ClawHub
Back to skill

Security audit

Feishu Custom Bot

Security checks for vulnerabilities and agentic risk

Overview

This skill appears to be a legitimate Feishu bot messaging helper, but its description is broader than its documented capabilities and should be reviewed before use.

Install only if you want an agent to send Feishu/Lark custom bot webhook messages through an OOMOL-connected account. Confirm every message payload before execution, and do not rely on this skill for reading or updating Feishu data unless the publisher narrows or expands the documented action set.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The skill description claims broad Feishu support for reading, creating, and updating data, but the documented connector actions only send webhook messages. This capability mismatch can mislead an agent into invoking the skill for tasks it cannot safely or correctly perform, increasing the chance of unintended writes, bad assumptions about available reads, or user-deceptive behavior.

Intent-Code Divergence

Low
Confidence
74% confidence
Finding
The safety section says untagged actions are reads, but the listed actions are all write operations. Even if no current read action exists, this inaccurate safety model can train an agent to assume unspecified actions are harmless, which is risky in a security-sensitive automation context.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The invocation text says to use this skill for ANY Feishu Custom Bot request and even for reading/updating data, which is broader than the documented functionality. Overbroad routing language can cause the agent to select this skill in inappropriate contexts, leading to wrong-tool use, accidental message sends, or failure to respect least-privilege behavior.

Static analysis

No suspicious patterns detected.