360 Secure SkillHub
360 安全技能中心

Supermarket Offer Watcher

该 Skill 用本地 Python 脚本维护商品监控清单，以及家庭位置、半径、日/周检索频率与时间时区等配置；随后按产品执行网络检索、打开结果页，校验商品、门店、有效期与价格后去重，并发送简洁的优惠提醒。定时运行依赖宿主提供的定时任务能力。

lukasosterheider
lukasosterheider
热度 56.40
最近更新 2026-05-11 17:30:28
分类：生活依赖：代码执行 / 读写文件 / 联网浏览搜索应用场景：购物 / 推荐 / 日程安排 / 自动化脚本 ...

安装技能

复制下面的话术发给你的 Agent，它会通过 secure-skills 检索并安装该技能。

执行命令安装此 skill： npx secure-skills-cli add supermarket-offer-watcher
Metadata
name
buy-it-cheaper
description
Save a personal grocery product watchlist and check nearby supermarket deals automatically by radius. Use when the user wants to track products (e.g., coffee pads), configure home location, choose radius, and set daily or weekly deal alerts with notifications.

Supermarket Offer Watcher

Overview

Use this skill to maintain a local product watchlist and run recurring nearby-offer checks across supermarket chains.

Configurable after installation:

  • Home location (text)
  • Radius (km)
  • Product list
  • Check frequency (daily or weekly)
  • Check time and timezone

Default data file:

  • /data/workspace/data/supermarkt-watchlist.json

Quick setup

Run from the skill folder.

  1. Initialize config:
python3 scripts/watchlist.py init --location "<ZIP CITY>" --radius-km 15 --mode daily --time 07:00 --tz Europe/Berlin
  1. Add products:
python3 scripts/watchlist.py add "Senseo Coffee Pads" --aliases "Senseo Pads" --stores "REWE,EDEKA,Lidl,Aldi,Kaufland,Netto,Penny"
  1. Inspect config:
python3 scripts/watchlist.py list

Manage configuration later

Update location/radius:

python3 scripts/watchlist.py set-home --location "00001 Berlin" --radius-km 20

Update schedule:

# daily at 07:00
python3 scripts/watchlist.py set-schedule --mode daily --time 07:00 --tz Europe/Berlin

# weekly on monday at 07:00
python3 scripts/watchlist.py set-schedule --mode weekly --weekday monday --time 07:00 --tz Europe/Berlin

Remove product:

python3 scripts/watchlist.py remove "Senseo Coffee Pads"

Offer-check workflow

  1. Load watchlist config from /data/workspace/data/supermarkt-watchlist.json.
  2. For each product, run 2–4 web_search queries (product + "offer/deal" + location + optional chain).
  3. Open promising result pages using web_fetch.
  4. Validate each hit before reporting:
    • Product (or alias) is explicit
    • Store/chain is clear
    • Date window is currently valid (today/in current week)
    • Price or discount is visible
  5. Deduplicate same product/store/date combinations.
  6. Send compact alert summary.

Alert format

Use this line format per valid deal:

✅ <Product> — <Store> — <Price/Discount> — valid until <Date> — <URL>

If no safe deals are found, send:

No reliable nearby offers found today.

Cron setup guidelines

Use cron.add with sessionTarget: "isolated" and payload.kind: "agentTurn".

  • Daily cron expression: 0 7 * * *
  • Weekly cron expression example (Monday): 0 7 * * 1
  • Always set timezone from config (e.g., Europe/Berlin).

In the cron prompt, instruct the agent to:

  • read watchlist config,
  • run offer checks,
  • report only validated deals.