cerbia.core.score_aggregators.MaxScoreAggregator implements worst-case aggregation: the aggregate is the largest severity-weighted blocking score.
aggregate = max(scores)
For example, [0.20, 0.85, 0.45] produces 0.85. An empty list produces 0.0. This is the simplest choice when one high-confidence blocking finding is enough to determine the result.
score_aggregator:
score_aggregator: cerbia.core.score_aggregators.MaxScoreAggregator