2. field_llm_judge / create_doc
{
"type": "field_llm_judge",
"tool": "create_doc",
"action_id": "create_roasted_chicken_evidence_sheet",
"path": "args",
"criterion": "The body is a compact decision-evidence sheet about the three roasted-chicken sale runs, rather than unrelated notes or a general menu summary.\n\nThe body clearly states that across all three runs, production was repeatable and stayed within the chef-owner's labor and yield thresholds. Missing or contradicting either repeatability or compliance with both types of threshold must fail.\n\nThe body clearly states that customer demand or sell-through varied across the runs. Any claim that demand or sell-through was uniform must fail.\n\nThe body does not make, recommend, or purport to reopen a decision about permanent, standing, weekly, private-event, or other future menu placement; it remains an evidence record of the sale runs.",
"check_id": "riley_108_01"
}Complete grading specification
{
"type": "tool_trace",
"config": {
"check_version": 2,
"today": "2028-01-01",
"semantic_judge_version": 2,
"assertions": [
{
"type": "field_equals",
"tool": "create_doc",
"action_id": "create_roasted_chicken_evidence_sheet",
"path": "result.ok",
"value": true,
"check_id": "riley_108_00"
},
{
"type": "field_llm_judge",
"tool": "create_doc",
"action_id": "create_roasted_chicken_evidence_sheet",
"path": "args",
"criterion": "The body is a compact decision-evidence sheet about the three roasted-chicken sale runs, rather than unrelated notes or a general menu summary.\n\nThe body clearly states that across all three runs, production was repeatable and stayed within the chef-owner's labor and yield thresholds. Missing or contradicting either repeatability or compliance with both types of threshold must fail.\n\nThe body clearly states that customer demand or sell-through varied across the runs. Any claim that demand or sell-through was uniform must fail.\n\nThe body does not make, recommend, or purport to reopen a decision about permanent, standing, weekly, private-event, or other future menu placement; it remains an evidence record of the sale runs.",
"check_id": "riley_108_01"
}
]
}
}