ClawHub
Back to skill

Security audit

Youtube Transcriber

Security checks across malware telemetry and agentic risk

Overview

This skill mostly does what it says, but it should be reviewed because it uploads audio to OpenAI without a runtime consent gate and has unsafe handling of custom output paths.

Review before installing. Use it only with YouTube audio you are allowed to send to OpenAI, avoid sensitive or confidential recordings, and avoid custom output paths until the path interpolation issue is fixed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill clearly instructs use of shell commands and reads/writes transcript files, but the manifest shown in SKILL.md does not declare permissions. That mismatch is a real security issue because users and policy systems cannot accurately assess or constrain the skill's capabilities, especially since it invokes external tools and writes to arbitrary output paths.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The README states that audio may be downloaded and sent to the OpenAI Whisper API, but it does not clearly warn users that third-party data transfer occurs and may include sensitive spoken content. In a transcription skill, users may process private meetings, interviews, or other confidential material, so the lack of an explicit privacy/data-sharing notice can lead to unintentional disclosure.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger phrases include broad natural-language requests such as 'get transcript' and 'what does this video say,' which can match ordinary conversation and cause unintended invocation. In this skill, accidental activation is more concerning because invocation leads to shell execution, external network access, possible API charges, and file creation.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
When subtitles are unavailable, the script uploads downloaded audio to the OpenAI transcription API without an explicit user-facing warning at execution time that media content is being transferred to a third party. In an agent skill context, this can expose private, copyrighted, sensitive, or regulated audio content unexpectedly, especially if users assume all processing is local.

External Transmission

Medium
Category
Data Exfiltration
Content
CURL_ARGS=(
  -s
  -X POST
  "https://api.openai.com/v1/audio/transcriptions"
  -H "Authorization: Bearer ${OPENAI_API_KEY}"
  -F "file=@${WHISPER_INPUT}"
  -F "model=whisper-1"
Confidence
90% confidence
Finding
https://api.openai.com/

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.