Injection-based testing while refactoring is a pattern that minimizes the need for manual editing when altering the behaviour of a code base. Neither does it rely on a compilation or a linking process nor does it make assumptions on the structure of the code. Thus, it can be particularly useful for refactoring code that has been written in scripting languages, and specifically targets the research and engineering context. We describe the pattern and propose a set of functions for its application. The applicability of code injection for refactoring is highlighted via specific examples for deriving unit and integration tests. Finally, we comment on the customizing of the pattern and give practical advice for its implementation.