools.

By Jonathan Reyes / Ethion Consulting


What Is PAM?

PAM is a blueprint for building AI agent systems. Not a ready-made system. A design philosophy and set of tools for creating your own, purpose-built for you.

PAM assumes you already have a brain (Claude, ChatGPT, Gemini, Mistral, or any LLM). What it provides is everything else: the harness you build around that brain to make it useful, reliable, and yours.

Read the article: The Professional Agent Model: A Blueprint for Building Your Own AI System

Read the full specification: PAM v0.5.2 (25,900 words)


Quick Start

Install as a Claude Code Plugin

Clone this repo into your project or a shared location:

git clone https://github.com/JonEliRey/pam-framework.git

Claude Code will automatically discover the plugin and its skills. You'll have access to:

  • /pam-skill-creator: create PAM-compliant skills (wraps /skill-creator with PAM standards)
  • /pam-hook-creator: create PAM-compliant hooks (wraps /hookify with criticality tiers)
  • /pam-agent-creator: create PAM-compliant agents with memory, cluster namespace, and skill preloading
  • /pam-cluster-creator: create complete PAM clusters with root context, domain.yaml, and registries

Build Your First Cluster

/pam-cluster-creator

The skill walks you through every decision: domain, shared context, agents, boundaries. When done, you have a working PAM cluster with all the pieces in place.


What's in This Repo

pam-framework/
  README.md                         ← you are here
  PAM-framework.md                  ← full specification (25,900 words)
  PAM-article.md                    ← companion article (3,500 words)
  .claude-plugin/
    plugin.json                     ← makes this a Claude Code plugin
  skills/
    pam-skill-creator/              ← skill creation with PAM standards
    pam-hook-creator/               ← hook creation with criticality tiers
    pam-agent-creator/              ← agent creation with memory + namespace
    pam-cluster-creator/            ← cluster creation with full structure

Core Ideas

  1. Agents are professionals, not tools. Think about potential, not fixed capabilities.
  2. The harness is everything around the brain. Skills, hooks, memory, context, guard rails.
  3. Clusters organize teams. Shared context, bounded domains, no information bloat.
  4. Memory makes it real. Without memory, every session starts from zero.
  5. Four executive seats govern the system. Chief of Staff (user advocacy), COO (operations), CTO (innovation), CISO (security).
  6. Start simple, grow on evidence. Don't build the full stack on day one.

License

MIT


Author

Jonathan Reyes · Ethion Consulting

PAM is a living framework. This document is the starting point, not the destination.

About

A framework for building AI agent systems as professionals, not tools

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors