Add Subtitle To Video Online

• Outcome: Review
• Latest audit: May 29, 2026 at 9:33 PM UTC
• Version:
  1.0.0

Overview

This appears to be a real cloud video captioning skill, but it is broader than its subtitle-focused name suggests and sends media to a third-party backend.

Install only if you are comfortable sending selected videos, prompts, and edit instructions to NemoVideo's cloud API. Avoid confidential, regulated, or third-party-sensitive footage unless you trust the provider's privacy and retention practices. Keep NEMO_TOKEN private, and confirm before using non-subtitle editing features such as BGM, overlays, timeline edits, or exports.

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
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (4)

Description-Behavior Mismatch

Medium

• Confidence: 94% confidence
• Finding: The skill is marketed as a subtitle-generation utility, but its instructions expose a much broader remote video-editing pipeline, including timeline manipulation, uploads, audio changes, state inspection, and export. This capability mismatch increases the chance of over-collection, unintended processing, and misuse because users and orchestrators may invoke a more powerful cloud editor than the declared purpose suggests.

Context-Inappropriate Capability

Low

• Confidence: 84% confidence
• Finding: The skill performs token acquisition, session creation, and credit/balance handling that go beyond a simple local subtitle tool and are only partially aligned with the declared purpose. While these functions may be operationally necessary for the backend, exposing them without clear necessity or boundaries expands the attack surface around authentication state and billing-related interactions.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The invocation rules route broad, generic video-editing language such as export, edit, add BGM, overlays, and audio-track actions to this skill, despite its declared subtitle-focused purpose. Overbroad matching can cause unintended activation for unrelated prompts, leading users to send media or trigger remote processing they did not intend for this specific skill.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill instructs the agent to upload user video files to a third-party cloud backend and export processed results, but it does not prominently warn users that their media leaves the local environment. For potentially sensitive videos, this omission creates privacy and compliance risk because users may not understand that remote storage and processing occur.

VirusTotal

65/65 vendors flagged this skill as clean.