Tax Compliance Self-Check

Problem

In month-end close, pre-filing checks, audit workpapers, and group compliance reviews, teams often lack a repeatable tax-risk path. Facts are scattered across vouchers, filings, contracts, and expense details, so manual review can miss high-frequency issues such as invoice risk, late filing or payment, missing withholding, and corporate income tax adjustments. This skill turns that check into structured inputs plus rule-based scanning.

How it works

The skill expects company details in company and tax facts grouped under vat, cit, iit, stamp, invoice, and filing. scripts/rules.json is a data-driven rule library supporting condition types such as flag_true, missing, diff_pct, meal_limit, deduct_limit, all, and any; adding or adjusting rules usually means editing JSON only. Running scripts/run_check.py produces a Markdown summary and an Excel report with compliance score, risk list, regulatory basis, remediation advice, and items pending confirmation. Scoring is severity-weighted: high = 3, medium = 2, low = 1, while pending items excluded from the denominator.

Boundaries

It is suitable for pre-filing pre-checks, cross-entity comparison, and auxiliary workpapers before tax advisors enter, but it is not a tax opinion. Material tax matters, incentive eligibility, and local policies must still follow the latest regulations and the competent tax authority. “Pending confirmation” is not a pass and should be rechecked after data is added; the baseline covers high-frequency risk points, not the full tax code or industry-specific rules.