Scenario Testing for AI Agents
Score how similar N agent outputs are — exact match, Jaccard token overlap, divergence point, composite 0-1 score. Stdlib-only.
