dws

Original：🇨🇳 Chinese

Translated

13 scripts

Manage DingTalk product capabilities (AI Table/Calendar/Address Book/Group Chat & Bots/Todo/Approval/Attendance/Logs/DING Messages/Workbench/Open Platform Documents, etc.). Use this when users need to operate table data, manage schedules and meetings, query address books, manage group chats, send messages via bots, create todos, submit approvals, check attendance, and submit daily/weekly reports (DingTalk log templates).

2installs

Sourcedingtalk-real-ai/dingtalk-workspace-cli

Added on2026-03-28

NPX Install

npx skill4agent add dingtalk-real-ai/dingtalk-workspace-cli dws

SKILL.md Content (Chinese)

View Translation Comparison →

DingTalk Full Product Skill

Manage DingTalk product capabilities via the

dws

command.

STRICTLY PROHIBITED (NEVER DO)

Never use methods other than the dws command to operate (prohibit curl, HTTP API, browsers)
Never fabricate identifiers like UUID or ID; must extract them from command returns
Never guess field names/parameter values; must query and confirm before operation

STRICTLY REQUIRED (MUST DO)

All commands must include
```
--format json
```
to get parsable output
Dangerous operations must be confirmed with the user first; add
```
--yes
```
to execute only after user consent
Batch operations cannot exceed 30 records at a time
All commands must strictly follow the parameter formats specified in the corresponding product reference documents (e.g., if there are parameter values, there must be at least one space between the parameter and its value)

Product Overview

Product	Purpose	Reference Document
`aitable`	AI Table: CRUD operations for tables/data tables/fields/records/template search	aitable.md
`approval`	Approval: Approval form/initiate instance/approval/revoke	simple.md
`attendance`	Attendance: Check-in record/schedule query	attendance.md
`calendar`	Calendar: Schedule/participant/meeting room/busy-free query	calendar.md
`chat`	Group Chat & Bots: Search groups/create groups/manage group members/modify group names/bot mass messaging/one-on-one chat/recall/Webhook/bot search	chat.md
`contact`	Address Book: User query(current user/search/details)/Department query(search/sub-departments/member list)	contact.md
`devdoc`	Open Platform Documents: Search development documents	simple.md
`ding`	DING Messages: Send/recall (in-app/SMS/phone call)	ding.md
`report`	Logs: Create via template/inbox/sent/view template/details/read statistics	report.md
`todo`	Todo: Create(including priority/deadline)/query/modify/mark as complete/delete	todo.md
`workbench`	Workbench: Application management	workbench.md

Intent Judgment Decision Tree

If the user mentions "table/multi-dimensional table/AI table/record/data" →

aitable

If the user mentions "approval/leave/reimbursement/business trip/overtime" →

approval

If the user mentions "attendance/check-in/schedule" →

attendance

If the user mentions "schedule/calendar/meeting room/appointment" →

calendar

If the user mentions "group chat/create group/group member/group management/bot send message/Webhook/bot mass messaging/bot one-on-one chat/notification" →

chat

If the user mentions "address book/colleague/department/organizational structure" →

contact

If the user mentions "development/API/call error document" →

devdoc

If the user mentions "DING/emergency message/phone reminder" →

ding

If the user mentions "log/daily report/weekly report/log statistics/write daily report/submit weekly report/send log/fill log" →

report

If the user mentions "todo/TODO/task reminder" →

todo

If the user mentions "workbench/application management" →

workbench

Key Distinctions: aitable(data table) vs todo(todo task) Key Distinctions: report(DingTalk log/daily/weekly report) vs todo(todo task) Key Distinctions: chat send-by-bot(send message as bot) vs send-by-webhook(custom bot Webhook alert)

For more confusing scenarios, see intent-guide.md

Dangerous Operation Confirmation

The following operations are irreversible or high-impact. You must first show the operation summary to the user and obtain explicit consent before execution. Add

--yes

to execute only after consent.

Product	Command	Description
`aitable`	`base delete`	Delete the entire AI Table, including all data tables and records
`aitable`	`record delete`	Delete records (supports batch operation)
`calendar`	`event delete`	Delete schedule, and all participants will be canceled synchronously
`calendar`	`participant delete`	Remove schedule participants
`calendar`	`room delete`	Cancel meeting room reservation
`chat`	`group members remove`	Remove group members
`todo`	`task delete`	Delete todo

Confirmation Process

Step 1 → Display operation summary (operation type + target object + scope of impact)
Step 2 → User explicitly confirms (e.g., "Confirm" / "Yes")
Step 3 → Add --yes to execute the command

Core Process

As an intelligent assistant, your primary task is to understand the user's true and complete intent, rather than simply executing commands. Before selecting a

dws

product command, you must strictly follow the four-step process below:

Intent Classification: First, determine which category the core verb/action of the user's instruction belongs to. This is more important than focusing on nouns.
Ambiguity Handling and Information Inquiry: If the user's instruction is ambiguous or contains keywords from multiple products, never guess. You must actively ask the user to clarify the intent. This is your core value as an intelligent assistant rather than a command executor.
Precise Product Mapping: After completing the first two steps and clarifying the intent, refer to the Product Overview and Intent Judgment Decision Tree to select the product.
Fully read the product reference documents, and implement the user's intent by writing code or directly calling commands.

Error Handling

When encountering an error, retry once with
```
--verbose
```
If
```
RECOVERY_EVENT_ID=<event_id>
```
appears in stderr, prioritize executing the recovery closed-loop according to recovery-guide.md
If it still fails, report the complete error information to the user; never attempt alternative solutions on your own
For authentication failures, refer to the authentication section in global-reference.md for handling
High-frequency errors and troubleshooting processes for each product can be found in error-codes.md

Detailed References (Read On Demand)

references/products/ — Detailed references for each product's commands
references/intent-guide.md — Intent Routing Guide (comparison of confusing scenarios)
references/global-reference.md — Global flags, authentication, output format
references/field-rules.md — AI Table field type rules
references/error-codes.md — Error codes + debugging process
references/recovery-guide.md — Recovery closed-loop,
```
RECOVERY_EVENT_ID
```
,
```
execute/finalize
```
specifications
scripts/ — Batch operation scripts for each product (AI Table/Calendar/Bot Messages/Address Book/Attendance/Logs/Todo, etc.)

dws