Read bank.csv and produce a cash position summary for today:
- opening balance, total inflows, total outflows, closing balance
- a breakdown of inflows/outflows by category (sales, suppliers, payroll, tax, other)
- flag any single transaction above $10,000 or any unusual category
Save the result as cash-summary-<date>.md

Compare actuals.csv with budget.csv for this month. Produce:
- a table: category | actual | budget | variance $ | variance %
- list the top 5 variances (largest absolute) with a likely reason for each
- flag any category more than 15% over budget
Save as variance-analysis-<period>.md

Read ar-aging.csv. Summarise receivables by bucket (current, 1-30, 31-60, 61-90,
90+ days) and by customer. List the top 10 debtors by total outstanding, with the
oldest invoice date. Draft a short, professional follow-up email for each of the
top 5 overdue debtors (90+ days) — polite but firm, mentioning invoice numbers and
amounts. Save the emails to ar-followups.md

Read payroll-<month>.csv. Produce:
- total gross pay, total CPF contributions, total net pay, headcount
- by department: headcount, gross pay total, average pay
- flag any employee whose pay changed by more than 20% vs the previous month
- list employees with no CPF contribution (flag for review)
Save as payroll-summary-<month>.md

Compare ledger.csv with bank-statement.csv and list all differences:
- amounts in ledger but not in the bank statement (and vice versa)
- date mismatches for the same amount
- duplicates (same amount twice in ledger)
Group them into: timing differences vs. potential errors. Show as a table.

Extract from each invoice file: supplier, invoice number, date, amount (excl. GST),
GST amount (9%), total, and PO number if present. Output one table per invoice,
then one combined summary table. Flag any invoice where GST is not 9% of the
net amount, or where the total doesn't add up.

Using sales.csv and purchases.csv, compute for the period:
- output tax (9% of taxable sales), input tax (9% of taxable purchases)
- GST payable/receivable = output − input
- list all line items where the GST amount is not exactly 9% of the net amount
- flag any zero-rated/exempt items that appear in the taxable lists
Save as gst-check-<period>.md

Compare the inter-company ledgers of entities A, B, C and D. For each pair,
match "due to" against the counterparty's "due from" and report:
- matched amounts
- unmatched amounts with the difference
- any amount that appears on only one side (potential error or unrecorded entry)
Output a reconciliation table per pair, and a summary of total unmatched amounts.

Based on the attached P&L, balance sheet and prior-year data, draft a management
commentary (max 1.5 pages) with:
- key highlights and lowlights vs last month and vs last year
- margin analysis (gross margin, operating margin) with drivers
- balance sheet commentary: cash, receivables, payables, borrowings
- 2-3 questions management should ask, and recommended focus for next month
Use plain language; cite the figure you are commenting on in brackets.
Save as management-commentary-<month>.md

Read fa-register.csv. For each asset, verify the monthly depreciation is consistent
with its cost and rate. Flag:
- assets with zero net book value still listed as active
- assets fully depreciated but still in use (list — these may need review)
- any asset missing a cost, date or rate
Output a clean review table with only the flagged assets.

Based on this audit checklist, build an audit schedule markdown table:
Area | Working paper name | Supporting docs needed | Owner | Status | Notes.
Pre-fill status "Not started" for every row and save as audit-schedule.md

Using fx-balances.csv and the attached month-end rates, revalue all foreign currency
balances to SGD: balance × rate per account. Show original, rate, revalued amount,
and gain/loss vs the previous month's revaluation. Flag any revaluation gain/loss
above $5,000 for review.