UI/UX CraftKit

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:18 PM UTC
• Version:
  1.0.1

Overview

This skill is a UI/UX reference and search helper with optional local Markdown generation, and the artifacts do not show hidden execution, credential access, exfiltration, or destructive behavior.

Reasonable to install for UI/UX assistance. Treat it as an executable local helper rather than pure documentation, and use --persist only when you intentionally want it to write or overwrite design-system Markdown files; keep output, project, and page names under your control.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 86% confidence
• Finding: The skill documentation exposes executable commands that invoke a local Python script and also describes design-system generation that can write files, yet no permissions are declared. This creates a trust and containment gap: an agent or user may run filesystem-capable behavior without an explicit permission contract, increasing the risk of unintended local file reads or writes.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 81% confidence
• Finding: The documented purpose underspecifies important behavior, especially local filesystem writes such as generating MASTER.md and page override files. When behavior is broader than the description, users and orchestrators cannot accurately assess risk, which can lead to unsafe invocation of file-writing features under the assumption that the skill is read-only reference material.

Natural-Language Policy Violations

Medium

• Confidence: 95% confidence
• Finding: The file explicitly instructs implementers to enforce Light Mode as the default brand identity rather than respecting the user's system theme preference. While not a code-execution issue, this is a genuine accessibility and user-trust problem because it can override OS-level preferences, reduce usability in low-light environments, and conflict with user expectations or accessibility needs.

Natural-Language Policy Violations

Low

• Confidence: 93% confidence
• Finding: The CSV includes a guideline entry whose usage text is written in Chinese and explicitly instructs behavior when selecting fallback icons. In an agent skill, this can steer model output into a language the user did not request, causing prompt/UX misalignment and reducing reliability, though it does not appear to create direct code-execution or data-exfiltration risk.

Natural-Language Policy Violations

Medium

• Confidence: 88% confidence
• Finding: Row 118 prescribes file-type color coding with English-centric labels such as 'PDF orange, doc blue, image purple', which can bias the UI guidance toward English naming conventions and may not localize cleanly for non-English users. In a UI/UX design skill intended for broad reuse across products and locales, this can propagate exclusionary or non-compliant internationalization patterns into downstream applications.

Natural-Language Policy Violations

Medium

• Confidence: 96% confidence
• Finding: Row 116 explicitly recommends 'language flag accent' for a translator app, which is a known problematic pattern because flags represent countries, not languages, and can misidentify or exclude users. In this skill's context, the file is reusable design guidance across many stacks, so embedding this pattern increases the chance that generated UIs will violate localization policy or create misleading language selection experiences.

VirusTotal

64/64 vendors flagged this skill as clean.