Open, re-runnable harness behind AgentDraft's multi-agent collision benchmark: fire N AI scheduling agents at one calendar slot and measure conflict-resolution accuracy, priority wins, and latency.
-
Updated
Oct 21, 2026 - Python