ClawHub
Back to skill

Security audit

AI图片加水印 | 批量水印 | 版权水印 | 品牌水印 | 青虎AI

Security checks for vulnerabilities and agentic risk

Overview

This skill provides local image watermarking instructions that match its stated purpose, with only minor activation-scope caveats.

Before installing, expect the agent to run local ImageMagick or ffmpeg commands on images you provide and to create output files such as a watermarked directory. Review batch commands before running them, especially in folders with many images, and install ImageMagick only from your normal trusted package manager.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger description contains broad phrases such as '加水印' and 'watermark' combined with '必须触发', which can cause the skill to activate on ordinary image-editing requests without sufficient disambiguation. Over-broad mandatory activation can route unrelated user tasks into shell-command-based workflows, increasing the chance of incorrect tool use or unintended file operations.

Vague Triggers

Low
Confidence
82% confidence
Finding
The activation guidance gives only a couple of positive examples and does not define clear boundaries for when the skill should not be used. This underspecification can lead to accidental invocation for adjacent tasks, especially because the skill includes executable command recipes and file-processing behavior.

Static analysis

No suspicious patterns detected.