Skip to content

Security

This document provides security usage recommendations for the Feishu plugin.

⚠️ Core Risk Warning

Important: This plugin connects to your work data (messages, documents, calendar, contacts) via Feishu APIs. What AI can read theoretically has a risk of leakage.

Security Recommendations

1. Prefer Personal Accounts

  • For now, prefer using personal Feishu accounts for testing and evaluation
  • Avoid connecting to sensitive data with company accounts
  • Expand usage scope after evaluation

2. Be Cautious in Group Chats

  • Recommendation: Avoid using in group chats to reduce data leakage risk
  • AI may read information it shouldn't in group chats
  • If needed in group chats, be sure to configure groupPolicy

3. Human Review for Important Operations

  • For important operations like sending, modifying, writing data
  • Be sure to "preview first, then confirm"
  • Never let AI operate in fully autonomous mode

4. Understand Permission Boundaries

  • User Authorization: Bot permissions are equivalent to user permissions
  • Application Authorization: Bot permissions are determined by app configuration
  • Periodically review and clean up unnecessary permissions

Operational Risks

AI Hallucination

  • AI may misunderstand your intent
  • May generate content that looks reasonable but is inaccurate
  • Verify important information

Irreversible Operations

  • Messages sent by AI on your behalf are sent in your name and cannot be recalled
  • Deleted documents cannot be recovered
  • Be cautious with delete operations

Principle of Least Privilege

App Permissions

  • Only enable necessary API permissions
  • Periodically review enabled permissions
  • Timely disable permissions no longer needed

Group Configuration

json
{
  "channels": {
    "feishu": {
      "groups": {
        "oc_xxxxx": {
          "groupPolicy": "allowlist",
          "requireMention": true,
          "tools": {
            "deny": ["feishu_drive_file"]
          }
        }
      }
    }
  }
}

Confirmation for Sensitive Operations

For the following operations, the plugin will require user confirmation:

  • Delete documents
  • Delete bitables
  • Batch messages
  • Batch operations

Audit and Monitoring

View Logs

Use diagnostic command to check plugin status:

bash
openclaw feishu-diagnose

Audit Recommendations

  • Regularly check message history
  • Pay attention to abnormal access behavior
  • Keep logs of important operations

Disclaimer

This software is released under the MIT License. When running, it calls Feishu Open Platform APIs, requiring compliance with:

Next Steps

Released under the MIT License.