Free Subtitle Generator

• Outcome: Review
• Latest audit: May 29, 2026 at 9:32 PM UTC
• Version:
  1.0.0

Overview

This skill is a real cloud subtitle/video tool, but it asks for broader remote editing and upload authority than a user may expect from a subtitle-only skill.

Install only if you are comfortable sending selected videos, prompts, and editing state to mega-api-prod.nemovideo.ai. Use explicit captioning requests, avoid private or regulated media unless you trust that provider, and be aware the skill can perform broader cloud video editing and export workflows, not just generate subtitles.

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 (6)

Description-Behavior Mismatch

Medium

• Confidence: 96% confidence
• Finding: The skill is presented as a narrow subtitle generator, but the documentation grants broad cloud video-editing capabilities including uploads, state inspection, rendering, overlays, audio changes, and general editing. This scope mismatch can cause users and calling agents to authorize data transfer and powerful actions they did not intend, increasing the risk of deceptive consent, overbroad invocation, and misuse of uploaded media or session state.

Context-Inappropriate Capability

Low

• Confidence: 83% confidence
• Finding: Including credit balance checks and subscription/export-tier handling goes beyond the stated subtitle-generation purpose and exposes account and billing-related state to the skill workflow. Even if not directly malicious, this expands access to user/account information and can be used to steer users into registration or paid flows unrelated to the narrow advertised task.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: Routing essentially all unmatched requests to this skill creates an overbroad trigger surface, allowing unrelated user prompts to be sent to a cloud backend with editing/export capabilities. In context, this is especially dangerous because the skill can upload media, create sessions, and perform remote actions, so accidental activation may expose user data or trigger unintended processing.

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The phrase "Or just tell me what you're thinking" is an overly vague invocation that encourages activation on general conversation rather than a specific subtitle task. This increases the chance that unrelated user content, including sensitive media plans or files, gets funneled into the skill's remote processing pipeline without clear intent.

Missing User Warnings

Medium

• Confidence: 97% confidence
• Finding: The skill asks users to share video files and describes cloud processing, but it does not provide a clear privacy warning about uploading potentially sensitive media to a third-party backend. Because videos can contain faces, voices, locations, screens, and other personal data, lack of disclosure undermines informed consent and materially increases privacy risk.

Missing User Warnings

Low

• Confidence: 84% confidence
• Finding: The setup flow silently uses an environment token or auto-generates an anonymous token without clearly warning the user that authentication material is being accessed or created. This can surprise users and operators, blur account boundaries, and cause actions to occur under credentials or trial accounts they did not knowingly authorize.

VirusTotal

64/64 vendors flagged this skill as clean.