Symptom

When using SafetyCriticalPolicy, severity 5 incidents get scores in the hundreds or thousands -- way beyond the expected 0-100 range.

Context

The scoring system uses a Template Method pattern. The base Policy class orchestrates the scoring pipeline:

1. baseScore() -- subclass computes initial score
2. applySeverityWeight() -- multiplies by clamped severity weight
3. Final rounding

The base applySeverityWeight() clamps the multiplier via Math.min(severity * 2, 10). SafetyCriticalPolicy overrides this method.