national-pension-workplace

This skill queries the National Pension Service's enrolled-workplace dataset (data.go.kr 3046071, V2) through the `k-skill-proxy` to gauge a Korean company's headcount and its month-over-month change. Searching by workplace name (with the first six digits of a business number as an optional prefix filter), it returns matching workplace candidates and, when a single workplace is pinned down, details such as the number of enrolled subscribers (`jnngpCnt`), the current-month notified pension amount (`crrmmNtcAmt`), and monthly new-acquisition/loss counts, plus a month-by-month time series. It solves the problem of estimating a firm's employee scale and staffing trend from an authoritative public source.

A privacy boundary shapes the design: the National Pension data exposes only the first six digits of a business number (the rest masked), so a workplace name is mandatory and the skill lists six-digit-plus-name candidates without asserting they are the same entity. Coverage skews toward corporations and workplaces above a certain employee threshold; small or sole-proprietor workplaces may be absent. The skill reports only upstream facts, avoiding invented scores or risk labels, and does not assert identity when several candidates match.

The published entry is a CLI stub deferring to `npx -y @nomadamas/k-skill@0 instruct national-pension-workplace` for live instructions and running `scripts/national_pension_workplace.py` with a required `--name`. It targets recruiters, investors, and business-research users sizing up companies. No user-side secret is required; the data.go.kr key lives only on the proxy, and hard rules forbid storing plaintext credentials or bypassing legal boundaries.