dingtalk-workflow-health-diary

Health Diary (dingtalk-workflow-health-diary)

Record diet and physical condition through conversation, and AI analyzes correlation patterns between food and physical discomfort. Compared with the Feishu version, DingTalk's exclusive advantage: can integrate attendance clock-in data to indirectly infer sleep and wake patterns.

Trigger Methods

"记录饮食"  "健康日记"  "吃了什么"  "身体不舒服"  "健康追踪"  "午餐记录"

Working Modes

Mode 1: Text Description Recording

Trigger: "Had braised beef noodles and an iced Americano for lunch" / "Had a sandwich for breakfast today"

Workflow:

Step 0: 确保数据表已初始化（见下方"数据初始化"章节）
    └── dws aitable base list → 检查是否存在"健康日记" Base
        不存在 → 执行完整初始化流程（创建 Base → 创建表 → 创建字段）

Step 1: 解析用户输入
├── 提取餐次（早餐/午餐/晚餐/加餐）
├── 提取食物列表和份量
└── 推断日期（"昨天"→-1天、"前天"→-2天，默认今天）

Step 2: AI 检查关联风险
└── dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --keyword "<食物名>" --yes -f json
    如果有高风险食物（关联强度="高"） → 立即提醒

Step 3: 写入饮食记录（⚠️ 必须包含"标题"主字段）
└── dws aitable record create --base-id <hd-base-id> --table-id <diet-table-id> --records '[{"cells":{"标题":"2026-04-06 午餐","日期":"2026-04-06","餐次":"午餐","食物列表":"红烧牛肉面, 冰美式"}}]' --yes -f json

Step 4: 如果有风险，发送提醒
└── dws ding message send --robot-code "bot_code" --users "manager6950" --content "..." --type "1" --yes

Output Format:

已记录饮食：
| 日期 | 餐次 | 食物 |
|------|------|------|
| 2026-04-06 | 午餐 | 红烧牛肉面(一大碗), 冰美式(大杯) |

⚠️ 提醒：之前 3 次喝冰美式后 1-2 小时内都出现了胃部不适，请注意观察。

Mode 2: Physical Condition Recording

Trigger: "My stomach feels a bit uncomfortable in the afternoon" / "I didn't sleep well today" / "I feel a bit dizzy"

Workflow:

Step 1: 解析身体状态
├── 状态描述（胃部不适/头痛/疲劳/失眠等）
├── 严重程度（轻微/中度/严重）
└── 时间（自动记录当前时间，格式 "YYYY-MM-DDTHH:mm:ss"）

Step 2: AI 回溯检查今日饮食
└── dws aitable record query --base-id <hd-base-id> --table-id <diet-table-id> --limit 50 --yes -f json
    → 从中筛选今日记录，检查时间差

Step 3: 查询历史关联模式
└── dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --limit 50 --yes -f json
    → 查看是否有食物关联

Step 4: 写入身体状态记录
└── dws aitable record create --base-id <hd-base-id> --table-id <health-table-id> --records '[{"cells":{"标题":"2026-04-06 胃部不适","时间":"2026-04-06T15:00:00","状态描述":"胃部不适","严重程度":"轻微"}}]' --yes -f json

Step 5: 如果发现新的关联模式，更新关联分析表
└── dws aitable record create --base-id <hd-base-id> --table-id <correlation-table-id> --records '[{"cells":{"标题":"冰美式-胃部不适","食物":"冰美式","状态":"胃部不适","时间差":"1-2小时","出现次数":1,"关联强度":"中"}}]' --yes -f json
    或更新已有记录：
    dws aitable record update --base-id <hd-base-id> --table-id <correlation-table-id> --records '[{"recordId":"<id>","cells":{"出现次数":3,"关联强度":"高"}}]' --yes -f json

Mode 3: Correlation Analysis

Trigger: "What foods tend to make my stomach uncomfortable?" / "Analyze the relationship between diet and health" / "Are there any food correlations"

Workflow:

