The Problem

A Git repository often contains only a part of a working system. A real-world website may depend on a database, uploaded files, object storage, users/Auth, runtime configuration, scheduled jobs, access policies, and the state of external services.

This is clearly visible in the previously researched MODX: a significant part of the website resides in the DB, and another part in the file system. The same issue arises when migrating modern AI-builder applications: having the source code does not yet mean having a reproducible system.

What We Want to Achieve

We need to form a practical model of the project state that allows us to understand what needs to be saved, restored, or transferred so that the project continues to work regardless of the current infrastructure.

A single mandatory set for all websites is not assumed. The goal is to learn how to discover significant parts of the state and explicitly capture them.

Related to runtime, backup/restore, secrets and external dependencies, and project portability. For MODX, a specific case is already being investigated in the task of restoring from dumps and files.