Test 2: Verifying action items count for Michal...
Found 5 ground truth file(s)

✓ 1.json: Action items count OK (ground truth: 1, output: 1)
✓ 2.json: Action items count OK (ground truth: 0, output: 1)
✓ 3.json: Action items count OK (ground truth: 0, output: 1)
❌ 4.json: Insufficient action items
    Expected at least 1, got 0
✓ 5.json: Action items count OK (ground truth: 0, output: 2)

============================================================
Results: 4/5 files pass
============================================================
FAILED: Some files have insufficient action items