Humanizer-Text AI写作痕迹去除

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:38 PM UTC
• Version:
  2.1.1

Overview

This is a simple text-rewriting skill with no code execution or credential access, but users should review edits because it is meant to make AI-like writing look more human.

Install only if you want an agent to rewrite text to sound less AI-generated. Use explicit prompts, review the output for changed facts or meaning, and avoid using it on legal, academic, compliance, or sensitive content unless rewriting provenance is appropriate.

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 (2)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The phrase "Or invoke directly when editing documents" is broad enough to let the skill trigger during ordinary editing workflows, not just when a user explicitly asks to humanize text. That increases the chance of unintended activation on unrelated content, which could silently rewrite text and undermine user intent, editorial integrity, or policy-sensitive wording.

Vague Triggers

Medium

• Confidence: 87% confidence
• Finding: The skill can be invoked on essentially any 'AI-generated text' without boundaries on content type, trust level, or safety-sensitive use cases. Broad trigger criteria increase the chance the skill is applied in inappropriate contexts, such as rewriting policy, legal, academic, or compliance-relevant text in ways that obscure provenance or alter meaning.

VirusTotal

66/66 vendors flagged this skill as clean.