Problem:

Currently Checkers check for plaintext. Sometimes it might be useful to branch from a Checker to further resolve the detected plaintext. The most common example probably would be a JSON where values consist of ctexts. It would be nice, if one could have the possibility to decode those values further.

Is there a possibility to achieve something like that. So far I haven't found any, but maybe someone knows more.

Solution:

One (1) could either allow branching of decoders, where decoders are allowed to output a list of ctexts to further decode or (2) one could allow checkers to start new runs for subfields like values in a dictionary. I think the second approach is conceptually cleaner.