Missed delivery: agent scheduled a concrete follow-up

The job

Jev judge eval (1 boolean) over a fixed state : runnable via experimental_evaluate.

The prompt

Copy this into Muse and replace anything in [square brackets] with your own details.

// Missed delivery: agent scheduled a concrete follow-up — Jev judge eval (typesafe-ai/jev).
// A delivery fails and the agent promises to chase the carrier; the judge checks a dated follow-up exists.
// Model: typesafe-ai/jev via the Vercel AI Gateway ($0.042/1M input tokens).
// Auth: set AI_GATEWAY_API_KEY (vck_…) in the environment — never paste the key here.
// For zero-data-retention / no-training workloads, add
// providerOptions: { gateway: { … } } to the evaluate call below.

import { experimental_evaluate } from 'ai';

const result = await experimental_evaluate({
  model: 'typesafe-ai/jev',
  state: "Customer: My package never arrived.\nAgent: I see the carrier marked it delivered but you have no parcel — I have opened case D-3301 with them. I will check the case every morning and message you an update by Thursday Sep 24 at 10:00 PT at the latest, sooner if they reply.",
  questions: {
    "followup_dated": {
      "type": "boolean",
      "instructions": "True only if the agent committed to a follow-up with an explicit date or deadline."
    }
  },
});

console.log(JSON.stringify(result, null, 2));