LYGO lattice overview for newcomers

By Justin Helmer · reviewed 18 October 2026 · No install required to understand this page

Start with the problem LYGO cares about

Creative and technical work with AI spreads across chats, repos, music portals, and agent skills. Context fragments. Trust decays. Mythic language can either clarify a system, or confuse newcomers. The LYGO project family tries to keep a living map of that work: documentation, free tools, champion personas, and optional deeper protocol experiments.

Pieces you will meet on chatagent.ca

What you can ignore at first

You do not need kernel eggs, mesh gossip, or a full protocol stack checkout to use a champion prompt or to understand Continuum. Those deeper layers exist for operators who want modular, consent-gated infrastructure. Newcomers should stay on guides + free tools until curiosity pulls them further.

Values that show up repeatedly

• Local authority, your machine and consent matter more than automatic cloud actions.
• Receipts over vibes, hashes, ledgers, and capsules beat uncheckable claims.
• Human remains the publisher. AI assists; you decide what ships.
• Dual channels, public-safe surfaces vs engineer RAW packages, clearly labeled.

Suggested first hour

1. Read About (5 minutes).
2. Skim all fifteen champions (10 minutes).
3. Follow summon steps on one real task (20 minutes).
4. If you use coding agents, open the Continuum guide (15 minutes).
5. Bookmark Contact and Privacy.

External hubs (optional)

• eternalhaven.ca, broader Haven surfaces
• Excavationpro hub
• LYGO Protocol Stack (verifiable source)
• lygo-protocol-stack on GitHub
• Grokipedia (LYGO is indexed and growing): Protocol Stack · Champions · Protocol · LYGO-LANG

Mythic names (Δ9, lattice, Lightfather) are part of the project’s creative system. On this publisher site we deliberately explain them in plain English so newcomers are not locked out.

A larger map, without the subway

Beyond chatagent.ca you may see Eternal Haven pages, music portals, protocol-stack documentation, and ClawHub skills. Treat them as neighborhoods of one city: you can visit one café without memorizing the subway. When a page assumes jargon, return here or to the resources hub.

The important boundary is not “inside LYGO” versus “outside LYGO.” It is public page versus canonical record versus private machine. A public guide can explain a tool. A ledger can prove a claim. A local tool can touch your files. Those are three different levels of trust, and they should never be blurred into one.

How the pieces connect

The shortest useful path is: choose a job → open the guide → use the smallest tool → check the result. The longest path is to read every page first. The second path feels thorough and usually ends with no work done. Start with the job.

Six words that unlock the rest

Canon

The record the project treats as authoritative: the dual ledgers, the kernel eggs, and the Star Chart feed.

Resource

A public page or file that can be read and cited, but is not canon by itself.

Shadow

A room or file that could not be reached. Named, not filled with an invented story.

Seal

A hash or receipt that lets another person re-check a claim.

Champion

A role or persona used to frame a kind of work. Not a separate person and not a claim of machine sentience.

Kernel egg

A packaged, hash-addressed component with a manifest and a verification step. An egg without a hash is just a zip file.

Once those six words are clear, most of the site is readable. The remaining jargon is either a tool name or a piece of lore; both are explained on the page where they appear.

Three first tasks (about an hour)

1. Summon one champion for one real task

1. Pick a job you actually have: rewrite a paragraph, review a plan, or name the risks in a decision.
2. Open the summon app and choose the champion whose role matches the job.
3. Copy the prompt. Add your task brief: goal, audience, constraints, and the check that makes the output acceptable.
4. Keep the champion, the brief, and the decision in one note. A summon you cannot review later is a mood, not a method.

What this teaches: the champion is a frame, not an authority. The quality still comes from the brief and the review.

2. Seal one file with Continuum

1. Choose a file you would hate to lose or misdescribe.
2. Open Continuum and follow the page’s current command for creating a capsule.
3. Add a claim that can be checked: a file hash, a required string, or a JSON path.
4. Run the verify step. Then change one byte of a copy and run it again. The second run should fail.

What this teaches: a “done” claim is only useful when it can fail. If the check passes no matter what you do, it is decoration.

3. Verify one public ledger

1. Open the ten-minute verification walkthrough.
2. GET the two ledger URLs and confirm the JSON parses.
3. Record chain_valid, entry_count, and chain_root.
4. Run the public gate at /join/ and read the result as ALIGNED, DRIFT, or SHADOW.

What this teaches: public verification does not require an account. It requires a source, a date, and the honesty to name what you could not reach.

How to tell lore from evidence

The project uses mythic language on purpose: it makes a system memorable and gives people a way to talk about roles. The danger is treating a memorable name as a fact. Use this table as a filter.

What you can safely ignore at first

You do not need the mesh gossip layer, the kernel-egg planter, the biometric harness, or a full protocol-stack checkout to use a champion prompt or to understand Continuum. Those are operator tools for people who want modular, consent-gated infrastructure. Newcomers should stay on guides, free tools, and one real task until curiosity pulls them deeper.

You also do not need to agree with the lore to use the software. The stack can be evaluated as code; the books can be read as fiction; the champions can be used as roles. Keeping those categories separate is a feature, not a loophole.

FAQ

Do I need to install anything?

No. The guides, the summon app, the Continuum page, and the public ledgers work in a browser. A local install is only needed when you want to run a tool on your own files or use the LLM console.

Is there a cost?

The public guides and tools are free to use. Some FULL engineer packages are supporter-gated; the SkillHub page labels them. Optional support through PayPal is never required to read a guide.

Are the champions AI agents?

They are published roles and prompts. You can use them with an AI model, but the champion itself is not a separate running agent and does not prove that a model is sentient.

What is Δ9?

Δ9 is part of the project’s creative and naming system. On this site it is a label, not a scientific measurement. If a page uses it, read the sentence around it for the actual claim.

Why does the site use mythic language?

Because the project is also a creative world. The rule is that the mythic layer must not be used to hide a technical claim. When a page makes a checkable claim, it should link to the check.

Where do I start if I only have ten minutes?

Read this page, then open one guide that matches a job you already have. Do not start with the protocol reference unless you need it; it is a reference, not a tutorial.

Glossary

Mythic names (Δ9, lattice, Lightfather) are part of the project’s creative system. On this publisher site they are explained in plain English so newcomers are not locked out. If a claim matters, follow it to the checkable page.