ClawHub
Back to skill

Security audit

IQ Skill

Security checks across malware telemetry and agentic risk

Overview

This is a coherent IQ and puzzle generator with no evidence of hidden credential use, network access, persistence, purchases, or destructive behavior, though users should control any generated files and treat scores as informal.

Install only if you want local IQ-style puzzles, brain games, or browser-based quiz files. Ask the agent to confirm filenames before writing files, do not overwrite important files, review generated HTML before sharing it, and treat any IQ ranges or performance labels as informal practice feedback rather than a validated assessment.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

Medium
Confidence
98% confidence
Finding
This is a true issue: both `format_as_markdown` and `format_as_text` include `**Answer:**` and explanations inline with every question, and markdown is the default output format. For an IQ-test generator, this defeats the core purpose of testing or challenging the user and can leak the answer key immediately to downstream users or agents expecting quiz-only content.

Vague Triggers

Medium
Confidence
76% confidence
Finding
The description is broad enough to match generic requests for quizzes, riddles, assessments, and interactive content, which can cause over-invocation outside the user's precise intent. Over-broad routing is dangerous because it may activate a skill with file-generation behavior in contexts where the user only wanted simple text content.

Vague Triggers

Medium
Confidence
81% confidence
Finding
The activation criteria list many common user intents without boundaries or negative examples, making accidental triggering more likely. In this skill, that matters because some workflows move beyond simple text generation into script use and file creation, so a routing mistake can escalate from benign chat output to local side effects.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The instruction to save the final HTML file does not warn the user that local files will be created or modified. This is dangerous because it creates a hidden side effect: an agent may write browser-executable content to disk without informed consent, potentially overwriting files or leaving unexpected artifacts.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal