ClawHub

OpenGFX

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only logo generation skill; it includes a visible service-mode upload note, but no code, credentials, hidden data access, or persistence mechanism.

Safe to install for logo generation, with one caveat: if you use it in an ACP/service workflow, require explicit confirmation before generated assets, brand prompts, or metadata are uploaded to R2 or any external endpoint.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The learnings file broadens the skill from generating logo SVGs into producing broader social media assets and running an end-to-end pipeline that includes external upload and delivery. This scope expansion is security-relevant because operators and users may invoke the skill expecting local logo generation, while the documentation encourages additional behaviors with different data-handling and permission requirements.

Context-Inappropriate Capability

Medium
Confidence
97% confidence
Finding
The instruction to 'Upload to R2' introduces remote data exfiltration capability that is not justified by a logo-generation skill description. If implemented, generated assets, prompts, brand concepts, or embedded metadata could be transferred off-platform without the user's informed expectation, creating confidentiality and compliance risk.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The autonomous workflow describes external upload and delivery without any user-facing warning or consent step, which is unsafe for a skill that appears to generate files locally. Silent third-party transfer can expose sensitive brand materials, internal project names, or customer data and undermines the principle of informed consent for network actions.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal