Security audit

Uiverse Design

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed UI component reference skill, with an optional remote library download and one insecure example users should review before reuse.

Safe to install for the bundled offline examples. Review any copied form code before production use, especially password fields, and do not run scripts/setup.sh unless you are comfortable replacing the reviewed subset with live, unpinned upstream GitHub content.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The documentation instructs users to run a setup script that downloads external content at runtime, expanding the skill beyond a static local reference library. This is risky because the fetched content is explicitly unreviewed and mutable, so the skill's effective contents can change after distribution and introduce malicious or unsafe files through a supply-chain path.

Context-Inappropriate Capability

Medium

Confidence: 91% confidence
Finding: Allowing the curated local dataset to be replaced by unreviewed external GitHub content undermines the trust model of the skill. An attacker controlling the upstream source, a compromised dependency path, or even normal upstream changes could alter component files and any downstream scripts or examples, leading to delivery of malicious HTML/CSS or other unexpected content to users.

Missing User Warnings

Medium

Confidence: 98% confidence
Finding: The login password field uses type="text", which displays entered credentials in clear view instead of masking them. This can expose sensitive information through shoulder-surfing, screen sharing, recordings, or screenshots; in a reusable UI component library, consumers may copy this pattern directly into production interfaces.

Missing User Warnings

Medium

Confidence: 99% confidence
Finding: The signup password and confirm-password fields are also implemented as type="text", causing new credentials to be visibly exposed during entry. This increases the risk of credential disclosure to nearby observers and makes the component an unsafe example in a design library where developers may reuse it without noticing the issue.

Static analysis

No suspicious patterns detected.