Agent tooling

jev-model-router-demo

@jcpsimmons2JavaScriptupdated 2026-09-17runnable

Throwaway Jev demo: route coding tasks to Grok Build or Codex Astra

jcpsimmons/jev-model-router-demo

Where it calls Jev

import { experimental_evaluate as evaluate } from 'ai';

router.mjs:7

The link points at the commit we read, so the line number still holds.

What it asks Jev

These question sets are lifted from this project's source exactly as written.

from router.mjs:82

  1. complexityscore

    Rate the reasoning and implementation complexity of this task.

    • 0mechanical: one obvious localized change with direct verification
    • 1contained: limited reasoning across a few known files or steps
    • 2systemic: multiple components, unclear diagnosis, or meaningful tradeoffs
    • 3high consequence: architecture, security, migration, or broad ambiguous work
  2. needsPlanningtrue/false

    Should a capable engineer plan or investigate before changing code?

Run it on your own text

The code
import { experimental_evaluate as evaluate } from 'ai';

const { answers } = await evaluate({
  model: 'typesafe-ai/jev',
  state,
  questions: {
    complexity: { type: 'score', instructions: 'Rate the reasoning and implementation complexity of this task.',
      criteria: ['mechanical: one obvious localized change with direct verification', 'contained: limited reasoning across a few known files or steps', 'systemic: multiple components, unclear diagnosis, or meaningful tradeoffs', 'high consequence: architecture, security, migration, or broad ambiguous work'] },
    needsPlanning: { type: 'boolean', instructions: 'Should a capable engineer plan or investigate before changing code?' },
  },
});

Other projects in this category