I would expect that SOLR could actually do the indexing but knowing what needs to be indexed is still a real issue. I think an ontology of software that includes more WHY particular things are done in programs or data needs to be included in current Software Engineering practices.
E.g. if someone built an API that used LibreOffice code to convert Office files to PDFs and hosted it using AWS, would they think to check to see if someone can reference the metadata URL and read their EC2 instance credentials? Or reference file:// URIs and read from the server's filesystem?