Security audit

Feishu Openclaw Integration

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only Feishu/OpenClaw integration, but it needs Review because it can process enterprise chats, history, and files while privacy boundaries are under-disclosed.

Install only after confirming where OpenClaw and any model provider process data, using least-privilege Feishu scopes, limiting approved chats and file types, protecting app secrets, disabling broad ALL-group broadcasts by default, and making chat-history and file analysis explicit admin-approved features with retention and deletion rules.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The document explicitly claims '数据不离开企业' while multiple examples transmit message text, sender identifiers, chat history, and uploaded files to OpenClaw endpoints over HTTP. This is a material mismatch between stated privacy properties and actual data flows, which can mislead operators into deploying the integration under false assumptions about confidentiality and data residency.

Description-Behavior Mismatch

Medium

Confidence: 85% confidence
Finding: The skill is presented as a bot/customer-service/workflow integration, but it also documents retrieving historical group messages for broader analysis. That expanded surveillance capability increases privacy risk because administrators may enable the skill for limited messaging use while overlooking that it can aggregate and analyze conversation history.

Context-Inappropriate Capability

Medium

Confidence: 93% confidence
Finding: The file-processing example downloads uploaded files from Feishu and forwards their contents for analysis, which is broader and riskier than a simple chat bot integration. This can expose confidential documents, regulated data, or intellectual property to another service without adequate scoping, consent, or handling safeguards.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The message-handling example sends chat text and a user identifier to OpenClaw but does not warn administrators or end users that conversation content and metadata are being shared. In an enterprise messaging context, undisclosed transfer of identifiers and message contents can create privacy, compliance, and trust issues.

Missing User Warnings

Medium

Confidence: 96% confidence
Finding: The group analytics example fetches message history and sends it to OpenClaw for summarization without clearly warning about the privacy impact. Historical group chats often contain sensitive operational, personnel, or customer information, so silent export for analysis materially increases confidentiality and compliance risk.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The uploaded-file analysis flow forwards file contents for external processing without any explicit warning that sensitive documents may be transmitted. Because enterprise file uploads can contain contracts, HR records, source code, or customer data, this omission can lead to severe confidentiality and regulatory exposure.

Static analysis

No suspicious patterns detected.