ClawHub
Back to skill

Security audit

Image Inpainting — Pro Pack on RunComfy

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed RunComfy image-inpainting helper that sends user-provided image inputs to a third-party model service for the stated editing purpose.

Install only if you are comfortable using RunComfy for image edits and sending the relevant image and mask URLs to that service. Check that your agent invokes it only for intentional image-inpainting or local-edit requests, especially for broad requests like removing an object or watermark.

SkillSpector

By NVIDIA
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 (1)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The trigger list includes broad phrases such as "remove from image," "remove object," and "patch the photo," which can match ordinary user requests outside a deliberate image-inpainting workflow. This can cause unintended auto-invocation of a skill that uploads user-supplied images and masks to a third-party service, increasing the chance of accidental data exposure or performing edits the user did not intend through this specific provider.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.