The sequencers that ModeShape provides currently process the BINARY content of whatever node is supplied as input. However, some repository designs may intermix files of different MIME types into the same area and not use (or even know about) file extensions, instead relying upon the content of the files (rather than the path or node names) to determine the type of file.

The sequencers themselves or sequencing framework should be able to deal with this. For example, a sequencer could just check the MIME type of the BINARY content, or even look for the "mix:mimeType" mixin (and its "jcr:mimeType" STRING property) to see if the sequencer can process that MIME type. Or, the Sequencer class could define a "getAcceptedMimeTypes(): Set<String>" method that an implementation could implement and the Sequencers class could check the MIME type before calling the sequencer.