DolphinBench
DiscordGitHub

Test 015

Sep 14, 2026 / 1 fact

YAML

Request

Add the appropriate CRM contacts from dinner. Lina Torres at BrightBank procurement requested a Q4 integration check-in; Casey Wu from the Stanford founder table offered a friend-of-friend introduction.

Required memory

Fact 80

Morgan Chen requested that the CRM rule be locked in: only investor, customer, and vendor contacts belong in the CRM; Stanford-crowd and friends-of-friends introductions must not be entered; investors must be tagged `investor`, not `network`.

Source evidence (1)

000451Apr 26, 2023 / 14:30 UTC-07:00

Don't put Stanford crowd / friends-of-friends intros in the CRM. Those stay personal, full stop. CRM is for investor, customer, vendor contacts only -- and if it's an investor, tag `investor`, not `network`. Near miss just now so lock that rule in.

Message 000451 in history

Expected tool calls

  • add_crm_row

Grading

1. field_equals / add_crm_row
{
  "type": "field_equals",
  "tool": "add_crm_row",
  "path": "result.ok",
  "value": true,
  "check_id": "morgan_015_00",
  "action_id": "morgan_015_add_crm_row"
}
2. field_equals / add_crm_row
{
  "type": "field_equals",
  "path": "result.row.name",
  "value": "Lina Torres",
  "tool": "add_crm_row",
  "check_id": "morgan_015_01",
  "action_id": "morgan_015_add_crm_row"
}
3. tool_success_count / add_crm_row
{
  "type": "tool_success_count",
  "tool": "add_crm_row",
  "count": 1,
  "check_id": "morgan_015_02"
}
Complete grading specification
{
  "type": "tool_trace",
  "config": {
    "check_version": 2,
    "today": "2026-09-14",
    "assertions": [
      {
        "type": "field_equals",
        "tool": "add_crm_row",
        "path": "result.ok",
        "value": true,
        "check_id": "morgan_015_00",
        "action_id": "morgan_015_add_crm_row"
      },
      {
        "type": "field_equals",
        "path": "result.row.name",
        "value": "Lina Torres",
        "tool": "add_crm_row",
        "check_id": "morgan_015_01",
        "action_id": "morgan_015_add_crm_row"
      },
      {
        "type": "tool_success_count",
        "tool": "add_crm_row",
        "count": 1,
        "check_id": "morgan_015_02"
      }
    ]
  }
}
App stateDownload JSON
Source file

tests/morgan/015.yaml

SHA-256: 1d3fd64b8aa22706437061efa5308deacbb3262577b08bdbedc365e93ddb8937