Keyframes Video Editing

• Outcome: Review
• Latest audit: May 29, 2026 at 9:35 PM UTC
• Version:
  1.0.0

Overview

This cloud video-editing skill appears purpose-aligned, but it deserves review because it can automatically connect to a third-party backend and route broad or ambiguous prompts to that service.

Install only if you are comfortable sending video files, editing prompts, timeline state, and render jobs to NemoVideo's cloud API. Use it for explicit video-editing requests, avoid sensitive media unless you understand the provider's privacy and retention terms, and prefer confirming before any upload or automatic anonymous token setup.

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 (4)

Context-Inappropriate Capability

Low

• Confidence: 88% confidence
• Finding: The skill derives local install-path/platform information and sends it to a remote service via attribution headers, even though that data is not necessary to perform video editing. This creates unnecessary host-environment disclosure that can be used for user/tool fingerprinting and service-side profiling without clear user consent.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The invocation examples and setup language are broad enough that common phrases like 'export' or generic editing requests may activate the skill in contexts where the user did not specifically intend cloud video processing. Because activation can lead to token acquisition, session creation, and eventual remote uploads, ambiguous triggering increases the risk of unintended data transfer and surprising side effects.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The catch-all routing rule sends 'everything else' to the SSE edit path, which is too ambiguous for a networked skill that can process user content remotely. In practice, loosely related prompts could be interpreted as editing commands, causing unintended backend interaction, state changes, or uploads with insufficient user intent verification.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill description and onboarding do not clearly warn users that their video files and editing instructions are sent to a cloud backend for processing. This is dangerous because media files can contain sensitive personal, commercial, or biometric information, and users may disclose such content without informed consent about third-party transmission and retention.

VirusTotal

66/66 vendors flagged this skill as clean.