With good tests and other static analysis and linters in place, you can let computers take care of pointing out the obvious. Human review can then focus on what people are good at.
pullapprove_conditions to make sure the author
takes care of everything your automated checks can help with
before bothering real people.
version: 3 pullapprove_conditions: - condition: "'*travis*' in statuses.successful" unmet_status: "failure" explanation: "Tests must pass before review starts"