Ai Video Editor Ghaus Editz

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:39 PM UTC
• Version:
  1.0.0

Overview

This is a cloud video-editing skill that sends media and edit prompts to NemoVideo, and its behavior is generally coherent with that purpose.

Install only if you are comfortable sending selected media files, edit prompts, URLs, session metadata, and a NEMO_TOKEN or anonymous trial token to mega-api-prod.nemovideo.ai. Avoid private or confidential footage unless you trust that provider's retention and privacy practices, and prefer a disposable or revocable token.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Description-Behavior Mismatch

Low

• Confidence: 93% confidence
• Finding: The skill manifest presents itself as a simple video-editing tool, but the instructions also direct the agent to automatically mint anonymous tokens and create remote sessions with a third-party API before user work begins. That hidden setup materially changes the skill's behavior and trust model because it performs network authentication and account/session creation not clearly disclosed in the high-level description.

Description-Behavior Mismatch

Medium

• Confidence: 90% confidence
• Finding: The manifest frames the capability as editing uploaded video clips, but the implementation claims support for additional image and audio formats such as jpg, png, gif, mp3, wav, and others. This scope expansion can surprise users and host platforms because the skill may ingest and transmit more types of content than expected, increasing privacy and policy risk.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: Routing 'Everything else' to the SSE editing action means a very broad set of unmatched prompts will be forwarded to the backend service. This increases the chance of accidental data exfiltration, unintended remote actions, or misuse because ambiguous or unrelated user input is treated as an instruction for the cloud editing pipeline.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill tells users to upload clips and describe edits, but it does not prominently warn that both media files and prompts are transmitted to a remote cloud processing service. In a media-editing context this is especially sensitive because uploaded videos may contain personal, confidential, or copyrighted content, so lack of disclosure undermines informed consent.

VirusTotal

64/64 vendors flagged this skill as clean.