Perfect

Security checks across malware telemetry and agentic risk

Overview

This skill is a fiction-based mental-health support tool, but it is under-scoped for suicide, self-harm, and eating-disorder situations.

Review carefully before installing. The skill may be useful for discussing themes in Perfect, but users should not rely on it for active suicide risk, self-harm, eating disorders, or other urgent mental-health situations unless it is revised to direct people to emergency services, crisis lines, trusted adults, and qualified professionals.

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 (6)

Vague Triggers

Medium

Confidence: 92% confidence
Finding: The trigger list is broad enough to activate on common emotional statements such as 'I'm not good enough' or 'I can't keep pretending,' which may cause the skill to intervene in situations where the user did not ask for this book-specific framework. In a mental-health-adjacent skill, unintended activation is risky because it can derail more appropriate safety handling, especially for users expressing distress or crisis-related language.

Missing User Warnings

High

Confidence: 98% confidence
Finding: This skill explicitly covers eating disorders, self-harm-adjacent behavior, and suicide-prevention awareness, yet it provides no explicit crisis disclaimer, emergency escalation guidance, or instruction to prioritize professional help. That omission is dangerous because users may treat the skill as sufficient support during acute mental-health situations, delaying contact with emergency services, crisis hotlines, or trusted adults.

Missing User Warnings

High

Confidence: 99% confidence
Finding: The Quick Start includes a suicide-related example ('Someone I know tried to die') but presents it as a normal onboarding prompt without any urgent-help framing or safety instructions. Because this guide must be shown proactively on first load, it can normalize crisis disclosure without simultaneously directing users to immediate support, increasing the chance of unsafe reliance on the skill.

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The use-case text explicitly targets eating disorders, coming out, and suicide concerns using broad support-oriented language rather than narrow literary discussion. That makes the skill likely to activate in real crisis or mental-health situations, where a fiction-themed toolkit may provide inappropriate or insufficient guidance and could displace safer crisis-specific handling.

Vague Triggers

Medium

Confidence: 84% confidence
Finding: The description is broad enough to overlap with general mental-health counseling, identity crises, and suicide-prevention conversations, despite the skill being based on a verse novel. In a sensitive domain, ambiguous scope increases the chance that users in distress receive literary or unqualified guidance instead of appropriate safety-focused support.

Missing User Warnings

Medium

Confidence: 92% confidence
Finding: This section contains explicit discussion of eating disorders, steroid misuse, and a suicide attempt response scenario without any content warning, crisis framing, or safety guidance. In a mental-health-adjacent skill likely to be triggered by vulnerable users, abrupt exposure to this material can worsen distress, normalize harmful behaviors, or provide unsafe guidance by omission.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal