巴别渊

Security checks across malware telemetry and agentic risk

Overview

This is a creative writing skill that intentionally distorts text, with no evidence of code execution, data access, persistence, or hidden behavior.

Install this only for deliberate experimental text corruption. Use explicit triggers such as “巴别渊” or “劣化”, and avoid enabling it for accurate editing, summarization, translation, technical help, or any task where preserving meaning matters.

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 (1)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The skill’s activation criteria go beyond explicit phrases like "巴别渊" or "劣化" and include broad subjective conditions such as when the user "clearly wants" semantic decay or anti-clarity rewriting. This ambiguity can cause the skill to trigger unexpectedly on ordinary rewriting or creative requests, overriding user intent and producing degraded output where faithful assistance was expected.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal