观势深度洞察引擎

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed Chinese-language analysis methodology skill with no code, dependencies, persistence, or data access behavior.

Install this if you want a Chinese-language strategic review framework that challenges existing conclusions. Be aware it may activate on broad requests like "think again" or "go deeper," so users who prefer narrow, explicit skill invocation should configure routing accordingly.

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 (2)

Vague Triggers

Medium

Confidence: 93% confidence
Finding: The trigger conditions include very broad phrases such as asking to 'go deeper', 'think again', or 'what else was not considered', which overlap with common conversational language and can cause the skill to activate unintentionally. Because this skill is a methodology layer rather than a narrowly scoped tool, accidental invocation can redirect normal conversations into an unsolicited analysis workflow, creating prompt-routing confusion and reducing user control over agent behavior.

Natural-Language Policy Violations

Medium

Confidence: 88% confidence
Finding: The skill metadata and body are written to operate in Chinese without offering a user-language choice or documenting a necessary locale restriction. This can cause inaccessible or unexpected behavior for non-Chinese users, and in multi-agent routing contexts may lead to responses in the wrong language, misunderstanding of outputs, or silent degradation of review quality.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal