The Agent Wrote a Test That Mocks the Function Under Test

When an AI agent writes unit tests, the most common failure isn't weak assertions — it's mocking the thing you actually want tested. Here's what to look for.