Step 1: 查询所有饮食记录和身体状态记录
├── dws aitable record query --base-id <hd-base-id> --table-id <diet-table-id> --limit 100 --yes -f json
└── dws aitable record query --base-id <hd-base-id> --table-id <health-table-id> --limit 100 --yes -f json

Step 2: 查询已有关联分析
└── dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --limit 50 --yes -f json

Step 3: AI 执行关联分析
├── 按食物分组统计身体状态出现频率
├── 计算时间差模式（吃X后多久出现症状）
├── 计算关联强度（高/中/低，基于出现次数和时间一致性）
└── 排除偶然关联（出现次数 < 2 的不报告）

Step 4: 更新关联分析表并输出报告
└── 对每对新的食物-状态关联，写入或更新 correlation 表

Mode 4: Weekly Report Generation

Trigger: "Generate this week's health report" / "Health weekly report"

Workflow:

Step 1: 查询本周数据
├── dws aitable record query --base-id <hd-base-id> --table-id <diet-table-id> --limit 100 --yes -f json
├── dws aitable record query --base-id <hd-base-id> --table-id <health-table-id> --limit 100 --yes -f json
└── dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --limit 50 --yes -f json

Step 2: 查询考勤数据（钉钉独有）
└── dws attendance record get --user "manager6950" --date "2026-04-06" --yes -f json
    → 推断作息规律（上班时间 → 间接了解起床时间）
    → 注意：dws attendance summary 不可用，仅使用 record get 获取考勤数据

Step 3: AI 生成周报

Weekly Report Format:

## 健康周报（MM/DD - MM/DD）

### 饮食概览
- 记录天数：N 天
- 总餐次：M 次
- 饮食规律性：较好/一般/较差

### 身体状态
- 不适记录：N 次
- 主要症状：胃部不适 X 次、疲劳 Y 次

### 作息规律（钉钉独有）
- 平均上班打卡时间：HH:MM
- 平均下班打卡时间：HH:MM
- 工作时长：约 N 小时

### 关联发现
| 食物 | 状态 | 出现次数 | 典型时间差 | 关联强度 |
|------|------|---------|-----------|---------|
| 冰美式 | 胃部不适 | 3次 | 1-2小时 | 高 |
| 辣味火锅 | 失眠 | 1次 | 6-8小时 | 中 |

### 建议
- 冰美式与胃部不适有强关联，建议减少摄入或改用热饮
- 辣味火锅建议避免晚餐食用

Mode 5: Proactive Reminder

Trigger: Automatically triggered when the user records a meal containing high-risk foods (embedded in Step 2 of Mode 1)

Workflow:

Step 1: 用户记录饮食 → 提取食物名称关键词

Step 2: 查询关联分析表
└── dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --keyword "<食物名>" --yes -f json

Step 3: 如果发现高关联强度记录（出现次数 ≥ 3 且关联强度="高"） → 立即提醒

Data Initialization

On first use, first check whether the Base already exists; if not, create it according to the complete process:

# Step 0: 检查是否已初始化
dws aitable base list --yes -f json | grep "健康日记"
# 如果不存在，执行以下步骤：

# Step 1: 创建 Base
dws aitable base create --name "健康日记" --yes -f json
# 返回 baseId，记为 <hd-base-id>

# Step 2: 创建 3 张表（⚠️ 每张表自动创建一个 primaryDoc 类型的"标题"主字段）
dws aitable table create --base-id <hd-base-id> --name "饮食记录" --yes -f json
# 返回 tableId，记为 <diet-table-id>
dws aitable table create --base-id <hd-base-id> --name "身体状态" --yes -f json
# 返回 tableId，记为 <health-table-id>
dws aitable table create --base-id <hd-base-id> --name "关联分析" --yes -f json
# 返回 tableId，记为 <correlation-table-id>

