Dianping Merchant Decision Assistant

Problem: a high score is not enough

Choosing on Dianping is less about finding stores and more about filtering noisy signals. A 4.6 rating may come from a small sample, a cheap group-buy may be restricted to weekdays, and repeated complaints about queues, upselling, or inconsistent food may not be visible in the star score alone. Practical judgment requires separating rating, review quality, transaction constraints, and use-case fit.

How it works: scenario first, evidence second

The skill turns the decision into explicit steps:

• Scenario matching: distinguish dating, work, family meals, first-time try, or repeat visits before judging the store.
• Evidence reading: weigh rating bands, review count, recent complaints, and page constraints captured via the shared browser-commerce-base workflow when live pages are available.
• Deal evaluation: check date limits, booking rules, scope, and hidden conditions instead of price alone.
• Structured output: for one store, produce conclusion, rationale, and risk; for multiple stores, rank them and list each store's fit and caveats.

Boundaries: no blind recommendation

This skill is suited to converting reviews, packages, and merchant pages into comparable decision inputs. It should not treat platform ratings as ground truth. For stores that depend heavily on live queue, temporary promos, or in-person experience, actual booking and visit conditions still matter.