The Spider plugin analyses responses to identify any links in the response body, or the "Location" header. If the URL represented has not been seen, the URL is added to a tree, and can be automatically downloaded when desired. WebScarab has two modes of fetching unseen links. "Fetch Tree" enumerates all currently unknown links below the selected node, and queues them for retrieval. "Fetch Selection" queues only the selected nodes for retrieval.
Future enhancements should also identify HTML forms, and present those to the user for manual modification and submission. This would be a great project for an interested user to implement.