AIButton

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent generator for Vision Node AI button files and ZIP packages, with expected local file output but no evidence of hidden, credential-stealing, destructive, or unrelated behavior.

Before installing, understand that using this skill will create generated control files and ZIP packages on your machine, typically under ~/Desktop/VNAIButton/. Review the generated index.html before deployment, especially because the button communicates with its parent page and should only be embedded in a trusted Vision Node environment.

SkillSpector

By NVIDIA

Vulnerability Patterns

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
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: The skill description promises local file creation and ZIP packaging but does not clearly warn the user that it will write artifacts to the local filesystem. Hidden or implicit file writes can surprise users, cause unintended persistence of generated content, and normalize agent-side filesystem modification without explicit consent.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The packaging section explicitly writes to ~/Desktop/VNAIButton/ and deletes temporary directories, yet the skill does not present these side effects as a clear consent boundary before execution. This makes the skill more dangerous because a user may invoke a content-generation workflow without realizing it includes local writes and deletion operations.

VirusTotal

59/59 vendors flagged this skill as clean.

View on VirusTotal