Security audit

品牌破圈五密码方法论

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only brand strategy skill with no code execution, credential use, or data access beyond its stated branding advice purpose.

Install if you want a Chinese-language brand breakthrough framework for marketing analysis. Be aware it may activate on broad brand-analysis requests and frame answers through this specific five-part methodology, so use a more specific prompt if you want a different analysis style.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (2)

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger conditions are broad enough to match common brand, marketing, and analysis requests, which can cause this skill to activate in situations where the user did not specifically ask for this framework. That creates prompt-squatting risk: the assistant may be steered into this skill’s prescribed methodology and outputs over more appropriate general-purpose reasoning or other domain skills.

Natural-Language Policy Violations

Medium

Confidence: 89% confidence
Finding: The output requirement '说人话：能用客户听得懂的语言' appears alongside a fully Chinese skill and effectively biases responses toward a fixed language/register without checking the user’s preference. In a multilingual assistant, forcing a language or style can override user intent, reduce usability, and in some cases cause the wrong skill behavior to dominate cross-language interactions.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.