writER

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:34 PM UTC
• Version:
  1.0.1

Overview

This is an instruction-only writing style skill with no evidence of hidden code, credential access, data exfiltration, or system control.

Install only if you want this ER-style poetic rewrite behavior to be available to your agent. Be aware that broad Chinese writing prompts may activate it unintentionally, and avoid pasting highly sensitive personal text unless you are comfortable with your agent and model provider processing it.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• 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

Findings (2)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The trigger keywords include broad, generic phrases such as '帮我写散文', '写成文章', and '改成我的风格', which can match many ordinary writing requests and cause the skill to activate unintentionally. This can override user intent, apply an unwanted persona/style transformation, and increase prompt-routing ambiguity in systems that auto-invoke skills from loose keyword matches.

Natural-Language Policy Violations

Medium

• Confidence: 84% confidence
• Finding: The skill metadata and content strongly constrain behavior to Chinese-language poetic rewriting without stating any fallback for user language preference. If auto-selected, it may force Chinese output or a Chinese stylistic mode even when the user's request is in another language, causing integrity issues in output and reducing user control over the agent's behavior.

VirusTotal

64/64 vendors flagged this skill as clean.