daangn-used-goods-search

Daangn Used-Goods Search is a read-only skill for querying Daangn (Korea's Karrot) second-hand marketplace using its public Remix `_data` JSON routes. It solves the problem of searching used-goods listings by keyword and region and viewing listing details through natural language, without needing to log in. It deliberately excludes any account-affecting or transactional automation such as login, chat, favoriting, offers, inquiries, reservations, or purchases, and it does not attempt to bypass CAPTCHAs, bot blocks, or login walls.

The skill drives a single Python helper (scripts/daangn_used_goods.py) that uses only the Python standard library — no external packages or proxy. Its workflow extracts keyword, region, and price/type filters from a request; resolves a region name to Daangn's internal region id via the public region-keyword API (with a defined disambiguation priority for duplicate place names); calls the category-specific search `_data` route; and optionally reads a detail route or public HTML metadata for a given listing URL. Output fields include title, price, status, region, and source URL, always preserving the source URL and the effective region applied. It documents clear failure modes (route/schema changes, ambiguous regions, zero results, deleted or private listings) and reports rather than works around blocking.

It is aimed at Korean-speaking users and agents who want quick, read-only marketplace lookups — for example finding a MacBook in a specific neighborhood or inspecting a shared listing URL. Prerequisites are an internet connection and Python 3.9+. It is explicitly scoped to observation only and does not depend on IP/cookie default location.