To reduce the chance of mistakes in test manifests (e.g. #230), we could introduce a test suite validation check.
One option would be to run all tests against a dummy system using a test suite runner (e.g. https://github.com/rubensworks/rdf-test-suite.js).