Implement incremental saving in the Spider service to prevent data loss during long discovery runs.

Changes:

• Spider.mjs: Save tracker.json (candidates) and visited.json (graph) after processing every page of search results (or every N repositories).