Create a duplicate track of the vocal. Run that track through the eq. Set the eq to boost the offending frequencies.
Set up a compressor on the other (actual) vocal track. Patch the output of the eq to the sidechain input of the compressor. Tweak the compressor settings so that it only responds to the sibilance.
I got this by downloading a Behringer manual for one of their compressors. It was one of the best manuals I've read believe it or not. You hook your input source into your compressor like usual (mic or track) and take the output of the compressor back into your recorder or PA or whatever. The side chain jack is really a stereo in/out jack so you need a stereo plug with 2 mono cables attached. On of the plugs goes to the EQ input and the other goes to the EQ output. The compressor now sends the signal to the EQ and receives it through the other side of the plug. The difference now is that the compressor pays no attention to the incoming signal from the mic or track; it works off the signal returned from the EQ. If the EQ is left flat then the compressor works the same as if it work working off the track or mic signal. If however you cut all frequencies accept the ones that punch out the syllabence like sass and t's. Say cut everything accept 3K-10K then it will compress those sounds and leave the others alone. This takes a lot of tweaking but it does work.
