ClawHub
Back to skill

Security audit

雄韬创作锚 - 创作就怕没有灵感

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed content-creation skill with optional local PDF/poster output and no evidence of hidden data access or persistence.

Installers should know this skill can strongly shape the assistant's writing style when invoked and can create local PDF or image files when those modes are requested. Use the file-generation modes only when you intend to create those artifacts and have the required local tools installed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • 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 (1)

Vague Triggers

High
Confidence
88% confidence
Finding
An overly broad activation phrase can cause the skill to trigger during ordinary conversation, pulling the model into this skill's instructions when the user did not explicitly intend to invoke it. In this case, the skill contains strong persona and workflow constraints plus file-generation behavior, so accidental activation could override normal assistant behavior and produce unexpected outputs or tool use.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.