jevbooks

← All projects

jev-judgment

HyunjunJeon/jev-judgment

Send closed coding-agent judgments to TypeSafe Jev

What it is

An Agent Skill for coding agents that sends closed judgments to TypeSafe Jev and acts on the returned probabilities, so the main coding model keeps writing code and interrupts the user only when a person must decide.

How it uses Jev

Jev is called at three points: before asking the user a closed question, before a command that may destroy data, leak secrets, or reach beyond the task, and after a command fails. The bundled scripts/jev.py posts {state, questions} to the TypeSafe API and uses the probabilities to proceed or ask.

Primitives:choicescorenoul

Technique worth stealing

Keep the main coding model writing code while a separate typed judgment model answers closed questions.

Try it

npx skills add HyunjunJeon/jev-judgment, then set TYPESAFE_API_KEY

View on GitHub

judged by Jevjev-1.13.0

Evidence

Each line is one question put to Jev about the README. ≥ 0.60 reads as yes, ≤ 0.40 as no; in between Jev is not making a call.

  • Jev-centricyes0.92
  • Shows a System One patternyes0.79
  • Handles uncertaintyno0.13
  • Measuredno0.32
  • Runnableyes0.97
  • Worth recommendingno0.39
  • Model replicano0.03
  • Problem scopescore on a 0–2 scale1.38
  • About Jevyes0.98

Signals by Jev jev-1.13.0, card written by DeepSeek V4.1 Flash from the README on 20 Sept 2026.