ClawHub

Free Ai Video Generator

Security checks across malware telemetry and agentic risk

Overview

This is a cloud video-generation skill that uses NemoVideo tokens and uploads user-provided prompts or media, with weak privacy wording but no evidence of hidden or malicious behavior.

Install only if you are comfortable with NemoVideo receiving your prompts, uploaded images or media, inferred language, and generated outputs. Avoid confidential or regulated content unless you trust the provider's terms, and keep NEMO_TOKEN private.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Intent-Code Divergence

Medium
Confidence
93% confidence
Finding
The documentation materially misrepresents the trust and authentication model by claiming no authentication is needed while earlier sections instruct the skill to obtain and use a NEMO_TOKEN and create backend sessions. This can cause users or operators to unknowingly permit outbound authentication and remote processing, undermining informed consent and safe deployment review.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs automatic backend connection and anonymous token acquisition on first use while explicitly keeping setup messaging brief, without a clear consent prompt describing network activity, token issuance, and remote transfer of user prompts or uploaded media. This is risky because sensitive user content may be sent to a third party before the user meaningfully understands or approves that behavior.

Natural-Language Policy Violations

Medium
Confidence
83% confidence
Finding
The skill sends a detected language value to the backend without documenting how detection occurs, what locale data is derived, or allowing the user to choose or override it. While lower impact than credential or code-execution flaws, it creates unnecessary metadata disclosure and can leak inferred user characteristics to the remote service.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal