built a claude code plugin around typesafe's jev, the new model that returns probabilities instead of prose jev-shadow asks it 5 hazard questions about every command claude proposes and logs the answers next to what claude code's own classifier decided. zero latency, changes nothing
github.comGitHub - dreyfus92/jev-shadow: claude code plugin that measures jev before trusting it: background verdicts, joined with the classifier's real denials, and a reportclaude code plugin that measures jev before trusting it: background verdicts, joined with the classifier's real denials, and a report - dreyfus92/jev-shadow
0 reposts2 likes