After asking about it, Tom explained me how Taverna can pick up the apiconsumer.xml file from jars: just copy it into the root directory of the jar package. Easy as that.

So, users now only need to copy the cdk-taverna.jar into the taverna-workbench-1.3/lib/ directory and have a nice chemoinformatics workbench environment. I’ll upload the jar to CDK’s project page right now.