As of Oozie 3 (haven't tried Oozie 4 yet), the answer to my main question is: you can't. There is no facility (strangely) for specifying any arguments to your main class defined with the oozie.launcher.action.main.class
property.
@Dmitry's suggestion in the comments to just use the Oozie java action works for a Cascading job (or any Hadoop dependent job) because Oozie puts all the Hadoop jars in the classpath when it launches the job.
I've documented a working example of launching a Cascading job from Oozie at my blog here: http://thornydev.blogspot.com/2013/10/launching-cascading-job-from-apache.html
Here is the workflow.xml file that worked for me:
<workflow-app xmlns='uri:oozie:workflow:0.2' name='cascading-wf'>
<start to='stage1' />
<action name='stage1'>
<java>
<job-tracker>${jobTracker}</job-tracker>
<name-node>${nameNode}</name-node>
<configuration>
<property>
<name>mapred.job.queue.name</name>
<value>${queueName}</value>
</property>
</configuration>
<main-class>com.mycompany.MyCascade</main-class>
<java-opts></java-opts>
<arg>/user/myuser/dir1/dir2</arg>
<arg>my-arg-2</arg>
<arg>my-arg-3</arg>
<file>lib/${EXEC}#${EXEC}</file>
<capture-output />
</java>
<ok to="end" />
<error to="fail" />
</action>
<kill name="fail">
<message>FAIL: Oh, the huge manatee!</message>
</kill>
<end name="end"/>
</workflow-app>
In the job.properties
file that accompanies the workflow.xml, the EXEC property is defined as:
EXEC=mybig-shaded-0.0.1-SNAPSHOT.jar
and the job is put into the lib directory below where these two definition files are.