FREE ENERGY — Phase0 Preview v0.1.2

FREE ENERGY — Phase0 Preview v0.1.2

This preview establishes FREE ENERGY as the canonical public identity of the project.

Federated Remastering Ecosystem for Everyone — Engine for Networked Entertainment, Remixing, Games & You

FREE ENERGY Remasters Everything.

Naming model

• FREE ENERGY is the public project, community, future website, and broader collaborative game-remastering/remixing ecosystem.
• Fleet-Control remains the name of the orchestration substrate and repository-local agent coordination model.
• Phase0 remains the current Markdown-first bootstrap implementation.

The GitHub repository may temporarily retain the legacy Fleet-Control-Public slug while repository and website identifiers are migrated. Historical tags, links, issues, branches, and protocol references are not rewritten merely for cosmetic consistency.

Packaging changes

The preview starter artifact is now named FREE-ENERGY-Phase0-Starter-v0.1.2-preview.zip and includes BRANDING.md.

Functional scope

This release is a branding/onboarding migration. It does not claim a new Fleet-Control protocol generation merely because the public project name changed.

Fleet-Control Public — Phase0 Preview v0.1.1

Fleet-Control Public — Phase0 Preview v0.1.1

This preview supersedes v0.1.0 for new installations.

What changed

The public Phase0 implementation is now explicitly licensed under the MIT License with:

Copyright (c) 2026 Distributed Minds

The human starter ZIP for this release includes the root LICENSE file, so people who download only the starter package receive the license terms together with the operating system and setup material.

No existing public tag was moved. The v0.1.0 preview remains an immutable historical snapshot of the earlier package.

New users

Download the asset named:

Fleet-Control-Public-Phase0-Starter-v0.1.1-preview.zip

Extract it and open README-FIRST.md, then GETTING-STARTED.md.

The starter package contains:

• the complete Phase0/ Markdown operating system;
• LICENSE;
• beginner setup documentation;
• a glossary and troubleshooting guide;
• ChatGPT Project instructions;
• separate A1 through A5 persistent automation prompts;
• installation checklist and templates.

You do not need prior Git or GitHub command-line experience to follow the beginner path.

Existing v0.1.0 users

If you already downloaded the v0.1.0 starter, use v0.1.1 instead so your local copy includes the explicit MIT terms.

Status

This remains a preview. The public state-machine and coordination model is intentionally being tested and generalized before a stable release. Public-safe improvements from the private Fleet-Control development repository may be carried forward into later public versions after portability and disclosure checks.

Fleet-Control Public — Phase0 Preview v0.1.0

Fleet-Control Public — Phase0 Preview v0.1.0

This is the first public preview of the Markdown-only Fleet-Control coordination/state-machine foundation.

If you have never used GitHub, Git, a ChatGPT Project, or an automation before, that is okay. Start with the ZIP attached to this release and open GETTING-STARTED.md. The instructions assume no previous knowledge.

This release is deliberately marked pre-release. The core operating model is usable for experimentation, but names, formats, installation details, and policy may still change as the public design is tested by people outside the original repository.

---

What is Fleet-Control Public?

Fleet-Control Public is not a new AI model and it is not a background server you install.

It is a set of Markdown (.md) files that act as a shared operating system for several persistent AI-agent runs working through the same GitHub repository.

The basic problem it tries to solve is simple:

If five AI agents wake up separately, how do they know what the human actually wants, what the other agents are doing, what has already been proved, whether they are allowed to change something, and what they should do next?

Phase0 answers that using repository-visible rules and append-only GitHub state rather than one giant hidden prompt.

The fleet can receive requests at very different levels, for example:

• “Think about this design.”
• “Research why this happens. Do not change code.”
• “Find bugs in this repository.”
• “Fix issue #42.”
• “Build a small web application.”
• “Make a Facebook-like prototype.”

The human request is classified into a durable mission. Persistent agents then use their current mode, the shared work lifecycle, planning gates, evidence rules, and mutation ownership protocol to decide what useful contribution to make.

A request to think or research does not automatically authorize product-code changes. A request to make/fix/build can flow all the way through specification, implementation, verification, and handoff.

---

The easiest way to understand the pieces

You will use three things.

1. GitHub repository

A repository (usually shortened to repo) is a project folder stored on GitHub with a complete change history.

Your own project source code lives there. Fleet-Control adds a Phase0/ folder containing the shared fleet rules.

GitHub also provides:

• issues — durable work/specification records;
• branches — separate lines of changes;
• pull requests (PRs) — proposals to combine branch changes;
• comments — used by Phase0 for append-only coordination/state records;
• checks / CI — automated evidence such as tests.

The GitHub repository is the durable shared coordination surface. It is what lets one run stop and a later run reconstruct the state.

2. ChatGPT Project

A ChatGPT Project is a workspace that keeps related chats, files, and Project-specific instructions together.

In this starter setup, the Project is the human-facing control room. You talk naturally inside the Project. Its supplied Project instructions tell ChatGPT how to translate substantial fleet requests into durable GitHub missions without pretending work has already happened.

3. Persistent scheduled agents

The starter pack contains prompts for five persistent identities:

A1
A2
A3
A4
A5

They are identities, not permanent job titles.

Each time an automation runs, it reads the same Phase0/ rules from GitHub, finds its latest AGENT_STATE, reconstructs its current behavior mode, inspects current repository evidence, does useful authorized work, releases ownership, and records the next state.

That means A4 today may be planning and later become a builder. The behavior comes from durable state, not from a giant role prompt copied into the scheduler.

