Humanizer 1.0.0

Security checks across malware telemetry and agentic risk

Overview

This is a text-editing skill that openly rewrites AI-sounding prose and does not contain hidden code, network behavior, persistence, or destructive instructions.

Install this only if you want an agent to rewrite pasted text or named documents. Specify exact files, review edits before accepting them, and do not use it to hide AI involvement where school, work, publishing, or legal rules require disclosure.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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)

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The skill explicitly aims to make text sound 'more natural and human-written,' which can facilitate concealment of AI assistance or bypass policies that require disclosure of machine-generated content. Although the stated purpose is editing quality, the capability is dual-use and can be abused for academic dishonesty, deceptive publishing, or evasion of AI-detection and provenance controls.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal