ClawHub
Back to skill

Security audit

TikTok for Business MCP

Security checks for vulnerabilities and agentic risk

Overview

This TikTok Business skill is mostly transparent, but it can run broad account-changing MCP tools while presenting itself as mainly read-only.

Review before installing. This skill should only be used where the user understands it may do more than read TikTok Business data. Any call_tool use should be treated as potentially account-changing unless the live tool schema proves it is read-only, and budget, ad, access, or deletion operations should require explicit confirmation of the exact target and payload.

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

High
Confidence
96% confidence
Finding
The manifest advertises the skill as only for 'searching and reading data,' but the documented `call_tool` action is a generic invocation path that may create ads, change budgets, revoke access, or delete assets. This mismatch can mislead an orchestrator or human operator into invoking the skill under a read-only trust assumption, enabling unintended state-changing operations.

Intent-Code Divergence

High
Confidence
98% confidence
Finding
The safety section says untagged actions are reads and safe to run directly, yet `call_tool` is untagged while its own description warns it may perform write or destructive effects. This contradictory guidance creates a direct path for unsafe execution without confirmation, especially if an agent relies on tags or the safety summary rather than the detailed action description.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger scope 'ANY TikTok for Business MCP request' is overly broad and can cause the skill to be selected for loosely related requests without enough scrutiny. In this skill, broad routing is more dangerous because the action surface includes a generic `call_tool` mechanism that may perform sensitive mutations, increasing the chance of inappropriate or over-privileged invocation.

Static analysis

No suspicious patterns detected.