Model Replacement Prompt Generator

• Outcome: Pass
• Latest audit: Jun 7, 2026 at 11:13 PM UTC
• Version:
  1.0.0

Overview

This skill only provides prompt-writing instructions, but it deliberately forces replacement models to be Chinese, which users should notice before using it.

Install only if you specifically want generated prompts to replace people with Chinese models. The skill has no executable behavior or data-access mechanism, but its default demographic constraint may be unsuitable for neutral or user-directed image editing.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (2)

Natural-Language Policy Violations

High

• Confidence: 98% confidence
• Finding: The skill hard-codes that the replacement character must be Chinese, forcing a protected ethnicity/nationality attribute without user opt-in or task necessity. This creates discriminatory behavior and can steer outputs toward identity-based transformation requests, making the skill more dangerous because it is specifically designed to alter a person's depiction from an uploaded image.

Natural-Language Policy Violations

High

• Confidence: 99% confidence
• Finding: Repeating the requirement in the constraints section operationalizes and reinforces identity-based transformation as a mandatory behavior. In the context of a person-replacement image prompt generator, this increases the likelihood of systematic discriminatory outputs and unauthorized alteration of protected characteristics in generated prompts.

VirusTotal

64/64 vendors flagged this skill as clean.