A verification layer that lets AI agents safely delete code in large codebases by assessing usage paths and providing risk statuses.
