ClawHub

Ramp Video

Security checks across malware telemetry and agentic risk

Overview

This skill is a cloud video-editing helper that sends user-provided media to NemoVideo for processing, with privacy considerations but no evidence of hidden or destructive behavior.

Install only if you are comfortable sending selected video files or URLs to nemovideo.ai for cloud processing. Protect NEMO_TOKEN, avoid private or confidential footage unless you trust the provider, and review provider terms or retention practices before using it for sensitive media.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The manifest presents a narrowly scoped speed-ramp tool, but the body of the skill exposes a much broader cloud video-editing and rendering interface, including generic editing, uploads, state inspection, SSE messaging, and export orchestration. This scope drift increases the chance that users and host systems grant trust, data access, or invocation privileges under a narrower expectation than what the skill actually performs.

Context-Inappropriate Capability

Low
Confidence
80% confidence
Finding
Including credit-balance checks and broader session/account management in a speed-ramp editing skill introduces unnecessary access to account-related backend functions beyond the core task. Even if not directly exploitable for account takeover, this expands the operational surface and may expose user metadata or enable unintended interactions with paid-service state.

Vague Triggers

Medium
Confidence
87% confidence
Finding
The startup prompt and example phrases are broad enough to trigger the skill on vague requests like "edit my raw video footage" or "export 1080p MP4," which can cause accidental activation. In this skill, accidental invocation is more sensitive because activation may lead to remote backend connection, token acquisition, session creation, and file upload handling without the user fully understanding that cloud processing is involved.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill encourages users to share raw video footage but does not prominently disclose in the description that files are transmitted to a third-party cloud backend for processing and export. This is a meaningful privacy and data-handling issue because raw videos often contain sensitive personal, location, biometric, or proprietary content, and users may assume local or first-party processing.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal