ClawHub
Back to skill

Security audit

Azure Ai Voicelive Py Free

Security checks across malware telemetry and agentic risk

Overview

This skill is not malicious, but it asks for broad command execution and handles microphone audio with insufficient user-facing privacy scoping.

Review before installing. Use this only in environments where broad command execution is acceptable, and make sure users understand microphone audio and transcripts may be sent to Azure. Avoid using it for confidential, regulated, or third-party conversations unless you add consent, recording indicators, and local controls around capture and upload.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (4)

Intent-Code Divergence

Medium
Confidence
90% confidence
Finding
The skill advertises itself as a limited VoiceLive/free-edition capability, yet the manifest enables `exec`, which grants general command execution beyond the documented purpose. This expands the attack surface substantially because a caller or downstream agent may invoke shell commands unrelated to audio streaming, secrets handling, or SDK use.

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
General command-execution capability is not justified by the stated function of connecting to Azure VoiceLive for WebSocket/audio/transcription tasks. In an agent setting, unnecessary `exec` privileges can be abused to run arbitrary local commands, inspect files, or chain into credential theft or persistence.

Missing User Warnings

High
Confidence
88% confidence
Finding
The skill describes microphone and transcription features without clearly warning users that live audio and transcript data are transmitted to Azure services. This can lead to unintentional disclosure of sensitive personal, corporate, or regulated information, especially in environments where users may assume processing is local.

Missing User Warnings

High
Confidence
92% confidence
Finding
The microphone example shows how to capture and upload live audio but omits any user-facing notice or consent step. In practice, copy-pasted example code could be deployed as-is, resulting in covert or poorly disclosed audio collection and transfer to a third-party cloud service.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.