Maker Free Best

• Outcome: Review
• Latest audit: May 29, 2026 at 9:34 PM UTC
• Version:
  1.0.0

Overview

This cloud video-editing skill is mostly purpose-aligned, but it can automatically connect to a remote backend and route broad user prompts or media to NemoVideo without clear upfront user control.

Review before installing. Use this only for footage and prompts you are comfortable sending to NemoVideo's cloud service, prefer a dedicated low-privilege token, and make sure the agent confirms uploads, exports, and ambiguous edit requests instead of automatically forwarding them.

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
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Context-Inappropriate Capability

Low

• Confidence: 89% confidence
• Finding: The skill instructs the agent to derive installation-source attribution data from local install paths and send it as headers to a third-party backend, even though that information is not needed to create videos. This creates unnecessary environment disclosure and user/tooling fingerprinting, which expands data sharing beyond the user’s expected task scope.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The example trigger phrases are extremely broad and natural-language-like, increasing the chance that ordinary conversation will unintentionally activate the skill. That can cause users to be connected to the backend and have prompts or files processed remotely without sufficiently clear intent.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The routing table contains a catch-all rule that sends 'everything else' to the SSE workflow, which effectively routes a wide range of vague user input to the remote editing backend. In a chat setting, this can over-collect prompts and trigger remote actions beyond what the user clearly requested.

Missing User Warnings

Medium

• Confidence: 94% confidence
• Finding: The skill encourages users to share raw footage and prompts but does not clearly warn that both media and instructions are sent to a remote backend service. This undermines informed consent, especially because uploaded videos may contain sensitive personal, biometric, or location data.

Missing User Warnings

Low

• Confidence: 80% confidence
• Finding: The skill notes internally that export jobs may continue remotely and become orphaned if the session closes, but it does not surface this risk clearly to the user. Users may believe processing stops when they leave, which can lead to confusion, unintended remote processing, or lost outputs.

Natural-Language Policy Violations

Medium

• Confidence: 76% confidence
• Finding: The skill hard-codes session creation with language set to English without user choice or justification. This can cause user prompts to be misinterpreted, especially for multilingual users, increasing the risk of unintended edits, exports, or mishandling of content.

VirusTotal

66/66 vendors flagged this skill as clean.