Security audit

policy-to-action

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent policy-analysis report generator; it asks for relevant user-provided context and uses web research, with no hidden code execution or credential access found.

Install only if you want a Chinese-language policy analysis workflow. When using it, provide anonymized or minimal subject details where possible, verify generated policy/tax/legal conclusions against official sources, and review any HTML report before sharing it externally.

SkillSpector

By NVIDIA

Vulnerability Patterns

Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
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

Findings (6)

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The trigger example is very broad: '帮我解读政策 [政策名称]，我是 [企业/个人]，[行业]，[资质]'. In a shared assistant environment, this overlaps with normal user requests about policies and could cause the skill to activate unintentionally, leading to unexpected handling of user input or overshadowing other more appropriate skills.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger pattern is broad enough to match ordinary workplace or policy-related conversation, which can cause the skill to activate without clear user intent. Unintended activation increases the chance of unnecessary data collection, unsolicited web access, and generation of misleading authoritative-looking reports in contexts where the user did not ask for this workflow.

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The markdown states that any mention of listed keywords should automatically activate the skill, but does not define scope boundaries or require confirmation. In this skill's context, that is risky because activation leads directly into structured profiling and external research steps, which may process sensitive organizational or personal information without a deliberate opt-in.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill explicitly asks for identifying information about a company or individual, plus industry and qualification details, without any privacy notice, minimization guidance, or sensitivity handling. Because the skill also instructs downstream web research and report generation, unnecessary collection of real names and organizational details could expose sensitive business posture or personal data beyond what is needed for analysis.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The trigger phrase '对我们有什么影响' is broad and conversational, so it can activate on many ordinary discussions rather than an explicit request to use this skill. Over-broad routing can cause unintended invocation, incorrect context handling, and accidental processing of sensitive business or personal policy questions under this skill when the user did not clearly opt in.

Natural-Language Policy Violations

Medium

Confidence: 82% confidence
Finding: The skill metadata is entirely Chinese-centric and presents the capability as Chinese-language by default, which can cause the system to route users into a language mode they did not request. While this is primarily a usability and consent issue rather than a classic security flaw, forced language behavior can lead to misunderstanding of policy advice and reduce informed user control over outputs.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.