The read capabilities on Archive are largely inadequate. You can get the content by an archive path or you can flip through all the content and filter the results by path name. However, there are much more logical structures inside the archive that ShrinkWrap does not expose an API to locate. For instance, I want to pull out one or more WebModules from an EnterpriseArchive.
However, if we were to introduce a closed set of read methods, the user would eventual outgrow them and find themselves in the same position. My recommendation is to follow the lead of Weld Extensions and use an Archive criteria API. I believe it's a good choice because it feels familiar to Java EE developers with it's JPA-like syntax and it's completely extensible.
ArchiveContentQuery<EnterpriseArchive> query = ArchiveContentQueries.<EnterpriseArchive>createQuery(EnterpriseArchive.class)
List<WebArchive> wars = query.getResuiltList();
This is so much simpler than using a filter on a tree walker. It's also crucial for manipulation of the Archive within Arquillian.