For assistant handoffs

AI Quick Start

This page gives AI assistants a safe reading order and the core boundaries needed before proposing Abyssal Engine changes.

Read First

  1. Start with Design Principles to understand how decisions are judged.
  2. Read Architecture and Architecture Diagrams for ownership boundaries.
  3. Use Assets, Editor, and Gameplay to locate the system being changed.
  4. Check Glossary before renaming concepts or inventing new terms.

Rules of Engagement

Preserve architecture, make small truthful changes, update stale call sites instead of adding compatibility aliases, and validate with TypeScript.

Asset Boundary

Assets store authored intent. Runtime Actors, Components, renderer caches, and temporary diagnostics are not asset data.

Editor Boundary

Editor panels present and author data. They should not become hidden gameplay authorities.

Runtime Boundary

Gameplay services resolve saved definitions into live Components and Actors. Do not serialize runtime-spawned match state into reusable assets.

Preferred Change Shape

Audit the current code path, change the smallest owner responsible for the behavior, run validation, and report risks that should be tested manually.

Related Reading