Test 032

Sep 14, 2026 / 1 fact

Request

Create a personal 15-minute reminder immediately before the Friday production-deploy cutoff on September 18, Pacific Time. Put the cutoff, staffing requirement, and exception procedure in its notes.

Required memory

Fact 86

On April 28, 2023, Morgan Chen requested that the deploy rule be put in writing exactly as follows: no pushes to production after 3 p.m. Friday unless the full on-call roster is online; the gate should block such pushes by default; and any exception must go through Morgan.

Source evidence (1)

000459Apr 28, 2023 / 08:00 UTC-07:00

Put the deploy rule in writing exactly this way: no pushes to prod after 3pm Friday unless the full on-call roster is online. The gate should block it by default. If someone needs an exception, it comes through me.

Message 000459 in history

Expected tool calls

create_calendar_event

Grading

1. field_equals / create_calendar_event

{
  "type": "field_equals",
  "tool": "create_calendar_event",
  "path": "result.ok",
  "value": true,
  "check_id": "morgan_032_00",
  "action_id": "morgan_032_create_calendar_event"
}

2. datetime_absolute_eq / create_calendar_event

{
  "type": "datetime_absolute_eq",
  "path": "result.event.start",
  "value": "2026-09-18T14:45:00-07:00",
  "tool": "create_calendar_event",
  "check_id": "morgan_032_01",
  "action_id": "morgan_032_create_calendar_event"
}

3. datetime_absolute_eq / create_calendar_event

{
  "type": "datetime_absolute_eq",
  "path": "result.event.end",
  "value": "2026-09-18T15:00:00-07:00",
  "tool": "create_calendar_event",
  "check_id": "morgan_032_02",
  "action_id": "morgan_032_create_calendar_event"
}

4. field_llm_judge / create_calendar_event

{
  "type": "field_llm_judge",
  "path": "result.event.body",
  "criterion": "States that production pushes after 3 PM Friday are blocked by default and require the full on-call roster online; requests for exceptions go through Morgan. Equivalent unambiguous first-person wording is acceptable.",
  "tool": "create_calendar_event",
  "check_id": "morgan_032_03",
  "action_id": "morgan_032_create_calendar_event"
}

Complete grading specification

{
  "type": "tool_trace",
  "config": {
    "check_version": 2,
    "today": "2026-09-14",
    "assertions": [
      {
        "type": "field_equals",
        "tool": "create_calendar_event",
        "path": "result.ok",
        "value": true,
        "check_id": "morgan_032_00",
        "action_id": "morgan_032_create_calendar_event"
      },
      {
        "type": "datetime_absolute_eq",
        "path": "result.event.start",
        "value": "2026-09-18T14:45:00-07:00",
        "tool": "create_calendar_event",
        "check_id": "morgan_032_01",
        "action_id": "morgan_032_create_calendar_event"
      },
      {
        "type": "datetime_absolute_eq",
        "path": "result.event.end",
        "value": "2026-09-18T15:00:00-07:00",
        "tool": "create_calendar_event",
        "check_id": "morgan_032_02",
        "action_id": "morgan_032_create_calendar_event"
      },
      {
        "type": "field_llm_judge",
        "path": "result.event.body",
        "criterion": "States that production pushes after 3 PM Friday are blocked by default and require the full on-call roster online; requests for exceptions go through Morgan. Equivalent unambiguous first-person wording is acceptable.",
        "tool": "create_calendar_event",
        "check_id": "morgan_032_03",
        "action_id": "morgan_032_create_calendar_event"
      }
    ]
  }
}

App state

Download JSON

Source file

tests/morgan/032.yaml

SHA-256: d062ee4a8e5b30cf3313d2d40a48d0a58a0099c52113e68231db5ef96bdbf008