wellness-su - SKILL.md Agent Skill

name: wellness-su description: This skill should be used when the user asks for "diet suggestions", "outfit recommendations", "wellness advice", "weather-based tips", or mentions nutrition, clothing, or health recommendations based on weather conditions. Provides workflows for delivering weather-based diet and outfit advice. version: 1.0.0

Wellness Su - Daily Diet and Outfit Advisor

This skill provides workflows for the wellness-su agent that delivers daily diet and outfit recommendations based on weather conditions.

When This Skill Applies

Activate this skill when:

User asks for diet suggestions or nutrition advice
User requests outfit or clothing recommendations
User mentions wellness, health, or lifestyle tips
User wants advice based on current or forecasted weather

Core Workflow

1. Fetch Weather Information

Get current day's weather data for the user's location.

Weather Data Required:

Temperature (current and forecast)
Humidity
Weather condition (sunny, cloudy, rainy, snowy, etc.)
Wind speed
UV index (if available)

Steps:

Determine user's location (default to configured location)
Use Weather MCP or API to fetch current weather
Get daily forecast if available

2. Generate Diet Recommendations

Provide diet suggestions based on weather conditions:

Cold Weather (< 10°C):

Recommend warm, hearty meals
Suggest hot soups, stews, or porridge
Include warming spices (ginger, cinnamon)
Recommend hot beverages (tea, warm water)

Moderate Weather (10°C - 25°C):

Balanced meals with variety
Fresh fruits and vegetables
Light soups or salads
Room temperature foods

Hot Weather (> 25°C):

Light, refreshing meals
Cold dishes and salads
Hydrating foods (watermelon, cucumber)
Avoid heavy, greasy foods

Rainy/Humid Weather:

Foods that boost immunity (vitamin C rich)
Warm soups
Avoid overly oily foods
Include ginger or garlic

3. Generate Outfit Recommendations

Provide clothing suggestions based on weather:

Temperature-based:

< 5°C: Down jacket, thermal underwear, scarf, gloves, warm hat
5-15°C: Heavy coat, sweater, long pants, light layers
15-20°C: Light jacket, sweater, jeans or trousers
20-25°C: T-shirt, light cardigan, comfortable pants
> 25°C: Light, breathable clothing, shorts, sun protection

Weather Condition-based:

Sunny: Sunscreen, sunglasses, hat
Cloudy: Light layers, comfortable shoes
Rainy: Raincoat or umbrella, waterproof shoes, extra socks
Snowy: Waterproof boots, thermal wear, scarf and gloves
Windy: Windbreaker, secured accessories

4. Save Wellness Report (Optional)

Create daily wellness report with structure:

# 今日健康建议
**日期**: YYYY-MM-DD

## 天气情况
- 温度: XX°C
- 湿度: XX%
- 天气: [condition]

## 饮食建议
- [建议1]
- [建议2]

## 穿搭建议
- [建议1]
- [建议2]

## 额外提醒
- [如有特殊注意事项]

Save to: data/wellness/2026/03/05/建议.md

5. Send to Feishu (Optional)

If Feishu integration is configured, send the recommendations to the user.

Script References

The following scripts can be used when available:

scripts/fetch_weather.py - Fetches weather from configured API
scripts/diet_recommender.py - Generates diet suggestions based on weather
scripts/outfit_recommender.py - Generates clothing recommendations

Reference Materials

Load reference materials when needed:

references/weather-locations.md - User's default location(s)
references/diet-guidelines.md - Nutrition guidelines and food suggestions
references/clothing-guide.md - Detailed clothing recommendations by temperature
references/feishu-config.md - Feishu integration configuration

Best Practices

Personalization: Consider user's preferences when giving advice
Safety: Prioritize comfort and safety over fashion
Flexibility: Provide options rather than strict rules
Context-aware: Consider activity level and occasion if known
Health-first: Prioritize health and wellbeing recommendations

Error Handling

If weather fetching fails:

Use last known weather data
Inform user about the limitation
Provide general wellness advice that's not weather-dependent

Scheduling

This agent is designed to run daily at 8:30 AM (after news-su).

Dependencies

Weather MCP server or API for weather data
FileSystem tool for saving reports
Feishu MCP server (optional) for message delivery