pencil-design-skill

• Outcome: Pass
• Latest audit: May 28, 2026, 9:02 AM
• Version:
  3.2.0

Overview

This is a Pencil design workflow skill with disclosed usability caveats, not evidence of hidden access or harmful behavior.

Install if you want a Pencil-focused design assistant and are comfortable with it writing .pen files or editing Pencil documents through MCP. For English or multilingual projects, explicitly state the desired UI language, and consider narrowing activation if you do not want it active for general UI prototyping or design-to-code requests.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (3)

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The manifest description is broad enough to trigger on generic UI, prototyping, or design-to-code requests, which can cause this skill to activate outside narrowly intended Pencil-specific tasks. Over-broad activation increases the chance that restrictive instructions in the skill override user preferences or influence unrelated work, creating unnecessary prompt-scope and policy risk.

Natural-Language Policy Violations

High

• Confidence: 97% confidence
• Finding: The skill mandates Simplified Chinese for all user-facing text unless the user explicitly overrides it, which improperly substitutes skill defaults for user intent. In a broadly triggered design skill, this can silently alter deliverables, degrade usability for non-Chinese audiences, and cause incorrect or harmful output in customer-facing, regulated, or accessibility-sensitive contexts.

Natural-Language Policy Violations

Medium

• Confidence: 90% confidence
• Finding: The 'Common Mistakes' section reinforces the same mandatory Simplified Chinese behavior, making the unsafe default operationally sticky and more likely to be followed by the agent. Repetition across the skill increases the likelihood of persistent language override even when user expectations or surrounding context indicate otherwise.

VirusTotal

VirusTotal findings are pending for this skill version.