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

The payroll record has:
- payroll_month: 2026-10
- allowed_deduction: 8.69
- gross_pay: MISSING

Per R7, check order:
1. Missing gross_pay → clarify with reason_code MISSING_GROSS_PAY, citing R1 and R7.

But wait — I should verify the reference row exists for 2026-10? The order says: 1. Missing gross_pay first. So the answer is clarify with MISSING_GROSS_PAY. No monetary fields. Explicit null for missing required values.

The rules say: "Missing required values include explicit null." But the allowed answer fields for clarify: decision, reason_code, clauses. An optional explanation string is permitted. So I should include explanation naming the unresolved blocker and asking for its resolution.

The rule says "For these cases do not return calculated monetary fields. Missing required values include explicit null." Hmm, but the answer fields for clarify only include decision, reason_code, clauses (plus optional explanation). So I shouldn't add gross_pay field. The explanation can mention gross_pay is missing.

Also, should I check the reference row for 2026-10? The check order is: missing gross_pay is checked first (step 1), so no need to check reference rows. But it might be prudent to confirm the month has a reference row — not needed since step 1 triggers first. However, to be safe, the primary blocker is missing gross_pay. I'll cite R1 and R7.

Should I call lookup_reference? Not necessary. The order of checks: 1. Missing gross_pay → clarify. That's decisive.

Return JSON:
{"decision": "clarify", "reason_code": "MISSING_GROSS_PAY", "clauses": ["R1", "R7"], "explanation": "..."}

The explanation should name the unresolved blocker and ask for its resolution: "The payroll record for 2026-10 is missing gross_pay, which is required under R1; please supply gross_pay so the payslip can be calculated."

Money must be strings with two decimals — not applicable here. No Markdown fences, no surrounding prose? The instruction says "Return one JSON object without Markdown fences or surrounding prose." So output just the JSON object.

Let me start by listing the documents to see what inputs we have.

The payroll record has:
- payroll_month: 2026-10
- allowed_deduction: 8.69
- gross_pay: missing!

R7 check order:
1. Missing gross_pay → clarify with reason_code "MISSING_GROSS_PAY", citing R1 and R7.

So the answer is a clarification. No monetary fields. Missing required values include explicit null — the instructions say "Missing required values include explicit null." So for clarify, we include reason_code. Do we include gross_pay as null? The rule says "For these cases do not return calculated monetary fields. Missing required values include explicit null." The allowed answer fields for clarify: decision, reason_code, clauses. An optional explanation string is permitted. So I should include decision, reason_code, clauses, and explanation. Possibly include gross_pay: null? The allowed answer fields listed are "calculate uses decision, currency, gross_pay, employee_contribution, allowed_deduction, taxable_pay, income_tax, net_pay, clauses; retrieve uses decision, field, value, clauses; clarify uses decision, reason_code, clauses." So clarify fields are decision, reason_code, clauses — plus optional explanation. I'll include explanation naming the unresolved blocker and asking for resolution.

Since gross_pay is missing, I don't need to check the reference row — but the check order says missing gross_pay is first. No need to call lookup_reference since we're clarifying. Actually, should I verify that 2026-10 has a reference row? The check order puts missing gross_pay first, so it's the blocker. I could still check, but the answer is determined by the first check. The rules say check in this order, and return on the first failing check. Missing gross_pay fails first. I don't need lookup_reference.

Output JSON:
{
  "decision": "clarify",
  "reason_code": "MISSING_GROSS_PAY",
  "clauses": ["R1", "R7"],
  "explanation": "..."
}

Money must be strings with two decimals — but no monetary fields here. No Markdown fences, no surrounding prose. Return one JSON object.

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

The payroll record has payroll_month 2026-10 and allowed_deduction 8.69, but no gross_pay. Wait, let me check. The record shows payroll_month: "2026-10" and allowed_deduction: "8.69". There's no gross_pay field.

Per R7: Missing gross_pay → decision "clarify", reason_code "MISSING_GROSS_PAY", citing R1 and R7.

But wait — the order of checks: 1. Missing gross_pay → clarify MISSING_GROSS_PAY. 2. Missing payroll_month. 3. Month without reference row. 4. Conflicting gross pay.

Here gross_pay is missing. So the answer is clarify with MISSING_GROSS_PAY, citing R1 and R7.

Hmm, but should I check the reference row first? The order says check in this order: 1. Missing gross_pay first. So MISSING_GROSS_PAY.

The month is 2026-10, which is covered by 2026 rows presumably. But since gross_pay is missing, we return clarify per check 1.

Actually, let me double-check: the record has payroll_month present, so check 2 passes. Check 1 fails first: missing gross_pay → clarify MISSING_GROSS_PAY.

The answer fields for clarify: decision, reason_code, clauses. Optional explanation string. Missing required values include explicit null — but "Do not include input-only fields such as payroll_month, contract_gross_pay, or amendment." So I should not include gross_pay field. Allowed answer fields for clarify: decision, reason_code, clauses. So just those plus optional explanation.

Should I still verify the reference row exists for 2026-10? The check order says missing gross_pay is checked first, so no need. But to be safe, the month is 2026-10 which is within 2026. The clarification is due to missing gross_pay regardless.

Let me not call lookup_reference since the decision is clarify at check 1. Actually, it might be prudent to confirm the month has a reference row, but the rule order mandates returning MISSING_GROSS_PAY when gross_pay is missing, before checking reference rows. The month is in 2026 so it's covered anyway.

Output JSON:
{
  "decision": "clarify",
  "reason_code": "MISSING_GROSS_PAY",
  "clauses": ["R1", "R7"],
  "explanation": "The payroll record for 2026-10 has no gross_pay; R1 requires gross_pay for calculation. Please supply the missing gross_pay."
}

No monetary fields for clarify cases. Return one JSON object without Markdown fences.