ClawHub

zmg-srew

Security checks across malware telemetry and agentic risk

Overview

This is a content-only storyboard writing skill with no code execution, credentials, file access, persistence, or hidden system behavior.

Safe to install as a writing aid. Users should request another language if needed and provide verified source material for news, policy, documentary, or compliance-sensitive scripts because the skill explicitly depends on user-supplied facts.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The invocation description is broad enough to capture many ordinary creative-writing or scripting requests, which can cause the skill to activate outside its intended narrow domain. That increases the chance of unintended routing, policy bypass-by-specialization, or the model adopting this skill’s rigid behavior when a general response would be safer and more appropriate.

Natural-Language Policy Violations

Medium
Confidence
90% confidence
Finding
Forcing a Chinese title and Chinese-oriented output without user opt-in can override user language preferences and create prompt-routing confusion, especially in multilingual environments. This can lead to unusable outputs, accidental disclosure through mistranslation, or failure to satisfy user intent when downstream consumers expect another language.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
A fixed Chinese template at the output stage operationalizes the language lock-in, making the skill consistently produce content in a language the user may not want. Because the format is mandatory and central to the skill’s response, this increases the likelihood of persistent misalignment and downstream workflow errors in collaborative or automated pipelines.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal