# Correct scope by adopting again, not by amending a run

Amended by [ADR 0022](https://repo-standards.com/docs/decisions/0022-replace-a-stopped-run-in-place/): the new run
replaces the stopped one in place, without abandoning it and resolving its
changes first.

Scope amendment let an active run add discovered scope without abandoning its
progress, at the cost of a preview command, an amendment format, a third
inspection format, revalidation, and replay. No real adoption used it. With
confirmation identity bound to content ([ADR 0008](https://repo-standards.com/docs/decisions/0008-bind-confirmation-identity-to-what-the-run-reads/)),
read-only inspection cheap, and fixes repeat-safe by contract, a mistaken scope
is corrected by abandoning the run, committing or discarding its changes, and
adopting again with a new confirmed scope. The trade-off is a rough edge: an
abandoned run leaves uncommitted changes that must be resolved before the next
start. That cost is paid rarely; the amendment machinery was paid on every
release.
