Security audit

Psyvector Pv16

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small persona-style guidance prompt with broad and assertive wording, but it does not include code, hidden behavior, credential access, persistence, or data movement.

Install only if you want an assertive advice persona. For legal, medical, financial, employment, or other high-impact decisions, treat its guidance as opinion and ask for uncertainty, alternatives, and risks explicitly.

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

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: 95% confidence
Finding: The description "Expert guidance" is extremely broad and does not constrain the situations in which the skill should activate. In an agent ecosystem, vague activation criteria can cause the skill to be invoked in inappropriate or higher-risk contexts, increasing the chance that dominant or unsafe guidance is applied where it should not be.

Vague Triggers

Medium

Confidence: 95% confidence
Finding: The persona language presents the skill as an authority offering optimal solutions and strong decision-making without clear activation limits. Combined with the "controlling/core" persona, this can bias an agent toward overconfident, unilateral advice in contexts requiring caution, user consent, or domain-specific safeguards.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: Setting `risk_reminder` to `False` suppresses safety signaling while the skill simultaneously adopts a forceful, low-caution advisory posture. This increases the likelihood that users receive assertive recommendations without warnings about uncertainty, tradeoffs, or potential harm, especially in sensitive decision contexts.

Static analysis

No suspicious patterns detected.