D007 - The target is a manifest source, and the path is asked for rather than assumed

decided - 2026-08-26 - adding pros manifest --from-target

The payload manager keeps its own repository description, and this project copied that schema rather than inventing one precisely so that a target which is already configured is already described. Reading it is one command now.

It also does something no other command here can: it is the only way this project will find out what that document actually looks like. Its field names are known; whether it is a list, an object keyed by name, or a wrapper around either is not, and neither is the format of its checksum field. Both open questions in D004 are answered by running this once against a real target.

The path is a parameter, not a default

/data/pldmgr/repository_cache.json is a good guess - the manager's autoload list is in that directory - and a guess is what it would remain. A default carries the authority of a measurement without having made one, and when it turned out to be wrong the failure would be this target has no repository rather than this program looked in the wrong place.

So it is asked for. When somebody runs it against real target and it works, the path becomes a default with a measurement behind it, and this entry is where that change gets its justification.

It reads over the file service rather than the dashboard

The manager also answers on a web port, and that would be the obvious route. It is the worse one: the endpoint paths that service answers on have not been measured either, so that version would be a guess wrapped around a guess. The file service needs no endpoint - a file has a path, and the path is the thing being asked for anyway.

What the test proves, and what it deliberately does not

A stand-in serves a manifest in this project's shape, and the command reads it, reports it, and judges its entries exactly as it judges one read off a disk. That proves the plumbing.

It proves nothing about the real file, and the wording of the test says so, because a test whose name suggests otherwise would be the same kind of claim this entry exists to avoid.