Can a decision model prune your coding agent's context? What fast-jev-compaction keeps, costs and drops

fast-jev-compaction prunes in a second, not 30. At its default threshold it keeps almost no tool results, and a scorer that always says no cuts as much.