Image To Video Create

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:41 PM UTC
• Version:
  1.0.0

Overview

This is a coherent cloud image-to-video skill, but users should understand that selected media, URLs, and prompts may be sent to NemoVideo for processing.

Install only if you are comfortable sending chosen images, prompts, URLs, and generated media to NemoVideo’s cloud API. Avoid sensitive private media, use a dedicated token if available, and ask the agent to confirm before uploading, fetching a URL, or exporting when your request is broad or ambiguous.

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

Medium

• Confidence: 93% confidence
• Finding: The skill is presented as a narrow JPG-photo-to-video tool, but the documented behavior allows URL ingestion and a much broader set of file formats. That scope expansion increases the chance of users or upstream agents sending remote content or unsupported media into an external service without informed consent, which can create privacy, policy, and data-handling risks beyond the advertised purpose.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The trigger phrases are broad and generic enough that the skill may activate on unrelated user requests such as 'export' or 'convert my images' without clear intent to use this external cloud tool. Unintended invocation is dangerous here because it can lead to unexpected uploads, token-backed API actions, or cloud processing of user media.

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The routing table contains an 'Everything else' catch-all that sends prompts to the SSE backend, effectively turning ambiguous input into arbitrary remote processing. In a skill with network access and media manipulation, this expands behavior beyond user expectations and increases the likelihood of accidental data disclosure or unintended service actions.

Missing User Warnings

Medium

• Confidence: 96% confidence
• Finding: The skill sends user images and prompts to a third-party cloud backend but does not clearly warn users up front that their content leaves the local environment for remote processing. This is a material transparency and privacy issue, especially for product photos or other potentially sensitive media.

VirusTotal

66/66 vendors flagged this skill as clean.