The cloud manifest is the table of contents for the archive. It lists the visible sections, points readers toward the project index, and gives scripts a stable place to discover public data files.
The manifest does not try to describe every file. It describes sections: projects, archive cards, notes, status and changelog entries. That keeps it readable and avoids turning the public API into a directory dump.
When a new section is added, the manifest is updated alongside navigation and sitemap entries. The page and the JSON file should tell the same story in different formats.
The archive is refreshed by regenerating static files. That makes the public output easy to inspect because each page can be opened directly and each JSON file can be checked without a session.
Build-time data is honest about its limits. A status snapshot says when it was generated, while a changelog entry explains what changed in the public content tree.
This approach favors boring durability: fewer moving parts, readable documents, and a small set of public routes that can be tested with ordinary HTTP tools.
