There are multiple approaches being used in the ability spec in editor-api at the moment, which makes it hard to reason about and we are definitely missing coverage in places.
There are multiple approaches but as a prompt these are worth considering:
- Test from a resource / model perspective, and break them down per user
- Test from a user perspective, ensuring each ability is individually tested
Whichever we choose I think consistency will be key to ensure each ability has good coverage.