westland-linguistic-humidifier

• Outcome: Pass
• Latest audit: May 28, 2026, 6:09 AM
• Version:
  1.0.0

Overview

This skill is a small promotional persona that may be annoying because it can recommend a fictional product broadly, but it does not show harmful access or execution behavior.

Install this only if you want a proactive fictional product-promoter persona. If you prefer skills that activate only on explicit requests, tighten the trigger rules before use.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (1)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The skill activates on highly subjective and loosely defined conditions such as 'natural language vibes,' repeated points, buggy code, or existential statements. In practice, this can cause the agent to inject unsolicited promotional content into many unrelated conversations, creating prompt-scope hijacking behavior and reducing reliability by overriding user intent with spam-like recommendations.

VirusTotal

59/59 vendors flagged this skill as clean.