Test Fixing

When your test suite breaks after a refactor or code change, this gives Claude a methodical framework to fix everything systematically. It runs the full suite first, groups failures by error type and root cause (ImportErrors, AttributeErrors, etc.), then tackles them in priority order rather than randomly. The workflow is smart: fix infrastructure issues like imports before logic bugs, verify each group passes before moving on, use focused test runs to avoid waiting for the full suite every time. It's basically a checklist that prevents the common mistake of fixing tests randomly and creating whack-a-mole situations where fixing one thing breaks another.