Security audit

Vpn Rotator

Security checks across malware telemetry and agentic risk

Overview

The skill mostly describes VPN rotation, but it asks for powerful local/system access while its activation, inputs, outputs, and generic automation claims are inconsistent and under-scoped.

Review carefully before installing. Only use this skill in a controlled environment where changing VPN routes, storing VPN credentials, and running OpenVPN-related commands are acceptable. Do not grant broad exec/write authority or passwordless sudo unless you have audited and narrowed the instructions to VPN-only operations with explicit user confirmation.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (6)

Intent-Code Divergence

High

Confidence: 97% confidence
Finding: The documented return schema describes code-review scoring and compliance grading rather than VPN rotation results, which is inconsistent with the stated skill purpose. In an agent setting, this kind of semantic mismatch can cause downstream automations to trust and act on incorrect output types, leading to unsafe branching, mis-execution, or concealment of the skill’s real behavior.

Intent-Code Divergence

High

Confidence: 96% confidence
Finding: The input parameters describe generic content-auditing fields like content and strict_level instead of VPN-specific inputs such as target region, provider, credentials source, or rotation policy. This mismatch can trigger the skill in unintended contexts and let broad, unrelated user input reach a tool-enabled workflow with exec/write capability.

Intent-Code Divergence

Medium

Confidence: 90% confidence
Finding: The skill advertises broad file processing, API integration, command execution, and generic automation features beyond the documented VPN functionality. Capability inflation in a tool-enabled skill increases the chance of overbroad invocation and can mislead users or agents into granting the skill authority for unrelated operations.

Vague Triggers

High

Confidence: 91% confidence
Finding: The manifest-style description mixes unrelated use cases and lacks clear invocation boundaries, making it unclear when the skill should run. This ambiguity is dangerous in the context of a privileged skill because an agent may select it for tasks outside its intended scope and expose sensitive operations unnecessarily.

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The manifest-style description mixes unrelated use cases and lacks clear invocation boundaries, making it unclear when the skill should run. This ambiguity is dangerous in the context of a privileged skill because an agent may select it for tasks outside its intended scope and expose sensitive operations unnecessarily.

Missing User Warnings

Medium

Confidence: 93% confidence
Finding: The documentation instructs users to handle VPN credentials, configure system software, and use exec/write-capable operations, including sudo-related setup, without a prominent safety warning about system impact. In a tool-using agent environment, under-disclosed privilege requirements can lead to risky local changes, credential exposure, or unsafe command execution by unsuspecting users.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.