That's a broader definition of legacy than I’ve usually seen used; usually, “legacy code” is a fuzzy categort of code that lacks some significant subset of the maintainability properties expected as the norm in the category of “existing, actively maintained code”, such that issues in that code requiring working around the code, replacing code units wholesale, or expending extra effort research, document, develope tests for, or otherwise bring the code into a maintainable state to enable diagnosing and resolving issues with it.
That's a broader definition of legacy than I’ve usually seen used; usually, “legacy code” is a fuzzy categort of code that lacks some significant subset of the maintainability properties expected as the norm in the category of “existing, actively maintained code”, such that issues in that code requiring working around the code, replacing code units wholesale, or expending extra effort research, document, develope tests for, or otherwise bring the code into a maintainable state to enable diagnosing and resolving issues with it.