Dumph grabs a tree of tasks from your Phabricator instance and puts them in a CSV spreadsheet or in a HTML file.
Just define and run a query on Phabricator and copy the query id from the result page passing it to Dumph.
$ # run a query on Phab and grab its id or use one of the predefined ones
$ QUERY_ID=open
$ dumph "$QUERY_ID" --output-format=html > results.html
Over time we grew some scripts to grab info from Phab, paying particular attention to parent/child relations. Surely extending Phabricator would have been a better choice, but writing a silly Python script is much easier.
$ python setup.py install
Dumph is licensed under the MIT license.