ClawHub
Back to skill

Security audit

Cloudflare Workers Ai Images

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent Cloudflare Workers AI image helper, with ordinary credential and third-party upload risks that users should handle carefully.

Before installing, confirm you are comfortable sending prompts and any img2img source images to Cloudflare Workers AI. Use a least-privilege Cloudflare API token, keep credentials out of checked-in compose files and chat logs, and prefer temporary output paths unless you explicitly want to save generated images.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Tp4

High
Category
MCP Tool Poisoning
Confidence
95% confidence
Finding
The skill promises temporary chat-handoff behavior, but the documented interface allows arbitrary output paths and leaves delivery and cleanup to the caller. That mismatch can lead an orchestrator or user to assume safer lifecycle controls than actually exist, causing unintended persistent file writes, weak data-handling guarantees, and broader-than-expected model/network usage.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The documentation instructs users to pass long-lived Cloudflare credentials directly as container environment variables and even suggests verifying them with `env | grep '^CF_'`, which can normalize unsafe handling of secrets. While this is a common deployment pattern, it increases the chance of accidental exposure through compose files, shell history, debugging output, process inspection, or logs if operators follow the example without stronger secret-handling guidance.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.