ClawHub
Back to skill

Security audit

AI图像提示词-专业版

Security checks for vulnerabilities and agentic risk

Overview

This skill is mainly an image-prompt workflow tool, but its trigger and command/write permissions are broader than that purpose clearly supports.

Review this skill before installing. It is not malicious from the inspected artifact, but only use it when you intentionally want prompt files, reports, or prompt repositories created or changed, and avoid letting its broad trigger apply to unrelated coding or deployment tasks.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Behavioral ASTexec() Call, eval() Call, Dynamic Import
  • 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)

Vague Triggers

High
Confidence
90% confidence
Finding
An overly broad trigger can cause the agent to invoke this skill in unrelated contexts, including coding, deployment, or filesystem-affecting tasks not appropriate for an image-prompt tool. Because the skill advertises exec, write, and Git-oriented operations, ambiguous routing increases the chance of unintended command execution or local state changes under the wrong user intent.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill documents export/save/convert operations without clearly warning that local files may be created, modified, or overwritten. In an agent environment with write and exec capabilities, silent filesystem effects can lead to accidental data loss, clobbering of user files, or persistence of unwanted artifacts.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The examples show writing prompts and reports to directories and files but omit any warning about filesystem changes or overwrite risks. In context, the skill exposes write and exec tools, so users may unknowingly authorize destructive or persistent file operations based on seemingly harmless examples.

Static analysis

No suspicious patterns detected.