Integrate Claude Code with Feishu CLI for Seamless AI‑Powered Office Automation

Overview

This article explains how to connect the newly open‑sourced Feishu CLI to Claude Code, enabling an AI agent to operate Feishu instant messaging, cloud documents, calendars, and other services.

Installation and Configuration

Official CLI repository: https://github.com/larksuite/cli

Prerequisites: Claude Code must already be installed, which implies Node.js and npm are also present.

Install the CLI

# Install CLI
npm install -g @larksuite/cli

# Install CLI SKILL (required)
npx skills add larksuite/cli -y -g

Login and Authorization

# 1. Initialize app credentials (interactive, one‑time)
lark-cli config init

# 2. Login with recommended permissions
lark-cli auth login --recommend

# 3. Verify authorization status
lark-cli auth status

Supported Feishu Capabilities

Calendar: view, create, invite participants, query busy status, get time suggestions.

Instant Messaging: send/reply messages, create/manage group chats, view chat history, search messages, download media.

Cloud Docs: create, read, update, search documents, read/write assets and whiteboards.

Cloud Space: upload/download files, search documents and knowledge bases, manage comments.

Multi‑dimensional Tables: create and manage tables, fields, records, views, dashboards, and perform data aggregation.

Spreadsheets: create, read, write, append, find, and export sheet data.

Tasks: create, query, update, complete tasks; manage checklists, subtasks, comments, and reminders.

Knowledge Base: create and manage knowledge spaces, nodes, and documents.

Address Book: search users by name/email/phone, retrieve user information.

Email: browse, search, read, send, reply, forward emails, manage drafts, listen for new mail.

Video Conference: search meeting records, retrieve minutes and recordings.

Claude Code Verification

Open Claude Code (restart if already open) and test whether the SKILL is usable. For example, ask the SKILL to send a message to a user:

Send a message to "Java一条人": "Meeting at 2:30 PM – Service Mid‑Platform Architecture Design"

The execution result in Claude Code is shown below:

The message received on the Feishu side appears as follows:

Additional Application Scenarios

Upload cloud documents

Edit multi‑dimensional tables

Read knowledge base content

Query the address book

Manage schedule tasks

Send email notifications

Extension: If the Lark desktop client is installed, the same SKILL can be used for scheduled tasks or to design Feishu dashboards.

Conclusion

By leveraging the open‑source Feishu CLI together with Claude Code, developers can achieve seamless integration with the Feishu ecosystem, enabling AI‑assisted office automation across messaging, document management, data processing, scheduling, and knowledge management scenarios.