Also, FWIW you can access an aggregated javadoc at some point in the dist module:

https://github.com/hibernate/hibernate-search/blob/1e2805ed73c2fed14825432212dbfa44b31f4fa5/distribution/pom.xml#L141-L171

Maybe you can move that to the documentation module, and apply your processor on the result? That way you would not generate the aggregated javadoc twice. That would require some more work so that the dist module correctly includes the javadoc in the zip, though.