Security audit

Learn-X

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed tutoring workflow skill with routing and language-fit caveats, but no hidden execution, data access, or destructive behavior.

Install this if you want a strict coaching-style tutor that asks questions before explaining. Avoid enabling it as a default for users who mostly need direct answers, specialized medical/legal/financial guidance, debugging, or English-only instruction unless the host agent can route around those cases.

SkillSpector

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 (3)

Vague Triggers

High

Confidence: 95% confidence
Finding: The trigger is extremely broad ('teach me', 'help me learn', 'understand', 'get good at') and explicitly domain-agnostic, so it can activate on a wide range of ordinary user requests. That creates a real routing/control vulnerability: the skill may intercept unrelated or higher-risk topics and force this rigid workflow instead of allowing a more appropriate specialized skill or safety policy to handle the request.

Natural-Language Policy Violations

Medium

Confidence: 84% confidence
Finding: The skill content is written in a specific Chinese/Chinese-English mixed instructional style and does not indicate adaptation to the user's language or locale preferences. While not directly a security exploit, this can cause misalignment, confusion, or reduced comprehension, especially because the skill is broadly triggered and may therefore affect many users unintentionally.

Natural-Language Policy Violations

Medium

Confidence: 95% confidence
Finding: The file is entirely in Chinese while the skill is domain-general and may be triggered for users who do not read Chinese. That creates a prompt-quality and safety issue: users may be routed into an instructional flow they cannot understand, causing them to miss constraints, misunderstand guidance, or receive unusable learning support. In this context, the skill is broadly applicable across programming, tools, and other topics, so lack of language choice increases the chance of silent failure across a wide audience.

Static analysis

No suspicious patterns detected.