Running Apex Tests

It defaults to narrow scope first so you're not waiting on full test runs while debugging, then widens to regression once stable. The 120-point scoring system and structured output format (what ran, what failed, root cause, next step) keep the feedback loop tight. Useful when you're stuck on a failing test class, pushing coverage over 75%, or need bulk and callout mock patterns without hunting through old projects. Delegates production code fixes to the generating-apex skill and stays out of Jest or Agentforce territory.