# Step 3: 为每张表添加业务字段
# ⚠️ 字段类型必须用英文名: text / date / number / select / multiselect
# ⚠️ 字段参数名必须用 "fieldName"（不是 "name"）

# 饮食记录表
dws aitable field create --base-id <hd-base-id> --table-id <diet-table-id> \
  --fields '[{"fieldName":"日期","type":"date"},{"fieldName":"餐次","type":"text"},{"fieldName":"食物列表","type":"text"}]' \
  --yes -f json

# 身体状态表
dws aitable field create --base-id <hd-base-id> --table-id <health-table-id> \
  --fields '[{"fieldName":"时间","type":"text"},{"fieldName":"状态描述","type":"text"},{"fieldName":"严重程度","type":"text"},{"fieldName":"睡眠质量","type":"text"}]' \
  --yes -f json

# 关联分析表
dws aitable field create --base-id <hd-base-id> --table-id <correlation-table-id> \
  --fields '[{"fieldName":"食物","type":"text"},{"fieldName":"状态","type":"text"},{"fieldName":"时间差","type":"text"},{"fieldName":"出现次数","type":"number"},{"fieldName":"关联强度","type":"text"}]' \
  --yes -f json

Table Structure Overview

Table 1: Diet Records — tableId: <diet-table-id>

Field Name	Type	Source	Description
Title	primaryDoc	Auto-created	Primary field, recommended to fill in "YYYY-MM-DD meal"
Date	date	Manually created	Record date, format "YYYY-MM-DD"
Meal	text	Manually created	Breakfast/Lunch/Dinner/Snack
Food List	text	Manually created	Food names and portions, separated by commas

Table 2: Physical Condition — tableId: <health-table-id>

Field Name	Type	Source	Description
Title	primaryDoc	Auto-created	Primary field, recommended to fill in "YYYY-MM-DD condition description"
Time	text	Manually created	Condition record time, format "YYYY-MM-DDTHH:mm:ss"
Condition Description	text	Manually created	Such as "stomach discomfort", "fatigue"
Severity	text	Manually created	Mild/Moderate/Severe
Sleep Quality	text	Manually created	Good/Average/Poor (optional, requires the user to provide it proactively)

Table 3: Correlation Analysis — tableId: <correlation-table-id>

Field Name	Type	Source	Description
Title	primaryDoc	Auto-created	Primary field, recommended to fill in "Food - Status"
Food	text	Manually created	Associated food
Status	text	Manually created	Associated body status
Time difference	text	Manually created	Typical delay (e.g., "1-2 hours")
Occurrence count	number	Manually created	Associated occurrence count
Association strength	text	Manually created	High/Medium/Low

dws CLI command reference

Create health diary (full initialization)

# 1. 创建 Base
dws aitable base create --name "健康日记" --yes -f json

# 2. 创建表（每张表自动生成"标题"主字段）
dws aitable table create --base-id <hd-base-id> --name "饮食记录" --yes -f json
dws aitable table create --base-id <hd-base-id> --name "身体状态" --yes -f json
dws aitable table create --base-id <hd-base-id> --name "关联分析" --yes -f json

# 3. 创建业务字段（类型用英文: text/date/number，参数用 fieldName）
dws aitable field create --base-id <hd-base-id> --table-id <diet-table-id> \
  --fields '[{"fieldName":"日期","type":"date"},{"fieldName":"餐次","type":"text"},{"fieldName":"食物列表","type":"text"}]' --yes -f json
dws aitable field create --base-id <hd-base-id> --table-id <health-table-id> \
  --fields '[{"fieldName":"时间","type":"text"},{"fieldName":"状态描述","type":"text"},{"fieldName":"严重程度","type":"text"},{"fieldName":"睡眠质量","type":"text"}]' --yes -f json
dws aitable field create --base-id <hd-base-id> --table-id <correlation-table-id> \
  --fields '[{"fieldName":"食物","type":"text"},{"fieldName":"状态","type":"text"},{"fieldName":"时间差","type":"text"},{"fieldName":"出现次数","type":"number"},{"fieldName":"关联强度","type":"text"}]' --yes -f json

