Putting Jev before expensive steps as a multiple-choice gate
Jev decides who to assign, whether a lead is worth following up, whether an email needs a reply, whether a source is worth reading, and whether to approve, reject or escalate, while Claude does the research and coding.
whoever leaked this is f*cking crazy save this for your next agent build. Jev handles decisions with a fixed set of answers: > which agent should take this? > is this lead worth a follow-up? > does this email need a reply? > is this source worth reading? > approve, reject or send to a human? Claude still does the research, coding and writing. Jev helps decide which tasks deserve that work. the useful bit: you can put it BEFORE an expensive step. before Claude reads 100 documents, screen t