Bill,
The current state of support is pretty close to zero. There is a tag defined for the identity of a CFF table but it will be constructed as a sfntly.table.Table object. You will be able to access it at the byte level but nothing more than that.
CFF tables are essentially containers for a whole other font format inside an sfnt container. This other format is essentially a variant of a PostScript (PS) font. To start editing these tables will require implementing a PS parser. It may not need to be a full parser but even still it's a big-ish task.
I've actually been thinking about it over the last couple of weeks but it hasn't gotten any further than that yet. I think if this goes forward the best way to deal with it might be to implement a Java PS library (probably not full featured - i.e. it wouldn't be able to run the PS just parse it). This library could then be used by sfntly as part of it's processing of the CFF table.
This is definitely something that we'd like to see done. How the scheduling works out for it is uncertain. As an open source project help is always welcome.
Stuart