← 返回 Skills 市场
108
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install chill-beijing
功能描述
北京下班及周末放松推荐助手 v1.4.0。工作日推荐电影、脱口秀、演出；周末推荐City Walk、京郊游、社交活动、演出。实时抓取猫眼/大麦/美团/小红书数据。
使用说明 (SKILL.md)
Chill Beijing - 北京放松指南
北京打工人下班及周末去哪儿？实时抓取猫眼、大麦、美团、小红书最新数据，智能推荐。
Features
- 实时数据抓取 - 猫眼、大麦、美团、小红书最新数据
- 智能分类 - 工作日/周末不同推荐策略
- 耐心等待 - 设置足够超时，确保数据完整抓取
- 最新日期 - 始终使用当日数据，拒绝历史/模拟数据
Quick Start
# 获取今日推荐（自动判断工作日/周末）
node ~/.openclaw/skills/chill-beijing/scripts/generate-recommendations.mjs
# 强制获取工作日推荐
node ~/.openclaw/skills/chill-beijing/scripts/generate-recommendations.mjs --workday
# 强制获取周末推荐
node ~/.openclaw/skills/chill-beijing/scripts/generate-recommendations.mjs --weekend
# JSON 格式输出
node ~/.openclaw/skills/chill-beijing/scripts/generate-recommendations.mjs --json
推荐分类
工作日（周一-周四）
| 分类 | 数据源 | 内容 |
|---|---|---|
| 🎬 电影 | 猫眼/美团 | 当日热映、评分、场次 |
| 🎤 脱口秀 | 大麦/美团 | 开放麦、专场演出 |
| 🎵 演出 | 大麦 | LiveHouse、音乐会、民谣酒馆 |
周五（工作日+周末预告）
输出工作日3类 + 周末4类完整推荐
周末（周六-周日）
| 分类 | 数据源 | 内容 |
|---|---|---|
| 🚶 City Walk | 小红书 | 北京 city walk 路线推荐 |
| 🏔️ 京郊/周边游 | 小红书/美团 | 可过夜景点、民宿 |
| 👥 社交活动局 | 小红书 | 桌游、剧本杀、社交聚会 |
| 🎭 演出类 | 大麦/猫眼 | 话剧、演唱会、展览 |
Data Sources
| 数据源 | 抓取内容 | 超时设置 |
|---|---|---|
| 猫眼电影 | 当日上映电影、评分、场次 | 60秒 |
| 大麦网 | 北京站演出列表 | 60秒 |
| 美团演出 | 脱口秀、LiveHouse | 60秒 |
| 小红书 | City Walk、京郊游、社交局 | 60秒 |
技术特性
- 防反爬：使用 Puppeteer 模拟真实浏览器
- 超时耐心：单源 60 秒超时，失败自动重试
- 当日数据：所有抓取带日期参数，确保最新
- 容错降级：主源失败时使用备用源
API Usage
import { generateRecommendations, getTodayRecommendations } from './scripts/generate-recommendations.mjs';
// 获取今日推荐（自动判断工作日/周末）
const recs = await generateRecommendations();
// 获取结构化数据
const data = await getTodayRecommendations({ format: 'json' });
Files
scripts/data-fetcher.mjs- 数据抓取模块（猫眼/大麦/美团/小红书）scripts/generate-recommendations.mjs- 推荐生成主程序scripts/formatters.mjs- 数据格式化工具
Changelog
v1.0.0 (2026-03-27)
- 🎉 初始版本
- ✨ 工作日推荐：电影、脱口秀、演出
- ✨ 周末推荐：City Walk、京郊游、社交局、演出
- ✨ 多源数据抓取（猫眼/大麦/美团/小红书）
- ✨ 智能工作日/周末判断
License
MIT
安全使用建议
This package appears coherent with its stated purpose, but before installing you should: (1) verify the source/origin because the registry metadata lacks a homepage; (2) audit and pin dependencies (puppeteer ^22.0.0) and run npm install in an isolated environment — puppeteer will download a Chromium binary and headless browsers open network access and can be large; (3) note the SKILL.md mentions scripts/formatters.mjs which is missing — ask the author or check the full source to ensure nothing was omitted; (4) run the skill in a sandbox/container (or non-root) because the code launches Chromium with --no-sandbox (common but reduces sandbox protections); (5) if you need to run on a shared system, consider network restrictions or a proxy to limit where the headless browser can connect. If any of these points are unacceptable or the author/source can't be verified, treat the package cautiously.
功能分析
Type: OpenClaw Skill
Name: chill-beijing
Version: 1.4.0
The skill is a recommendation engine for leisure activities in Beijing, providing detailed suggestions for movies, shows, and city walks. It uses Puppeteer in `scripts/data-fetcher.mjs` to scrape real-time data from legitimate platforms like Maoyan and Damai, and fetches weather data from wttr.in. The code logic is transparent, lacks high-risk behaviors like shell execution or credential theft, and the extensive hardcoded content in `scripts/generate-recommendations.mjs` consists of legitimate points of interest and cultural descriptions aligned with the skill's stated purpose.
能力评估
Purpose & Capability
The name/description (Beijing leisure recommendations) align with included code: data-fetcher scrapes 猫眼/大麦/美团/小红书 and wttr.in for weather, and generate-recommendations formats suggestions. Declared dependency on puppeteer is expected for the described browser scraping.
Instruction Scope
SKILL.md instructs running the generate-recommendations script and documents use of Puppeteer and 60s timeouts — that matches the code. The instructions do not ask for unrelated files/credentials or to exfiltrate data. Minor mismatch: SKILL.md references a scripts/formatters.mjs file which is not present in the file manifest; also many recommendation texts in generate-recommendations.mjs are static content rather than purely derived from live scrapes (not a security issue but an implementation note).
Install Mechanism
There is no install spec in the registry (instruction-only), but package.json lists puppeteer, which will cause npm to fetch large Chromium artifacts when dependencies are installed. No suspicious third-party download URLs are present in the repo files, but running npm install will trigger remote downloads (normal for puppeteer).
Credentials
The skill requests no environment variables, no credentials, and accesses only public websites (wttr.in and the listed Chinese event sites). There is no evidence of secret exfiltration or access to unrelated services or config paths.
Persistence & Privilege
Flags show always:false and default autonomous invocation allowed (normal). The package does not request permanent presence, does not modify other skills, and does not require special agent-level privileges.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install chill-beijing - 安装完成后，直接呼叫该 Skill 的名称或使用
/chill-beijing触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.4.0
v1.4.0: 重写推荐理由，每段都深入分析为什么值得去；实时抓取猫眼/大麦/美团/小红书数据；智能工作日/周末推荐
元数据
常见问题
Chill Beijing 是什么？
北京下班及周末放松推荐助手 v1.4.0。工作日推荐电影、脱口秀、演出；周末推荐City Walk、京郊游、社交活动、演出。实时抓取猫眼/大麦/美团/小红书数据。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 108 次。
如何安装 Chill Beijing？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install chill-beijing」即可一键安装，无需额外配置。
Chill Beijing 是免费的吗？
是的，Chill Beijing 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Chill Beijing 支持哪些平台？
Chill Beijing 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Chill Beijing？
由 caoyachao（@caoyachao）开发并维护，当前版本 v1.4.0。
推荐 Skills