ClawHub
Back to skill

Security audit

ChatGPT Image Handoff

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed image-generation handoff workflow that uses local files and optional approved ChatGPT browser automation in ways that fit its purpose.

Before installing, expect this skill to create a local handoff folder, read approved reference images, scan a downloads folder for recent image outputs, and optionally use your already-authenticated ChatGPT browser session after confirmation. Keep the handoff directory out of public commits because it may contain local paths and working files.

SkillSpector

Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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 (1)

Lp3

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding
The skill instructs the agent to read and write local files via multiple Python scripts and handoff directories, but the skill metadata does not declare corresponding permissions. That mismatch can bypass user/operator expectations and weakens policy enforcement, especially because the workflow handles downloads, imports, ZIP archives, and browser-assisted external handoff with local state persistence.

Static analysis

No suspicious patterns detected.