Dianping Query Moss

查询大众点评餐厅信息。使用场景：(1) 用户要查询某家餐厅的评分、人均消费、地址 (2) 用户要搜索特定菜系或区域的餐厅推荐 (3) 用户要获取附近美食推荐。登录账号：一定S。重要：搜索时URL需包含城市ID（如hangzhou页面），账号定位北京时需使用URL参数方式绕过。

Security Scan

This is an instruction-only skill that describes how to use a browser tool to scrape Dianping search pages for restaurant information; its declared requirements and runtime instructions are generally consistent with that purpose.

Purpose & Capability

The name/description (query Dianping restaurant info) match the instructions: open the Hangzhou page, search, and extract restaurant cards. The skill does not request unrelated binaries, installs, or cloud credentials, so the capability aligns with the declared purpose.

Instruction Scope

The SKILL.md instructs the agent to use the browser tool (open https://www.dianping.com/hangzhou, inspect AI-search results, use browser.tabs, scroll to load content) and extract restaurant name, rating, average price, address, highlights, and review summaries — all within the stated scope. One notable instruction: it references a login account '一定S' and says if not logged in the user must provide a phone number and captcha to log in. That could cause the agent to prompt for potentially sensitive personal data; the instructions do not direct reading of unrelated system files or environment variables.

Install Mechanism

No install spec and no code files are present. This is instruction-only, so nothing is written to disk and there are no external packages or download URLs to evaluate.

Credentials

The skill declares no required environment variables or credentials, which is proportional. However, the SKILL.md includes an account name ('一定S') and instructs obtaining phone+captcha input from the user if not logged in. Because credential handling and collection of phone/captcha are not declared in metadata, users should be cautious about providing such information to the agent.

Persistence & Privilege

always is false and the skill is user-invocable with normal autonomous invocation allowed. It does not request persistent system privileges or modify other skills' settings.

Assessment

This skill is instruction-only and appears to simply scrape Dianping search pages for restaurant details. Before installing: (1) note the SKILL.md may prompt you to log in; it mentions an account name but provides no credentials — avoid sharing your phone number, verification codes, or passwords unless you trust the source; prefer using your own account in a browser session if necessary; (2) confirm you are comfortable with the agent opening and scraping public web pages via its browser tool; (3) absence of source/homepage means you have less provenance — if provenance matters, ask the publisher for details or use a verified alternative.

Dianping Query

Quick Start

1. 用 browser 工具打开大众点评杭州首页：https://www.dianping.com/hangzhou
   • 无需登录即可访问杭州地区内容
   • URL中的 /hangzhou 或 cityId=3 表示杭州
2. 确认登录状态（显示"一定S"即已登录）
3. 在搜索框输入餐厅名称或关键词
4. 重要：搜索结果在新标签页打开，URL格式为 https://www.dianping.com/ai-search?keyword=...
5. 使用 browser.tabs 查看所有打开的页面
6. 从AI搜索结果页面提取餐厅信息卡片

登录账号

• 用户名：一定S
• 如果未登录，需要用户配合输入手机号和验证码

城市定位问题处理

问题现象

• 账号显示定位在北京，无法切换到杭州
• 搜索结果会跳转到北京地区

解决方案（按优先级）

方案1：使用杭州URL（推荐）

• 直接访问 https://www.dianping.com/hangzhou 而非首页
• 搜索结果会自动关联杭州地区
• 无需账号切换，cookie定位不影响URL访问

方案2：搜索时指定城市

• 搜索词加上城市名，如："杭州滨江区好吃的餐厅"
• URL参数会包含城市信息：keyword/3/0_杭州滨江区...

方案3：浏览器地址栏修改

• 在地址栏手动修改城市ID
• 北京是 cityId=2，杭州是 cityId=3

城市ID参考

城市	cityId	URL示例
杭州	3	dianping.com/hangzhou
北京	2	dianping.com/
上海	1	dianping.com/shanghai

搜索结果提取

从AI搜索结果页面提取以下信息：

• 餐厅名称（link元素）
• 评分（⭐数字星）
• 人均消费（¥数字）
• 地址/位置
• 特色描述和推荐菜品
• 用户评价摘要

常见搜索场景

按名称查询：

搜索词：靠北泰式打抛饭
结果：店铺名称、评分4.2星、¥37、地址在拱墅区银树湾

按区域查询（使用杭州URL）：

URL: https://www.dianping.com/hangzhou
搜索词：滨江区好吃的餐厅
结果：7,799个相关商户，包含TOP推荐列表

按菜系查询：

搜索词：杭州泰式料理
结果：多家泰式餐厅列表，包含评分、人均、地址

注意事项

• 搜索结果页面可能需要滚动加载完整内容
• 店休、营业时间等特殊信息也会显示在结果中
• 点击餐厅名称可查看详情页
• 使用杭州URL可以绕过账号定位问题