Business Ops Ordering Assistant

Problem It Solves

Dasheng targets cake, snack, and flower ordering workflows in business operations. It packages address selection, product search, spec confirmation, order-link generation, and order status checks into a step-by-step skill instead of manual web or API calls.

How It Works

• Token and address: The user provides a user_token, stored in dgss_token.md, and the skill reads default_city and default_aid to reuse a saved address. City names must normalize to the city level.
• Search and order: It calls sweets_lst with cat_id (1=cake, 5=snacks, 8=flowers) and keyword, then returns brand, product, spec, and price details before getOrderaddr generates an order link.
• Order status: It extracts scene from the order URL and calls getOrderStatus for status.
• Protocol details: The API uses JSON-RPC 2.0; the outer payload returns result.content[0].text, containing an inner {code, data, msg} object. Windows should use a Python script to avoid curl Chinese-argument encoding issues.

Boundaries

• Only one spec_id per order; multi-spec stacking is not supported.
• It does not handle refunds, cancellations, logistics, invoices, or membership points.
• It must not expose endpoint names, domains, internal fields, or raw errors to the end user.