health-coach 💪

一个开源的 AI 健康管理技能，让你的 AI agent 变成专业的私人健康教练。

临床级营养分析、医学指标解读、运动编程、身体数据追踪——所有数据 100% 本地存储，零上传。

功能

• 🍽️ 饮食分析 — 拍照或文字描述 → 热量、宏量素、营养评估（内置中国菜热量数据库）
• 🏥 医学指标解读 — 血常规、生化全套、血脂、甲功、FeNO、激素、尿检
• 💪 运动指导 ��� 伤病感知的训练计划、心率区间分析、动作库
• 📊 体重追踪 — 体重/体脂/腰围趋势，7天均值，进度预测
• 💊 补剂建议 — 三级证据分类、药物交互提醒
• 💉 减重药物指南 — GLP-1（司美格鲁肽/替尔泊肽）临床数据、副作用、中国市场可获得性、个性化用药建议
• 📱 Apple Health 导入 — 一键解析 iPhone 健康数据（体重、运动、步数、心率、睡眠）
• 📈 进度报告 — 每日记录、周报、月报自动生成

快速开始

安装

npx skills add H1an1/health-coach -g -y

初始化个人档案

cd 你的工作目录
bash ~/.agents/skills/health-coach/scripts/init.sh .

交互式填写身高体重年龄，自动计算 BMR/TDEE/目标热量，生成个人档案。

导入 Apple Health 数据

1. iPhone → 健康 → 头像 → 导出所有健康数据
2. 把 export.zip 传到电脑
3. 解压后运行：

python3 ~/.agents/skills/health-coach/scripts/apple_health.py export/apple_health_export/export.xml --output health/ --days 30

生成报告

bash ~/.agents/skills/health-coach/scripts/report.sh . --weekly   # 周报
bash ~/.agents/skills/health-coach/scripts/report.sh . --monthly  # 月报

知识库

references/ 目录包含临床级参考资料：

隐私

所有健康数据 100% 存储在本地工作目录，不上传、不分享、不传输。技能完全通过本地文件和 AI agent 上下文运行。

项目结构

health-coach/
├── SKILL.md                    # 技能入口
├── config/                     # 用户配置模板
│   ├── profile.template.md
│   ├── goals.template.md
│   └── reminders.template.md
├── references/                 # 临床知识库
│   ├── nutrition.md
│   ├── medical-markers.md
│   ├── exercise.md
│   ├── supplements.md
│   └── apple-health.md
├── templates/                  # 报告模板
│   ├── daily-log.md
│   ├── weekly-report.md
│   └── monthly-report.md
└── scripts/
    ├── init.sh                 # 初始化脚本
    ├── report.sh               # 报告生成
    └── apple_health.py         # Apple Health 数据解析

更新日志

v0.2.0 — 2026-03-05

🆕 新增

• 中国品牌食品热量数据库 (references/cn-brands.md) — 611行，覆盖中国市场常见食品
  • 🧋 奶茶：乐乐茶、喜茶、霸王茶姬、瑞幸、星巴克、奈雪、茶百道（含少糖/缓糖版本）
  • 🍶 酒类：盒马果酒、啤酒(8品牌)、威士忌、伏特加、白酒、红酒 + 减脂饮酒排名
  • 🍡 甜品/烘焙：糖葫芦、好利来、鲍师傅、泸溪河
  • 🏪 便利店：全家、7-Eleven、便利蜂
  • 🛒 超市：山姆会员店（烘焙/熟食/冷冻/零食）、Costco开市客、盒马
  • 🥡 快餐：麦当劳、肯德基、汉堡王、赛百味、必胜客、吉野家、真功夫
  • 🥘 中式菜品：家常菜27道、粥粉面9种、外卖套餐8类
  • 🍢 火锅：19种食材 + 5种蘸料热量对比
  • 🔥 烧烤：11种常见串
  • 🌅 早餐：煎饼果子、油条、包子等13种
  • 🍮 甜品饮品：8种甜品 + 碳酸饮料/果汁/功能饮料26种
  • 🍫 零食品牌：良品铺子、三只松鼠、百草味
  • 🥛 乳制品：牛奶5品牌、酸奶5品牌、蛋白粉4品牌
  • 🍜 方便食品：11种（螺蛳粉、自热锅、拉面说等）
• 免责声明 — 中英双语免责声明（SKILL.md + README.md）
• 模型兼容性说明 — 标注基于 Claude Opus 4.6 开发测试

🔧 改进

• SKILL.md 饮食分析workflow增加 cn-brands.md 引用

v0.1.0 — 2026-03-03

• 初始版本发布
• 完整知识库：nutrition、medical-markers、exercise、supplements、medications、apple-health
• 脚本：init.sh、report.sh、apple_health.py
• 支持中英双语

贡献

欢迎 PR！特别需要：

• 📦 食品数据 — 更多品牌/菜系（日料、韩餐、印度菜、东南亚菜等）
• 🔗 数据源集成 — FatSecret API、Open Food Facts 等开放数据库对接
• 🏋️ 训练计划生成器
• 📱 Garmin/Fitbit 数据导入
• 🌍 更多语言支持

如何贡献食品数据

1. Fork 本仓库
2. 编辑 references/cn-brands.md，按现有格式添加数据
3. 数据来源请优先使用产品包装营养标签
4. 提交 PR，注明数据来源

许可

MIT

---

English

A comprehensive, open-source personal health management skill for AI agents.

