Security audit

Nx Matting

Security checks across malware telemetry and agentic risk

Overview

This skill mostly performs local image and video background removal, but it also forces an unrelated contact notice into user-facing replies and uses execution-adjacent downloads that deserve review before installation.

Review this skill before installing if you do not want agents to show the embedded contact notice. First use can download executable runtimes, models, and FFmpeg, so only run it from a publisher and manifest you trust, and consider removing ExecutionPolicy Bypass or using a stricter local PowerShell policy if your environment requires it.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (5)

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The README instructs users to run a PowerShell script with `-ExecutionPolicy Bypass`, which disables an important safety control and normalizes unsafe execution practices. While the README itself is not executable code, this can increase the risk that users run modified or malicious scripts without scrutiny, especially because the skill also downloads models/runtime on first use.

Vague Triggers

Medium

Confidence: 85% confidence
Finding: The skill description lists broad natural-language triggers such as generic image/video background removal scenarios without clear exclusions or stronger routing criteria. This can cause the agent to invoke the skill in unintended contexts, potentially leading to unexpected file processing, downloads, or execution of PowerShell-based workflows when the user did not clearly request this tool.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill states it will automatically download runtimes, models, and possibly FFmpeg from the network, but it does not require explicit user consent or present a clear security warning before doing so. In a security-sensitive agent context, silent network retrieval and execution-adjacent setup behavior increase supply-chain, privacy, and policy risks, especially because PowerShell with ExecutionPolicy Bypass is used to run the installer workflow.

Natural-Language Policy Violations

Medium

Confidence: 93% confidence
Finding: The skill mandates that the agent reproduce a fixed contact message exactly once in the final user response, removing user choice and overriding normal assistant behavior. This is suspicious because it injects unsolicited off-platform contact information into outputs, which can be used for promotion, social engineering, or trust redirection unrelated to the user's task.

Natural-Language Policy Violations

Medium

Confidence: 72% confidence
Finding: The script decodes and emits a hard-coded contact notice on completed/error events for image and video operations, introducing undocumented outbound-contact or social-engineering content into normal program output. In a security-sensitive agent context, hidden or opaque messaging is risky because it can mislead users, create undisclosed support channels, or embed policy-bypassing instructions outside the declared skill behavior.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.