iaworker

Security checks across malware telemetry and agentic risk

Overview

This skill needs review because it presents placeholder visual analysis as diagnosis and uses online text-to-speech by default without clear privacy disclosure.

Install only if you are comfortable treating this as a prototype helper, not a reliable visual diagnostic tool. Do not rely on its repair instructions for vehicles, electrical systems, structural parts, or other safety-critical work without human verification. Use offline TTS or disable speech if the text may contain private details, and be aware that camera/local media access and optional model downloads are part of the intended workflow.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Description-Behavior Mismatch

High

Confidence: 98% confidence
Finding: The function advertised as image analysis does not inspect image contents in any meaningful way and instead returns canned text templates. In this skill's context, that can cause users to rely on fabricated diagnoses and repair guidance for physical tasks, creating real-world safety risks and deceptive system behavior.

Intent-Code Divergence

Medium

Confidence: 91% confidence
Finding: The module documentation states that the engine analyzes visual inputs and feeds results to step generation, but the implementation is largely placeholder-driven. This mismatch is dangerous because operators and downstream components may trust the output as grounded in real visual evidence when it is not.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill states that `gtts` is the default TTS engine but does not warn that generated spoken text may be sent to an external Google service for synthesis. Because this skill can narrate diagnostics, repair instructions, and potentially user-provided contextual data from images/videos, sensitive information could be disclosed off-device without the user's awareness.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: When gTTS is used, the spoken text is sent to Google's external service for synthesis. In this skill, the text may contain user-provided repair instructions, diagnostic details, or safety-related content, and the file does not provide any disclosure, consent flow, or privacy guardrails before transmitting that data off-device.

VirusTotal

67/67 vendors flagged this skill as clean.

View on VirusTotal