Deepgram Voice Workflow

• Outcome: Pass
• Latest audit: May 29, 2026 · 8:18 PM UTC
• Version: 0.1.0

Overview

This skill is a coherent Deepgram voice workflow, but users should understand that audio/text is sent to Deepgram and outputs are saved locally.

Install only if you are comfortable sending chosen audio and reply text to Deepgram. Use a scoped Deepgram API key, confirm whether /root/.openclaw/.env is an acceptable secret source, and clean up transcript, JSON, and MP3 outputs when they contain sensitive content.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 93% confidence
• Finding: The skill advertises and instructs use of shell scripts, environment variables, and file I/O, but it does not declare corresponding permissions. This creates a trust and transparency gap: a caller may invoke the skill without understanding that it reads secrets, accesses local files, and executes shell-based workflows, which can lead to unintended data exposure or unsafe execution in agent environments.

Context-Inappropriate Capability

Medium

• Confidence: 91% confidence
• Finding: The script automatically falls back to reading a Deepgram API key from a fixed privileged file under /root/.openclaw/.env. That behavior expands the script's access beyond explicitly supplied inputs and can cause unexpected credential use in contexts where the caller did not intend to grant this script access to root-scoped secrets.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill documentation encourages sending audio for transcription and text for speech synthesis to Deepgram, but it does not clearly warn users that voice content and generated text leave the local environment and are transmitted to a third-party service. In voice workflows, this can expose sensitive personal, biometric, or business information without informed user consent, especially in bot automation contexts.

Missing User Warnings

Medium

• Confidence: 89% confidence
• Finding: The script uploads the provided audio file to Deepgram's external API, which may expose sensitive voice content to a third party. In a voice automation skill, this transmission is functionally expected, but the lack of an explicit warning or confirmation increases the risk of unintentional disclosure of personal or regulated audio data.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: This script sends input audio to Deepgram for transcription and optionally sends reply text for TTS, but it provides no user-facing notice, consent check, or privacy guardrail before transmitting potentially sensitive voice content to a third-party service. In the context of voice-message automation for Telegram/QQ/OneBot workflows, the content may include personal data, private conversations, or regulated information, so silent exfiltration to an external API creates a real privacy and compliance risk.

VirusTotal

66/66 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.