Psychology Master

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed psychology, learning, UX, and ethical marketing guidance skill with no evidence of hidden execution, persistence, credential access, or data exfiltration.

Install this as an educational and optimization aid, not as a license to manipulate users. Pay particular attention when using it for marketing, minors, vulnerable groups, or personality-based segmentation: require consent, minimize data, avoid dark patterns, keep claims truthful, and get legal or professional review for regulated or high-stakes contexts.

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 (3)

Vague Triggers

Medium

Confidence: 71% confidence
Finding: Several trigger phrases such as conversion optimization, persuasive messaging, and customer decision-making are broad enough to activate on common user requests without clear boundaries. Over-broad activation can cause the skill to be selected in unintended contexts, leading to unnecessary exposure to marketing-persuasion behaviors and possible misuse in sensitive or manipulative scenarios.

Natural-Language Policy Violations

Medium

Confidence: 91% confidence
Finding: The content explicitly maps cognitive biases to marketing and UX tactics such as setting defaults, leveraging loss aversion, social proof, and scarcity, with only minimal ethical qualification. In a skill positioned for optimization, persuasion, and conversion, this can enable manipulative dark-pattern design that nudges users against their interests or informed consent.

Missing User Warnings

Medium

Confidence: 91% confidence
Finding: The section explicitly recommends deriving personality traits from behavioral data such as language use, purchase history, engagement patterns, and social media activity, which are all forms of profiling-sensitive data. Although it includes a brief caution against manipulation, it does not warn about consent, privacy compliance, accuracy limits, bias, or the risks of covert psychological profiling for persuasion and segmentation, making misuse more likely in a marketing-oriented skill.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal