Human Voice Content Editor

Security checks across malware telemetry and agentic risk

Overview

This skill appears to be a straightforward text-humanizing helper that may write a user-requested output file but shows no evidence of hidden, networked, destructive, or credential-related behavior.

Before installing, be aware that the skill may rewrite content and create an output file such as humanized.md. Use it on documents you intend to edit, choose output paths deliberately to avoid overwriting work, and review the rewritten text before publishing.

SkillSpector

By NVIDIA

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
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)

Lp3

Medium

Category: MCP Least Privilege
Confidence: 91% confidence
Finding: The skill instructs users to run a local script that writes an output file (`--output humanized.md`), which is a file-write capability, but the skill metadata declares no permissions. Undeclared write access is dangerous because it hides the skill's true capabilities from reviewers and can enable unintended modification or overwrite of local content when the skill is invoked.

Vague Triggers

Medium

Confidence: 82% confidence
Finding: The trigger text is very broad, such as 'use whenever the user mentions content sounding like AI' or 'checking for AI feel,' which can match many ordinary writing requests. Overbroad activation increases the chance the skill runs in contexts the user did not intend, causing unnecessary rewriting, tone alteration, or propagation of file-writing behavior into general workflows.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal