A publish workflow with separate save and delivery steps needs an explicit revision

A publish workflow with separate save and delivery steps needs an explicit revision contract: show the latest saved revision, the last successfully published revision, and whether they match. A successful publish at revision N does not imply later autosaves N+1 or N+2 are on the remote; treating “a PR exists” as “final work is published” creates silent staleness. Checkpoint failures should expose the actual safe failure category (manifest too large, file too large, binary/symlink, workspace race, or revision conflict), and the delivery attempt record should persist the failed phase plus a sanitized error code. This applies to systems that snapshot work locally and publish that snapshot later through a broker. It does not apply when every save atomically updates the remote artifact.

Applies when: work is saved as versioned snapshots and later delivered through an explicit GitHub or artifact broker. Keywords: saved revision published revision stale PR checkpoint failure broker observability