UI/UX Pro Max

• Outcome: Pass
• Latest audit: Jul 10, 2026 · 2:05 AM UTC
• Version: 0.1.0

Overview

This skill provides UI/UX guidance and an optional local design-system file generator, with no evidence of exfiltration, hidden persistence, or malicious behavior.

Install only if you want a broad UI/UX assistant that may consult bundled reference data and, when asked to persist a design system, create markdown files in your workspace. Review commands using --persist, --output-dir, project names, or page names before running them.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands

Findings (7)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill instructs the agent to read bundled files and optionally run a local Python script, which implies file-read capability and potential file-write side effects, yet no permissions are declared. This creates a trust and containment gap: a host may allow behavior the user did not clearly consent to, and the skill’s broad UI/code scope increases the chance those capabilities are exercised during normal use.

Description-Behavior Mismatch

Medium

• Confidence: 95% confidence
• Finding: The skill can write persistent files and create directories in the local workspace, which exceeds pure advisory behavior and introduces side effects on the host filesystem. If exposed through an agent without strict consent and path controls, a user or prompt could cause unintended artifact creation, workspace pollution, or overwriting of expected project documentation.

Context-Inappropriate Capability

Medium

• Confidence: 92% confidence
• Finding: Local file-generation capability is present in a skill described primarily as UI/UX guidance, so the implementation has broader authority than the stated purpose implies. In agent settings, this mismatch is risky because downstream components may invoke the skill expecting read-only guidance while it can materially alter the workspace.

Vague Triggers

Medium

• Confidence: 84% confidence
• Finding: The activation description is very broad, covering UI design, UX, design systems, accessibility, critique, and concrete frontend code generation across multiple frameworks. Overly broad routing can cause the skill to trigger on many generic coding or product requests, increasing the chance it accesses local assets/scripts unexpectedly or influences tasks outside its narrow intended scope.

Vague Triggers

Medium

• Confidence: 93% confidence
• Finding: The README states that the skill auto-activates for broadly phrased UI/UX requests such as 'build', 'design', 'create', 'implement', 'review', 'fix', and 'improve'. In an agent environment, this wide trigger surface can cause the skill to engage on many ordinary development tasks, expanding the chance that unreviewed prompts, instructions, or generated code influence unrelated workflows.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The workflow activates on very broad phrases like 'design, build, create, implement, review, fix, improve,' which can cause the skill to trigger for ordinary frontend requests without clear user intent to use this specific workflow. Over-broad activation increases the chance of unintended tool/script execution paths and can steer conversations into unnecessary command usage or filesystem persistence behavior.

Missing User Warnings

Low

• Confidence: 87% confidence
• Finding: The write site creates and saves files without any user-facing warning at the moment of persistence, reducing transparency around side effects. This can lead to surprising state changes and makes abuse or accidental misuse harder for users and orchestrators to detect in time.

VirusTotal

64/64 vendors flagged this skill as clean.

Static analysis

No suspicious patterns detected.