A remote service bridge that enables AI agents to interact with the Feishu ecosystem through the Model Context Protocol (MCP) via HTTP APIs.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install feishu-mcp-remote
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install feishu-mcp-remote using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Feishu MCP Remote Service?
The Feishu MCP Remote skill is a high-performance integration designed for developers who need to connect their AI agents with the Feishu (Lark) collaboration suite. By utilizing the Model Context Protocol over a standard HTTP interface, this skill allows for deep integration with cloud documents, user directories, and file management systems without the need for local deployment.
As part of the Openclaw Skills ecosystem, this tool simplifies the complexity of enterprise API management. It provides a standardized way for LLM-powered agents to perform actions such as creating documents, searching for team members, and managing comments, all while maintaining strict security through User and Tenant authentication tokens.
Feishu MCP Remote Service Use Cases
- Automatically generating and updating project status reports directly within Lark cloud documents.
- Empowering AI assistants to find contact information for internal employees using natural language.
- Reading and synthesizing information from multiple Feishu documents to create consolidated summaries.
- Managing document feedback loops by programmatically reading and adding comments to specific doc blocks.
How Feishu MCP Remote Service Works
- The AI agent or client initiates a session by sending an initialize request to the Feishu MCP endpoint with valid UAT or TAT credentials.
- The system validates permissions and returns the protocol version and session parameters.
- The agent calls the tools/list method to discover available capabilities such as search-doc or create-doc.
- Specific tasks are executed via the tools/call method, passing the required arguments in a JSON-RPC 2.0 format.
- The remote service processes the request via the Feishu Open Platform and returns the structured result or error code to the agent.
Feishu MCP Remote Service Setup
To use this integration within the Openclaw Skills framework, follow these steps:
- Create a custom app on the Feishu Open Platform and obtain your App ID and App Secret.
- Apply for the necessary API permissions (e.g.,
docx:document:readonly,contact:user:search). - Generate an access token (TAT for app-level or UAT for user-level access).
- Configure your client to point to the remote endpoint:
https://mcp.feishu.cn/mcp.
# Example of setting up the required headers for a request
export FEISHU_TAT_TOKEN="t-gxxxxxxxxxxxxxx"
export ALLOWED_TOOLS="create-doc,fetch-doc"
Feishu MCP Remote Service Data Schema & Taxonomy
The skill communicates using JSON-RPC 2.0. Below is the organization of typical data exchange:
| Component | Type | Description |
|---|---|---|
jsonrpc |
String | Must be "2.0" |
method |
String | The MCP action (initialize, tools/list, tools/call) |
params |
Object | Contains the tool name and specific arguments (e.g., docID, content) |
headers |
Object | Includes authentication (X-Lark-MCP-TAT/UAT) and allowed tool filters |
result |
Object | The output of the tool execution, including text content or file metadata |
Feishu MCP Remote Service Advanced Features
- Multi-Identity Support: Seamlessly switch between User Access Tokens (UAT) and Tenant Access Tokens (TAT) based on the workflow requirements.
- Security Filtering: Use the X-Lark-MCP-Allowed-Tools header to restrict an agent's access to a specific subset of available tools.
- Rich Document Operations: Supports complex block-level updates and multi-format document searching (doc/docx).
- Comprehensive Error Handling: Detailed JSON-RPC error codes ranging from authentication failures to rate-limiting (429) triggers.
SKILL.md
Loading
Related Openclaw Skills
Automation Scripts
A comprehensive suite of automation tools designed to handle bulk file operations, data security, and scheduled task execution seamlessly.
Emoji Picker
A context-aware emoji and sticker selection skill that helps AI agents add visual flair and emotional tone to digital communications.
Web Scraper Pro
A professional-grade web scraping tool designed for automated data harvesting, multi-format table exports, and scheduled task execution.
PhoneBase Cloud Phone Control
A comprehensive CLI-driven skill for controlling Android cloud phones, enabling deep mobile app automation and UI interaction.
AI Avatar Generator
An AI-powered generation tool for creating personalized high-definition avatars and mobile wallpapers in various artistic styles.
Memory Fusion Lite
An efficient incremental memory management system that uses rolling distillation and weekly governance to maintain long-term agent context.
Featured*
