ClawHub
Back to skill

Security audit

Question Methodology

Security checks across malware telemetry and agentic risk

Overview

This is a small markdown-only thinking framework with no executable code, network behavior, credential use, or hidden install mechanism, though it asks agents to use a broad checklist and optionally keep task records.

Install this if you want a general-purpose questioning and completion checklist. Be aware it may encourage the agent to apply the framework often, and do not let it write sensitive task details into memory files unless you intentionally want that record retained.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

Description-Behavior Mismatch

Low
Confidence
86% confidence
Finding
The skill instructs the agent to create persistent records in a memory file even though its stated purpose is only question clarification. Unnecessary persistence can cause inadvertent storage of user data, internal reasoning artifacts, or task metadata beyond what is needed, increasing privacy and data-retention risk.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The activation conditions are very broad, including generic situations like before significant work or when the user says to use judgment. This can cause the skill to trigger in many unrelated contexts, unexpectedly steering agent behavior and introducing its persistence/output rules where they were not requested.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.