Scenario / test fixture: Tink

https://github.com/npm/tink

Should work like existing `multiple-roots` scenario.

Tasks:

- [ ] The abstract dependency graph is not going to be automatically inferrable because everything is or or less flat.
- [ ] Figure out if this is an actual problem and if we should do anything about it (docs, logic, special report, etc.)
- [ ] Investigate if there is metadata from which we _could_ recreate the abstract dependency graph. (Probably could from `yarn.lock` or `package-lock.json` if nothing else).