Security audit

五运灵名

Security checks across malware telemetry and agentic risk

Overview

This is a text-only Chinese naming and fortune-telling skill with no code execution, persistence, credential use, or data exfiltration, though it asks for personal birth details and includes cultural health/fate claims.

Install only if you want a Chinese traditional-metaphysics naming assistant. Share the least precise birth and family information that still satisfies your request, avoid relying on its health or fate statements for real decisions, and explicitly ask for neutral or modern naming preferences if you do not want gendered traditional conventions applied.

SkillSpector

By NVIDIA

Vulnerability Patterns

Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (7)

Vague Triggers

Medium

Confidence: 86% confidence
Finding: The skill description is broad enough to trigger on ordinary conversations about names, renaming, or fortune-telling, which can cause the skill to activate outside the user's true intent. Over-broad activation increases the chance of inappropriate instruction takeover, unwanted persona switching, and misleading deterministic advice in unrelated contexts.

Natural-Language Policy Violations

Medium

Confidence: 82% confidence
Finding: The skill hard-forces Chinese output and a rigid response format without considering the user's language preference, which can override the surrounding assistant policy and reduce user comprehension. In security terms, this is a prompt-control issue: a skill that unconditionally dictates output behavior can interfere with higher-priority instructions and cause unsafe or unusable responses for users who do not read Chinese.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The description is very broad and only states what the skill can do, not when it should or should not activate. In an agent environment, ambiguous activation boundaries can cause the model to invoke this skill for loosely related requests, leading to unnecessary collection or processing of sensitive personal data such as birth details, names, and identity-related information, and can also cause capability overreach into adjacent advice domains.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The handbook instructs collection of full birth date/time, location, sex, and name data to perform analysis, but provides no minimization, consent, retention, or handling guidance. This creates a real privacy and safety risk because birth details combined with location are sensitive personal data and can be unnecessarily exposed, stored, or reused by downstream systems or operators.

Natural-Language Policy Violations

Medium

Confidence: 87% confidence
Finding: The document hard-codes gender-specific naming expectations such as 'boys should be strong' and 'girls should be gentle' without presenting them as optional user preferences. This can lead to discriminatory or exclusionary outputs, especially for users who want neutral naming styles or whose identity does not fit the binary assumptions embedded in the guidance.

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The FAQ and checklist reinforce mandatory gendered conventions and impose extra restrictions on female names, such as special avoidance rules, without user choice or justification. In skill context, this is more dangerous because it operationalizes biased policy into repeatable outputs, increasing the chance of systematic discriminatory recommendations to many users.

Missing User Warnings

Medium

Confidence: 90% confidence
Finding: The file gives health-related advice and deterministic life-outcome claims tied to birth timing, including health warnings and personality/fate assertions, without any disclaimer that the content is non-medical, non-scientific, and not a substitute for professional advice. In the context of a baby-naming skill, this can mislead users into treating superstition as actionable health or life guidance, especially for decisions affecting infants or parental anxiety.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.