Ant Social Security Nanjing Calculator

Problem

Nanjing employer social security calculations look simple for a single base, but real payroll checks often get stuck on details: whether the monthly base hits the annual floor or ceiling, whether the pension, medical, maternity, unemployment, and work-injury rates are correct, whether the work-injury rate follows the unit’s assessed industry rate, and whether the serious-illness assistance fee should be added. A single number is not enough to compare employer cost with employee withholding, and manual spreadsheets can separate totals, line items, and policy notes. This skill turns that calculation into a reusable conversational workflow.

How it works

The skill extracts the social security base from the user’s message, then confirms optional parameters: the year defaults to 2026, and the current 2025 and 2026 rates are the same; the work-injury rate defaults to 0.4% for a category-two industry baseline, but can be set to 0.2% or 0.7% if the unit’s assessed rate differs; the serious-illness assistance fee is excluded by default, or can be added as CNY 10/month. The script then clamps the base to the Nanjing 2025-2026 annual range, from CNY 4,952 to CNY 24,762, and calculates employer and employee contributions for pension, medical, maternity, unemployment, and work-injury insurance. It generates a PNG infographic with the base, employer total, employee total, grand total, line items, policy notes, and the Ant Social Security logo.

Scope and caveats

The skill is aimed at Nanjing enterprise employee social security calculation, which fits payroll checks, cost estimates, and itemized image output. It is not a general nationwide social security calculator and does not cover housing fund, individual income tax, medical pooled accounts, cross-city enrollment, or special population policies. If the unit uses an assessed work-injury rate, local subsidies, historical annual rules, or includes the serious-illness assistance fee, specify the parameters before generating the result. Runtime requires Python 3 and Pillow; output is an image rather than a structured API response.