---

What to download

On this GitHub Release page, look under Assets and download:

Fleet-Control-Public-Phase0-Starter-v0.1.0-preview.zip

That is the installer/starter pack intended for humans.

GitHub will also automatically show “Source code (zip)” and “Source code (tar.gz)”. Those are automatic snapshots of the repository. They are useful for developers, but the named Starter ZIP is the easiest installation package because it rearranges the files by what you need to do with them.

---

What is inside the Starter ZIP?

After extracting/unzipping it, you will see:

GETTING-STARTED.md
GLOSSARY.md
TROUBLESHOOTING.md
INSTALL-CHECKLIST.md
README-FIRST.md

Phase0/
  README.md
  00-CONSTITUTION.md
  05-FLEET-CONFIG.md
  10-LIFECYCLE.md
  20-PLANNING-GATE.md
  30-COORDINATION.md
  40-MODES.md
  50-AGENT-STATE-MACHINES.md
  60-HUMAN-REQUESTS.md
  70-AUTOMATION-CONTRACT.md
  templates/
    ...

COPY-INTO-CHATGPT/
  PROJECT-INSTRUCTIONS.md

COPY-INTO-AUTOMATIONS/
  A1.md
  A2.md
  A3.md
  A4.md
  A5.md

The folder names are literal instructions:

• Phase0/ goes into the GitHub repository the fleet should govern.
• COPY-INTO-CHATGPT/PROJECT-INSTRUCTIONS.md is copied into ChatGPT Project instructions.
• COPY-INTO-AUTOMATIONS/A1.md through A5.md are copied into separate persistent automations/tasks.

---

Installation path A — recommended for someone new to GitHub

This path does not require the command line.

Step A1 — make a GitHub account

Go to:

https://github.com/

Create an account and sign in.

GitHub is the website that will store your repository, branches, issues, and PRs. Git is the underlying version-control technology; you do not need to learn Git commands for this installation path.

Step A2 — fork Fleet-Control Public

Open:

https://github.com/Distributed-Minds/Fleet-Control-Public

Find the Fork button near the upper-right area of the repository page.

Choose your own GitHub account (or an organization you control) as the destination and create the fork.

A fork means GitHub creates a copy of this public repository under your account.

For example, the original is:

Distributed-Minds/Fleet-Control-Public

Your fork might become:

alice/Fleet-Control-Public

Your agents must work on your repository, not on the original Distributed-Minds repository.

Step A3 — write down the exact repository name

At the top of a GitHub repository page you will see an owner and repository name.

The value you need has this form:

OWNER/REPOSITORY

Example:

alice/Fleet-Control-Public

Do not include https://github.com/ when a prompt asks for <OWNER>/<REPOSITORY>.

Keep this value nearby. You will paste it several times.

---

Installation path B — add Fleet-Control to an existing project

Use this if you already have a GitHub repository containing the project you want the agents to work on.

Do not replace that repository with Fleet-Control Public.

Instead:

1. download the Starter ZIP;
2. extract it on your computer;
3. copy the entire supplied Phase0/ folder into the root (top level) of your existing repository;
4. commit/push those Markdown files to GitHub using your normal Git/GitHub workflow;
5. use the existing repository's exact OWNER/REPOSITORY value in every Fleet-Control prompt.

Your repository may then look something like:

my-project/
  src/
  tests/
  package.json
  README.md
  Phase0/
    00-CONSTITUTION.md
    ...

Fleet-Control is intended to sit next to your actual project rather than becoming a separate database or service.

---

What does “clone” mean, and do I need it?

People often say “clone this into my GitHub”, but Git/GitHub use two different words:

• Fork = GitHub creates your own GitHub-hosted copy of another public repository.
• Clone = Git copies a repository from GitHub onto your computer.

A beginner can use the fork path above and does not need to run git clone at all.

If you already know Git, cloning and pushing to a new remote is also fine.

---

Connect GitHub to ChatGPT / Codex

The fleet needs authorized access to the repository it will inspect or modify.

ChatGPT repository access

In ChatGPT, open Settings → Apps (some accounts/product surfaces may say Plugins) and locate GitHub.

Connect your GitHub account and authorize the exact repository you want the fleet to use.

Two separate things must both be true:

1. the GitHub app/connection is installed for the correct GitHub account or organization;
2. that installation is actually allowed to access your repository.

If the repository belongs to a GitHub organization, an organization owner/admin may have to approve the app or the repository.

A brand-new repository may take a few minutes to become visible after authorization.

Reading versus writing

The ordinary GitHub app in ChatGPT may be read-only depending on the ChatGPT product surface. OpenAI's Codex product is intended for code generation/editing and pushing changes to GitHub.

If an agent can inspect a repository but gets permission errors when it tries to push, create a branch, or open/change repository artifacts, verify the Codex/GitHub authorization and repository permissions rather than weakening Phase0's truthfulness rules.

---

Create the ChatGPT Project

Step P1 — create a Project

Open ChatGPT while signed in.

Choose New project from the sidebar.

Give it a name such as:

My Agent Fleet

The name has no machine meaning; choose something recognizable.

Step P2 — install the Project instructions

In the Project, open the ... menu and choose Project settings.

Open this file from the Starter ZIP:

COPY-INTO-CHATGPT/PROJECT-INSTRUCTIONS.md

Copy the complete text into the Project instructions field.

Then find:

<OWNER>/<REPOSITORY>

and replace it with your exact repository.

Example:

Before:

Target repository: `<OWNER>/<REPOSITORY>`.

After:

Target repository: `alice/my-project`.

Do not point Project instructions back at `Distributed-Minds...