Keeping individual file results reduces the cost of future scans and lets you review file-level findings later.
Use deletion only when the codebase is sensitive and is not open source or otherwise publicly accessible. Results shared with another cached repository, such as a common open source library, are preserved.