Clinical-grade nutritional analysis, medical marker interpretation, exercise programming, and longitudinal health tracking — all running locally with zero data leakage.

Features

• 🍽️ Meal Analysis — Photo or text → calories, macros, nutritional assessment
• 🏥 Lab Interpretation — Blood panels, FeNO, thyroid, lipids, urinalysis
• 💪 Exercise Programming — Injury-aware training plans, heart rate zones
• 📊 Body Tracking — Weight, body fat, waist trends over time
• 💊 Supplement Guidance — Evidence-based, interaction-aware
• 💉 Weight Loss Medications — GLP-1 agonists (semaglutide, tirzepatide), clinical trial data, side effects, China market availability, personalized assessment
• 📱 Apple Health Import — Parse full XML export (weight, workouts, steps, HR, sleep)
• 📈 Progress Reports — Daily logs, weekly & monthly summaries

Quick Start

Install as a Skill

npx skills add H1an1/health-coach -g -y

Initialize Your Profile

cd your-workspace
bash ~/.agents/skills/health-coach/scripts/init.sh .

This creates a health/ directory with your personal profile, goals, and reminder config.

Import Apple Health Data

1. iPhone → Health → Profile → Export All Health Data
2. Transfer export.zip to your computer
3. Unzip and run:

python3 ~/.agents/skills/health-coach/scripts/apple_health.py export/apple_health_export/export.xml --output health/ --days 30

Generate Reports

bash ~/.agents/skills/health-coach/scripts/report.sh . --weekly
bash ~/.agents/skills/health-coach/scripts/report.sh . --monthly

Knowledge Base

The references/ directory contains clinical-grade knowledge:

Privacy

All health data stays 100% local in your workspace. Nothing is uploaded, shared, or transmitted. The skill operates entirely through local files and your AI agent’s context.

Changelog

v0.2.0 — 2026-03-05

🆕 New

• Chinese Brand Food Nutrition Database (references/cn-brands.md) — 600+ entries covering the Chinese market
  • Bubble tea (LELECHA, HEYTEA, CHAGEE, Luckin, Starbucks, Nayuki, ChaPanda)
  • Alcohol (craft beers, whisky, vodka, baijiu, wine + fat-loss ranking)
  • Supermarkets (Sam’s Club, Costco, FRESHIPPO/Hema)
  • Fast food (McDonald’s, KFC, Burger King, Subway, Pizza Hut, Yoshinoya)
  • Chinese dishes (27 home-style, 9 noodle/congee, 8 takeout combos)
  • Hotpot ingredients (19 items + 5 dipping sauces with calorie comparison)
  • BBQ, breakfast, desserts, snack brands, beverages
  • Dairy, protein powders, convenience store items, instant foods
• Disclaimer — Bilingual (CN/EN) disclaimer in SKILL.md and README.md
• Model Compatibility Notice — Developed and tested on Claude Opus 4.6

🔧 Improved

• SKILL.md meal analysis workflow now references cn-brands.md

v0.1.0 — 2026-03-03

• Initial release

Contributing

PRs welcome! Areas that need help:

• 📦 Food data — More brands/cuisines (Japanese, Korean, Indian, SEA, etc.)
• 🔗 Data source integration — FatSecret API, Open Food Facts
• 🏋️ Workout plan generators
• 📱 Garmin/Fitbit data import
• 🌍 Localization

How to contribute food data

1. Fork this repo
2. Edit references/cn-brands.md following the existing format
3. Prefer product packaging nutrition labels as data source
4. Submit a PR with source noted

Disclaimer / 免责声明

⚠️ This skill is for informational and educational purposes only. It does not provide medical diagnosis, treatment, or professional health advice. Always consult a qualified healthcare provider for medical concerns.

⚠️ 本技能提供的所有健康、营养、运动建议仅供参考，不构成医疗诊断或治疗建议。如有健康问题，请咨询专业医生。

Model Compatibility / 模型兼容性

This skill was developed and tested on Claude Opus 4.6. Different models may vary in accuracy — particularly for image-based tasks like meal photo analysis and lab result interpretation. If you’re using a different model, please verify outputs and adjust prompts as needed for your setup.

本技能基于 Claude Opus 4.6 开发和测试。不同模型的分析能力可能存在差异，尤其是图像识别类功能（如食物照片分析、化验单解读）。使用其他模型时，请结合自身情况验证结果，必要时调整提示词。

License

MIT