Groq Voice Transcriber

• Outcome: Review
• Latest audit: May 29, 2026, 9:23 PM
• Version:
  1.0.0

Overview

This skill does what it says by processing Telegram voice messages with Groq, but it automatically sends sensitive audio and transcripts to remote AI services and includes a hardcoded fallback API key.

Review before installing. Only use this in Telegram chats where participants understand that voice audio and transcripts are sent to Groq-hosted AI services, and remove the embedded fallback API key so the skill fails closed unless you provide your own GROQ_API_KEY.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill description states that incoming Telegram audio is intercepted, transcribed via Groq Whisper API, and then sent to an LLM, but it does not clearly warn users that their voice data and derived text leave Telegram and are transmitted to third-party services. This creates a real privacy and consent risk because users may unknowingly submit sensitive personal, business, or regulated information to external providers.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The skill sends raw audio to Groq's remote transcription API without any visible consent, disclosure, or data-handling notice. This creates a privacy and compliance risk because sensitive voice content may leave the local environment unexpectedly, especially if users assume processing is local.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The transcribed text is forwarded to a remote LLM service without any visible warning that user content is being transmitted off-system a second time. This compounds privacy exposure because spoken content may contain personal, confidential, or regulated information that users did not expect to be shared beyond transcription.

VirusTotal

63/63 vendors flagged this skill as clean.