ClawHub
Back to skill

Security audit

Gpt Review

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed ChatGPT browser-automation review tool, with privacy and file-overwrite cautions but no evidence of hidden or malicious behavior.

Install only if you are comfortable sending the reviewed article and prompt contents to ChatGPT/OpenAI through your logged-in Brave session. Avoid confidential, regulated, proprietary, or unpublished material unless your rules allow ChatGPT use, keep Brave remote debugging bound locally, and choose output paths carefully because existing files may be overwritten.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill automates sending article content to ChatGPT via a browser session and saving the returned content locally, but the description does not clearly warn users that their input will be transmitted to a third-party service and written to disk. This creates a real confidentiality and data-handling risk, especially if users submit unpublished, sensitive, or regulated content assuming the review stays local.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The script reads arbitrary prompt content from --prompt or --prompt-file and automates submission to chatgpt.com using a logged-in browser session, but it provides no explicit consent gate or warning that the content will be sent to an external third-party service. In an agent/skill context, this increases the chance that sensitive local data is exfiltrated to a remote provider without the operator fully realizing it.

Missing User Warnings

Low
Confidence
87% confidence
Finding
The script writes the extracted model response directly to a user-supplied output path without warning that the file may contain sensitive prompts, summaries, or copied data from a live authenticated session. This can lead to unintended local disclosure, accidental overwrite of existing files, or insecure retention of sensitive content.

Static analysis

No suspicious patterns detected.