Security audit

[Nyx Archive] Art Philosophy

Security checks for vulnerabilities and agentic risk

Overview

This art-coaching skill is not malicious, but it asks the agent to continuously infer and remember personal creative preferences and self-evaluation patterns without clear limits or user controls.

Install only if you are comfortable with the agent building an ongoing profile of your art preferences, creative motivations, critique style, and decision patterns from art-related conversations. Prefer using explicit commands and ask the agent not to retain or reuse inferred preferences if you do not want that profiling behavior.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

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

Findings (3)

Missing User Warnings

High

Confidence: 97% confidence
Finding: This section explicitly says the skill 'observes how you talk about art, what you create, what you respond to' and adapts to the user's values, philosophy, and behavior, which amounts to ongoing profiling. Because there is no warning, consent flow, retention policy, or sensitivity boundary, the skill normalizes collection of preference, psychological, and self-evaluation data that could be misused or exposed.

Missing User Warnings

High

Confidence: 98% confidence
Finding: The 'Passive Learning (Always On)' section instructs continuous inference of visual preferences, aesthetic values, creative philosophy, and decision patterns, including how the user evaluates themselves and handles blocks. These are sensitive behavioral inferences, and collecting them continuously without prominent disclosure or bounded use creates substantial privacy, manipulation, and over-collection risk.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The skill defines a vague passive trigger ('When art comes up in conversation') combined with 'Always On' behavior, which creates unclear activation boundaries and encourages persistent background monitoring. Ambiguous scope makes it easy for the agent to collect and infer user information outside a narrowly consented interaction, increasing privacy and behavioral-profiling risk.

Static analysis

No suspicious patterns detected.