Record diet (⚠️ must include the "Title" primary field)

dws aitable record create --base-id <hd-base-id> --table-id <diet-table-id> \
  --records '[{"cells":{"标题":"2026-04-06 午餐","日期":"2026-04-06","餐次":"午餐","食物列表":"红烧牛肉面(一大碗), 冰美式(大杯)"}}]' \
  --yes -f json

Record body status (⚠️ must include the "Title" primary field)

dws aitable record create --base-id <hd-base-id> --table-id <health-table-id> \
  --records '[{"cells":{"标题":"2026-04-06 胃部不适","时间":"2026-04-06T15:00:00","状态描述":"胃部不适","严重程度":"轻微"}}]' \
  --yes -f json

Query records

# 查询饮食记录
dws aitable record query --base-id <hd-base-id> --table-id <diet-table-id> --limit 50 --yes -f json

# 按关键词查询
dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --keyword "冰美式" --yes -f json

# 查询身体状态
dws aitable record query --base-id <hd-base-id> --table-id <health-table-id> --limit 50 --yes -f json

# 查询关联分析
dws aitable record query --base-id <hd-base-id> --table-id <correlation-table-id> --limit 50 --yes -f json

Update records

# 更新关联分析记录（需要 recordId）
dws aitable record update --base-id <hd-base-id> --table-id <correlation-table-id> \
  --records '[{"recordId":"<recordId>","cells":{"出现次数":3,"关联强度":"高"}}]' --yes -f json

Query attendance (work-rest reference, DingTalk exclusive)

# 获取打卡记录（测试通过；未设考勤组时返回空记录，属正常现象）
dws attendance record get --user "manager6950" --date "2026-04-06" --yes -f json

# 注意：dws attendance summary 服务端报"统计类型错误"(code C0002)，不可用

DING urgent reminder

dws ding message send --robot-code "bot_code" --users "manager6950" --content "提醒：你刚吃了冰美式，之前多次在喝完后出现胃部不适" --type "1" --yes

Scheduled tasks

Task	Cron	Description
Health weekly report	`0 20 * * 0`	Generate weekly report every Sunday at 20:00

CronCreate: cron="0 20 * * 0", prompt="生成本周的健康报告并保存到知识库", recurring=true, durable=true

Linked Skills

Linked Skill	Data interaction	Description
morning-brief	Provide dietary reminders	Add dietary notes to the morning briefing
business-advisor	Provide health data	The advisory team focuses on the relationship between work efficiency and health

Notes

All dws commands must include the --yes flag
aitable field creation:
- The type must use theEnglish name：text / date / number / select / multiselect(cannot use the Chinese "text" or the numeric code 1/2)
- Parameter names must use "fieldName"(cannot use "name"）
Each table automatically creates a primaryDoc type "Title" primary field; when creating records,must includethis field
Correlation analysis requires at least 2 data points before reporting, to avoid false positives
Association strength determination: 3 or more occurrences = High, 2 occurrences = Medium, 1 occurrence = Do not report
Attendance data is only used as a work-rest reference, not as a basis for health judgments
DING messages are only used for high-risk food reminders, to avoid frequent interruptions
The DingTalk CLI has no image recognition command (unlike Feishu); users need to describe food in text
Attendance query dws attendance record get works normally, but dws attendance summary returns a server-side error and is unavailable
When the organization has not set up an attendance group, attendance record get returns empty records, which is normal

Known issues

dws attendance summary The server reports "statistics type error" (code C0002) and is unavailable. For attendance data, use dws attendance record get instead

About this skill

Trigger Methods

Working Modes

Mode 1: Text Description Recording

Mode 2: Physical Condition Recording

Mode 3: Correlation Analysis

Mode 4: Weekly Report Generation

Mode 5: Proactive Reminder

Data Initialization

Table Structure Overview

dws CLI command reference

Create health diary (full initialization)

Record diet (⚠️ must include the "Title" primary field)

Record body status (⚠️ must include the "Title" primary field)

Query records

Update records

Query attendance (work-rest reference, DingTalk exclusive)

DING urgent reminder

Scheduled tasks

Linked Skills

Notes

Known issues