ClawHub

Video Screen

Security checks across malware telemetry and agentic risk

Overview

This is a cloud video-editing skill that sends user-provided media to a third-party rendering service, with privacy and metadata caveats but no evidence of hidden, destructive, or unrelated behavior.

Install only if you are comfortable sending your videos, images, audio, prompts, and remote media URLs to the third-party video service for processing. Avoid uploading confidential screen recordings or private footage unless you have checked the service's retention, billing, and privacy terms, and keep any NEMO_TOKEN value out of chat and logs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The skill directs the agent to infer the host platform from the local install path and transmit that attribution on every request, which is unnecessary for basic video editing and reveals host-environment metadata to a third party. While not directly yielding code execution, it expands data collection beyond user expectations and can aid fingerprinting, tracking, or environment-specific targeting.

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The skill's documented upload flow supports arbitrary remote URLs and a broad set of media/file types beyond the stated screen-recording use case. This scope expansion can enable unintended third-party fetching, processing of unrelated content, and possible misuse as a generic file-transfer/proxy mechanism without clear justification or guardrails.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The skill tells users to send video clips and states that processing occurs on remote GPU nodes, but it does not present a clear user-facing privacy notice or consent step before transferring potentially sensitive screen recordings off-device. Screen recordings often contain credentials, personal data, internal documents, or other confidential material, making undisclosed remote upload a meaningful privacy and compliance risk.

Missing User Warnings

Low
Confidence
80% confidence
Finding
The skill instructs the agent to use an environment token directly or obtain an anonymous token, but gives no guidance on protecting the credential, limiting its exposure, or avoiding disclosure in chat/logs. This is a weaker issue than direct secret exfiltration, yet it still normalizes unsafe credential handling practices around bearer tokens tied to service access and credits.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal