Currently the automatically extracted licensing information for frontend dependencies has mixed quality. This requires Legal to manually find and add the missing data before a release. We should improve this in one following directions:

1. Review the frontend implementation and compare it to the solution for the modeler. The quality for the modeler is much better according to Uli
2. Switch from a fully automated to a semi-automated solution, i.e. make it possible that licensing information can be manually provided and is then always included in the license book (so Legal only has to review new/changed dependencies)