Let me start by listing the documents to get the case inputs.

The payroll record has gross_pay "5796.92" and allowed_deduction "0.00", but no payroll_month. Per R7, missing payroll_month → clarify with reason_code "MISSING_PAYROLL_MONTH", citing R1 and R7.

Wait — let me double check. The record only has gross_pay and allowed_deduction. No payroll_month. So R7 check order:
1. Missing gross_pay? No, it's present.
2. Missing payroll_month? Yes → clarify with MISSING_PAYROLL_MONTH.

The answer format: clarify uses decision, reason_code, clauses. Optional explanation string permitted. Money fields not included. Missing required values include explicit null — but that applies to "these cases" for calculated monetary fields? It says "For these cases do not return calculated monetary fields. Missing required values include explicit null." Hmm, that's about not guessing values. For clarify, the allowed fields are decision, reason_code, clauses (plus optional explanation). I should not add gross_pay etc.

Return JSON without Markdown fences.

Let me start by listing the documents to retrieve the case inputs.

The payroll record has gross_pay = "5796.92" and allowed_deduction = "0.00". But there's no payroll_month! Wait, let me check again. The record contains only gross_pay and allowed_deduction. No payroll_month field.

Hmm, but wait — maybe I should check if there's more information. The document only has those fields. There's no payroll_month. According to R7:

1. Missing gross_pay: clarify MISSING_GROSS_PAY
2. Missing payroll_month: clarify MISSING_PAYROLL_MONTH
3. Month without reference row: clarify NO_APPLICABLE_REFERENCE
4. Conflicting gross pay: clarify CONFLICTING_GROSS_PAY

The record has gross_pay but is missing payroll_month. So the answer should be decision: "clarify", reason_code: "MISSING_PAYROLL_MONTH", citing R1 and R7.

Wait, but let me double-check the document. The record is {"gross_pay": "5796.92", "allowed_deduction": "0.00"}. No payroll_month, no amendment, no contract_gross_pay, no version. Hmm, actually the rules say "Tasks identify this rule set as `3`." — the prompt doesn't mention a version, but the harness handles that. The record itself lacks payroll_month.

Per R7 order: check missing gross_pay first (it's present), then missing payroll_month (it's missing). So clarify with MISSING_PAYROLL_MONTH, citing R1 and R7.

For clarification: fields are decision, reason_code, clauses, plus optional explanation. "Missing required values include explicit null" — hmm, that's about not returning calculated monetary fields; "Missing required values include explicit null" — I think that means when a required value is missing, it should be represented as explicit null in the input description. But the answer fields for clarify are decision, reason_code, clauses. The optional explanation string is permitted.

So my answer:

{
  "decision": "clarify",
  "reason_code": "MISSING_PAYROLL_MONTH",
  "clauses": ["R1", "R7"],
  "explanation": "The payroll record supplies gross_pay (5796.92 LCR) and allowed_deduction (0.00) but no payroll_month (YYYY-MM). R1 requires payroll_month to select the reference row; without it no reference row can be selected and no payslip can be calculated. Please supply the payroll month."
}

I should not call lookup_reference or calculate since we can't determine the month. No need for further tool calls. The answer is a clarification.

The instructions say to return one JSON object without Markdown fences or surrounding prose. So my final response should be just the JSON object.

Let me start by listing the documents to get the case inputs.

The payroll record has gross_pay = 5796.92 and allowed_deduction = 0.00, but there's no payroll_month. Wait, let me re-read the document. It has "gross_pay": "5796.92", "allowed_deduction": "0.00" but no payroll_month field.

Per R7:
1. Missing gross_pay → MISSING_GROSS_PAY
2. Missing payroll_month → MISSING_PAYROLL_MONTH

The payroll_month is missing. So the decision is clarify with reason_code "MISSING_PAYROLL_MONTH", citing R1 and R7.

Wait — let me double-check the document. The record is: {"kind": "payroll", "record": {"gross_pay": "5796.92", "allowed_deduction": "0.00"}}. Indeed, no payroll_month. So R7 step 2 applies: return decision clarify, reason_code MISSING_PAYROLL_MONTH, citing R1 and R7.

No monetary fields should be returned. Missing required values include explicit null — but for clarification, the answer fields are decision, reason_code, clauses, plus optional explanation. So I should return:

{
  "decision": "clarify",
  "reason_code": "MISSING_PAYROLL_MONTH",
  "clauses": ["R1", "R7"],
  "explanation": "..."
}

The explanation should name the unresolved blocker and ask for its resolution.