food-travel — Eat-First Travel Planner
One-liner: Input a dish, a craving, or a city — get a complete travel plan built around eating.
This skill solves the full "eat → where → go → stay → route" chain for food lovers.
Scenario Detection
Identify which scenario the user wants, then follow the corresponding workflow:
| Trigger pattern | Scenario | Example |
|---|
| A dish/cuisine + no city | A: Pick a destination | "我想吃烤鸭" "想吃海鲜去哪" |
| A city + food intent | B: City food map | "成都有什么好吃的" "杭州美食攻略" |
| A city + duration + food intent | C: Full itinerary | "3天吃遍西安" "周末广州美食游" |
If unclear, ask the user to clarify.
Scenario A: Pick a Destination for Food
Input: a dish, cuisine, or flavor preference
Output: best city recommendation + food list + travel logistics
Steps
- Web search:
"{dish/cuisine} 最正宗 去哪个城市吃" to identify the top 2-3 cities.
- For each city, web search:
"{city} 必吃 {dish} 餐厅推荐" to get restaurant data.
- Search flights (if user provides origin):
flyai search-flight --origin "{origin}" --destination "{city}" --dep-date {date}
- Search hotels:
flyai search-hotel --dest-name "{city}" --check-in-date {date} --check-out-date {date}
Output format
# 为了{dish}，去{city}！
## 为什么选{city}
（One-paragraph reason）
## 必吃清单
| 餐厅 | 招牌菜 | 人均 | 地址 | 推荐理由 |
|------|--------|------|------|----------|
| ... | ... | ... | ... | ... |
## 怎么去
（Flight options table with booking links）
## 住哪里
（Hotel options near food districts, with booking links）
Scenario B: City Food Map
Input: a city name
Output: meal-by-meal restaurant map organized by time of day
Steps
- Web search:
"{city} 必吃餐厅推荐" + "{city} 特色小吃 推荐" + "{city} 夜宵 推荐".
- Organize results into 4 time slots: 早餐, 午餐, 晚餐, 夜宵/下午茶.
- keyword-search supplement:
flyai keyword-search --query "{city} 美食券 餐厅"
Filter for food-related items only.
Output format
# {city}美食地图
## 🌅 早餐
| 餐厅 | 推荐 | 人均 | 地址 |
|------|------|------|------|
## ☀️ 午餐
...
## 🌆 晚餐
...
## 🌙 夜宵
...
## 可预订美食产品
（Filtered keyword-search results with images and booking links）
> 餐厅数据来自网络搜索，美食券来自 fly.ai 实时结果。
Scenario C: Full Food-Driven Itinerary
Input: city + duration (e.g. "3天吃遍西安")
Output: day-by-day schedule with every meal planned + attractions between meals + transport + hotel
Steps
- Web search:
"{city} {N}天美食攻略" + "{city} 必吃餐厅推荐".
- Search hotels:
flyai search-hotel --dest-name "{city}" --check-in-date {date} --check-out-date {date}
- Search flights (if origin provided):
flyai search-flight --origin "{origin}" --destination "{city}" --dep-date {date}
- Search attractions to fill between-meal time:
flyai search-poi --city-name "{city}"
- Organize into a day-by-day plan where every meal is the anchor.
Output format
# {N}天吃遍{city}
## Day 1
### 🌅 早餐 — {restaurant}
- 推荐：{dishes}｜人均：{price}｜地址：{addr}
### ☀️ 上午 — {attraction}（吃完溜达消食）
（POI info with booking link）
### 🍜 午餐 — {restaurant}
- 推荐：{dishes}｜人均：{price}｜地址：{addr}
### 🌆 下午 — {attraction/activity}
### 🔥 晚餐 — {restaurant}
- 推荐：{dishes}｜人均：{price}｜地址：{addr}
### 🌙 夜宵 — {restaurant}
- 推荐：{dishes}
## Day 2
...
## 交通
（Flight options with booking links）
## 住宿
（Hotel options with booking links, prefer hotels near Day 1 dinner area）
## 预算估算
| 项目 | 预估费用 |
|------|----------|
| 机票 | ¥xxx |
| 住宿 | ¥xxx |
| 餐饮 | ¥xxx |
| 门票 | ¥xxx |
| **合计** | **¥xxx** |
> 餐厅数据来自网络搜索，机票酒店来自 fly.ai 实时结果。
General Rules
- Food comes first — every itinerary section starts with a meal, attractions fill the gaps.
- Web search for restaurants — flyai has no restaurant database; always use web search for dining data.
- flyai for logistics — use
search-flight, search-hotel, search-poi, keyword-search for transport, accommodation, attractions, and bookable dining products.
- Always include booking links — for every flight, hotel, and POI result, show
[Click to book]({url}).
- Always include images — show
 or  when available.
- Practical details — include price, address, opening hours when available.
- Source attribution — "餐厅数据来自网络搜索，机票酒店来自 fly.ai 实时结果。"
