Currently (as of version 2.3.0), the Search and Replace task only supports the replacement of a single character or a collection of contiguous characters (such as a word or sentence). So, the only way to achieve this effect within GoAnywhere is to add one Search and Replace task for each character or word you wish to replace. This approach, however, would cause the project to scan the entire file once for each search and replace task within the project.
All data translation tasks, such as CSV, XML, Fixed-Width, etc... support an encoding option that allows you to specify the character set of the file you are working with. This allows GoAnywhere to work with these special characters, which, in many cases, eliminates the need to remove such characters from the data.
Would you provide more information as to why you need to change the character set? And what outcome are you trying to achieve? There may be an alternative approach depending on what you need.