Linkedin Humanizer

Security checks across malware telemetry and agentic risk

Overview

This skill does not run code, but it is explicitly designed to make AI-written LinkedIn content look human and pass detector checks.

Review carefully before installing. Use it only as an editing aid for content you are comfortable publishing, require explicit approval for any added facts, numbers, anecdotes, or voice details, avoid pasting sensitive DMs unless necessary, and do not use it to bypass disclosure rules, platform authenticity expectations, or AI-detection systems.

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 (2)

Vague Triggers

Medium

Confidence: 94% confidence
Finding: The skill is described with broad, high-level rewriting language and is positioned as a general-purpose way to make AI-generated LinkedIn content appear authentically human and pass AI detectors. That increases the chance it will be invoked for generic rewriting requests beyond a narrowly scoped editing task, including deceptive evasion-oriented use cases, which is risky given the explicit focus on bypassing detection signals.

Vague Triggers

Low

Confidence: 89% confidence
Finding: The activation condition 'when a draft feels off' is vague and subjective, making the skill easier to invoke in loosely related contexts without clear safeguards. In this file's context, that ambiguity compounds the broader deceptive purpose of rewriting text to look human and evade AI detection, so even a low-severity ambiguity meaningfully expands misuse opportunities.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal