ClawHub
Back to skill

Security audit

Seedream 5.0 Lite 广告图片

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed AI image-generation helper that sends user-provided prompts and optional reference images to AI Hive and saves generated images locally.

Before installing, treat prompts and selected reference images as data sent to AI Hive and its upload storage flow. Use only assets you are allowed to upload, protect the AI_HIVE_API_KEY or ~/.ai-hive/config.json credential, and set --output-dir if the default Downloads/AiHive location is not appropriate.

SkillSpector

Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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)

Lp3

Medium
Category
MCP Least Privilege
Confidence
83% confidence
Finding
The skill advertises executable behavior involving environment access, shell execution, file reads/writes, and network calls, but does not declare permissions or clearly scope those capabilities. This creates a trust gap: users may invoke the skill without understanding it can read local inputs, store config/results, and contact a remote API, increasing the risk of unintended data exposure or unsafe execution in agent environments.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The description frames the skill as an ad-image/landing-page alignment aid, but the documented behavior includes API key setup, remote uploads, task polling, metadata queries, and local downloads. That mismatch can mislead users about the operational and data-handling scope, causing them to provide sensitive assets or run the skill in environments where network transfer and local persistence were not expected.

Missing User Warnings

Low
Confidence
72% confidence
Finding
The skill states that generated images are downloaded but does not clearly tell users where files will be written or what naming/location rules apply. In agent or shared environments, undisclosed local writes can cause accidental overwrites, leakage into monitored directories, or retention of sensitive creative assets without user awareness.

Static analysis

No suspicious patterns detected.