the only copy was on the branch being deleted

A clean-up was scheduled to retire a set of working copies and the branches behind them. Somewhere in that set were result artifacts that had never been merged. They existed only on a branch, inside a working copy due for disposal, and nowhere else at all.

The recovery was not clever: land the artifacts on the mainline first, with a note recording why they were being rescued and where they had come from. The interesting part is that this had to be found by looking. Nothing in a branch listing marks a file as the only copy of itself; the test is whether the content exists elsewhere, and that requires checking rather than assuming.

The guard came after the incident rather than before it, which is honest about how the knowledge arrived. A deletion path now refuses to remove a branch that is not merged, from a working copy as well as from the central point, and the refusal was proven by attempting the deletion and watching it decline. The sanctioned clean-up route performs the worktree and merge checks before it will proceed.

What it proves: history is treated as primary evidence and deletion as a dangerous tool, and the preference is to make an irreversible action refuse rather than to be careful around it.