A 6-step agent production loop with Jev as judge
The author outlines a six-step loop where an LLM writes, Jev judges and code decides: packing state, asking Jev several typed questions in parallel, drafting per branch, verifying, three-tier gating and logging replayable receipts.
99% of AI agents in production are fundamentally broken one model writes the plan, approves its own tool calls, evaluates its own output, and ships it. zero audit trail. zero safety. the stack that fixes it: llm writes. jev judges. code decides. here is the exact 6-step loop: → state - code builds the smallest evidence packet and strips secrets → ask - jev answers several typed questions at once (Choice, Score, Noul) over the same state → draft - the llm only sees the context for the branch jev picked → verify - parsers, tests and allowlists run first, then jev checks narrow properties of the output → gate - thresholds turn probabilities into three outcomes: auto, review, stop → receipt - state, question version, model, route and outcome get logged and can be replayed what makes it different: • a typed answer comes with a probability, so uncertainty is measurable • one request can carry many questions (TypeSafe's cookbook reports roughly 10x faster than separate calls in a 13-question example) • rollout is staged: offline → shadow → assist → limited → expand the breakthrough insight: a confident answer is not permission. when jev lands in the middle band, the harness asks for more evidence or a human instead of letting the model guess. the engineer who ships agents with a kill switch and a receipt for every decision is the one who gets trusted with production. save this architecture for your next agent build.
This post quotes another post: x.com/0xtrackmind/status/2106788767960850798