jev-benchmark

A reproducible benchmark testing TypeSafe's Jev on a real agent-infrastructure task — classifying a proposed tool call as readonly, destructive, privileged, or exfiltration — measuring accuracy, latency, and whether the confidence score is actually reliable enough to route decisions on.

unverified — sourced from awesomejev.com; description independently written by AionEdge from the repo’s own GitHub metadata and README, not GitHub-API relevance-checked