User Request → MCP Tool Call → Feishu API → Return Result

User Request Examples:
- "Create a customer management system"
- "Help me make a project task management table"
- "Build an inventory management table"

Assistant Execution Process:

1. Analyze user requirements and design a reasonable table structure
   - Identify business type (CRM/project management/inventory/sales, etc.)
   - Determine core entities and attributes
   - Design field relationships and data flows

2. Call create_bitable to create the base table

3. Obtain the returned app_token

4. Call get_tables(app_token) to get table_id

5. Call add_table_field() to add fields: [Dynamically generated based on requirements]
   - Main table: Core business fields
   - Select field types (text/number/date/single-select/multi-select/people, etc.)
   - Set required fields and default values

6. Set field types, options, and validation rules
   - Single-select fields: Configure option lists
   - Number fields: Set range limits
   - Date fields: Configure time formats
   - Associated fields: Establish inter-table relationships

7. Add sample data

8. Add users as collaborators

9. Create associated tables (if needed)
   - Sub-table design (e.g., order details, task comments)
   - Inter-table association configuration
   - Data relationship description

User: Create a customer management system

Assistant Execution:

1. Requirement Analysis
   - Business Type: CRM Customer Management
   - Core Entities: Customers, Follow-up Records
   - Data Flow: Customer Information → Follow-up Interaction → Conversion

2. Call create_bitable(name="Customer Management System")

3. Obtain the returned app_token (e.g.: app_xxxxxxxxx)

4. Call get_tables(app_token) to get table_id (e.g.: tblxxxxxxxx)

5. Call add_table_field() to add fields:
   - Customer Name (field_name="Customer Name", field_type=1) Text
   - Contact Person (field_name="Contact Person", field_type=1) Text
   - Contact Phone (field_name="Contact Phone", field_type=11) Phone
   - Customer Source (field_name="Customer Source", field_type=3) Single-select
   - Customer Stage (field_name="Customer Stage", field_type=3) Single-select
   - Industry (field_name="Industry", field_type=3) Single-select
   - Estimated Transaction Amount (field_name="Estimated Transaction Amount", field_type=2) Number
   - Creation Time (field_name="Creation Time", field_type=5) Date
   - Remarks (field_name="Remarks", field_type=1) Text

6. Set field configurations
   - Customer Source Options: Online Promotion/Referral/Exhibition/Proactive Development
   - Customer Stage Options: Potential Customer/Interested/Negotiating/Converted/Lost
   - Industry Options: Internet/Finance/Manufacturing/Retail/Other
   - Estimated Amount: Set minimum value to 0

7. Add sample data
   Call add_record(app_token, table_id, {
     "Customer Name": "Sample Tech Company",
     "Contact Person": "Zhang San",
     "Contact Phone": "13800138000",
     "Customer Source": "Online Promotion",
     "Customer Stage": "Interested",
     "Industry": "Internet",
     "Estimated Transaction Amount": 50000,
     "Creation Time": 1704067200000,
     "Remarks": "Potential high-quality customer"
   })

8. Add user as collaborator
   Call get_user_by_email(email="user@example.com") to get open_id
   Call add_collaborator(app_token, "user", open_id, "edit")

9. Create associated table "Follow-up Records"
   Repeat steps 4-5 to create follow-up record table, including:
   - Follow-up Time (field_type=5)
   - Follow-up Method (field_type=3, options: Phone/WeChat/Email/In-person)
   - Follow-up Content (field_type=1)
   - Next Follow-up Time (field_type=5)
   - Customer Name (associated field linked to customer information table)

10. Return results
    - Table link: https://feishu.cn/base/{app_token}
    - Field description document
    - Usage suggestions

User: Create a project task management table

Assistant Execution:

1. Requirement Analysis
   - Business Type: Project Management
   - Core Entities: Projects, Tasks
   - Data Flow: Project Decomposition → Task Assignment → Execution Tracking

2. Call create_bitable(name="Project Task Management System")

3. Obtain the returned app_token

4. Call get_tables(app_token) to get table_id

5. Call add_table_field() to add fields to the "Project List" table:
   - Project Name (field_name="Project Name", field_type=1)
   - Project Number (field_name="Project Number", field_type=1)
   - Project Status (field_name="Project Status", field_type=3) Options: Not Started/In Progress/Completed/Delayed
   - Start Date (field_name="Start Date", field_type=5)
   - End Date (field_name="End Date", field_type=5)
   - Project Priority (field_name="Project Priority", field_type=3) Options: High/Medium/Low
   - Project Progress (field_name="Project Progress", field_type=15) 0-100%
   - Project Manager (field_name="Project Manager", field_type=1)

6. Add sample data
   Call add_record() to add 2-3 sample projects

7. Add user as collaborator

8. Create associated "Task List" table
   - Task Name (field_type=1)
   - Belongs to Project (associated field → Project List)
   - Task Owner (field_type=1)
   - Task Status (field_type=3) Options: To Do/In Progress/Completed/Canceled
   - Task Priority (field_type=3) Options: Urgent/High/Medium/Low
   - Start Time (field_type=5)
   - Deadline (field_type=5)
   - Estimated Hours (field_type=2)
   - Actual Hours (field_type=2)
   - Task Description (field_type=1)

9. Return dual-table structure description, association diagram, and usage suggestions

User: Create a purchase order management system

Assistant Execution:

1. Requirement Analysis
   - Business Type: Procurement Management
   - Core Entities: Purchase Orders
   - Data Flow: Order Creation → Approval → Execution

2. Call create_bitable(name="Purchase Order Management System")

3. Obtain the returned app_token

4. Call get_tables(app_token) to get table_id

5. Call add_table_field() to add fields:
   - Purchase Order (field_name="Purchase Order", field_type=1)
   - Unit Price (field_name="Unit Price", field_type=2)
   - Quantity (field_name="Quantity", field_type=2)
   - Total Amount (field_name="Total Amount", field_type=2)
   - Purchase Time (field_name="Purchase Time", field_type=5)
   - Purchaser (field_name="Purchaser", field_type=1)

6. Add sample data
7. Add user as collaborator
8. Return table link and usage instructions

User: Add 10 purchase records to table xyz

Assistant:
1. Call get_tables(app_token) to confirm table_id
2. Loop call add_record() to add data
3. Return addition results

User: Add user@example.com as table administrator

Assistant:
1. Call get_user_by_email(email="user@example.com") to get open_id
2. Call add_collaborator(
   app_token="...",
   member_type="user",
   member_id="<open_id>",
   perm_type="full_access"
)

feishu-integration/
├── plugin.json              # Plugin manifest
├── SKILL.md                 # This document
├── mcp-server/              # MCP Server
│   ├── index.py             # FastMCP implementation
│   ├── requirements.txt     # Python dependencies
│   └── README.md            # MCP documentation
├── scripts/                 # Python scripts (backup)
│   ├── create_feishu_app.py
│   ├── create_purchase_order_bitable.py
│   ├── add_feishu_collaborator.py
│   └── ...
└── .mcp.json                # MCP configuration

{
  "mcpServers": {
    "feishu": {
      "command": "python3.11",
      "args": ["/path/to/mcp-server/index.py"],
      "env": {
        "FEISHU_APP_ID": "your_app_id",
        "FEISHU_APP_SECRET": "your_app_secret"
      }
    }
  }
}

cd scripts
python3 create_feishu_app.py
python3 add_feishu_collaborator.py

pip install requests lark-oapi