Shadow AI Monitor

• Outcome: Pass
• Latest audit: May 28, 2026 at 6:37 AM UTC
• Version:
  1.0.0

Overview

This skill is a disclosed local dashboard generator, but users should handle its employee-monitoring reports and optional external delivery carefully.

Install only if you are authorized to analyze employee AI-tool usage. Treat generated reports as sensitive, store them in approved locations, avoid the CDN-backed dashboard in strict offline environments, and do not monitor workplace communications or send metrics to WhatsApp without legal, HR, privacy, and security approval.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Intent-Code Divergence

Medium

• Confidence: 95% confidence
• Finding: The documentation states there are 'No External API Calls,' but the dashboard technology section explicitly says Chart.js is loaded from the jsDelivr CDN. That creates a network dependency and a supply-chain exposure surface, and it can mislead users operating under strict privacy or offline assumptions.

Context-Inappropriate Capability

Medium

• Confidence: 86% confidence
• Finding: The weekly report example adds WhatsApp delivery even though the skill is described as a local dashboard/report generator. Sending metrics through a third-party messaging channel can leak compliance and employee-risk information outside the intended local processing boundary.

Missing User Warnings

Medium

• Confidence: 88% confidence
• Finding: The workflow advises monitoring Slack/Teams/Email for AI tool mentions and logging them to JSON, but it omits consent, minimization, legal review, retention, and employee-notice safeguards. In a monitoring/compliance context, this omission materially increases privacy, labor, and regulatory risk.

Ssd 3

Medium

• Confidence: 91% confidence
• Finding: The documentation proposes collecting employees' AI-tool mentions from workplace communications and logging them into a dataset. That is a concrete surveillance/data-collection pattern involving potentially sensitive employee and message metadata, and it can expose private communications if implemented without strict governance.

VirusTotal

65/65 vendors flagged this skill as clean.