Variant Design Skill

• Outcome: Pass
• Latest audit: Sep 12, 2026 · 1:58 AM UTC
• Version: 0.1.0

Overview

This design-helper skill is benign: it contains Markdown guidance for generating UI variations and does not add hidden execution, credential access, or persistence beyond disclosed design context notes.

Install this when you want a design-generation assistant. Be aware it may read nearby project files to infer brand and UI context, and it may add design-context comments to generated files; review generated code and external font imports before using output in production.

A.I.G

Vulnerability Patterns

• Skill Instruction Hijacking: Alters the agent's session goals or safety constraints when the skill loads
• Agent Memory Poisoning: Writes attacker-controlled rules into memory that affect later sessions
• Remote Payload Retrieval and Execution: Fetches external code whose behavior can change after review
• Embedded Malicious Code: Ships malicious scripts inside the skill and executes them locally
• Unauthorized Access and Privilege Escalation: Obtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (10)

Vague Triggers

Medium

• Confidence: 94% confidence
• Finding: The usage documentation explains what kinds of prompts the skill handles but does not define when it should abstain. Without negative boundaries, an orchestrator may over-apply the skill to any design-related request, increasing prompt-injection surface and causing unintended system-prompt influence over unrelated conversations or tasks.

Vague Triggers

Medium

• Confidence: 91% confidence
• Finding: The example invocations are ordinary natural-language requests like 'design a dashboard' that are indistinguishable from normal conversation. In chat environments where skills may auto-activate from prompt matching, this can cause unintentional invocation, leading the skill to take over responses, load additional prompt context, or steer output when the user did not explicitly request the skill.

Vague Triggers

Medium

• Confidence: 95% confidence
• Finding: The skill advertises very broad trigger phrases such as generic requests for design help and variation commands, which can cause the skill to activate in contexts the user did not specifically intend. That creates a prompt-routing and least-surprise problem: unrelated conversations may be pulled into this skill, leading to inappropriate file scanning, context persistence, or generation behavior that the user did not request.

Ae4

Medium

• Category: analysis-evasion
• Confidence: 80% confidence
• Finding: Suspicious Unicode normalization or mixed-script content

Natural-Language Policy Violations

Medium

• Confidence: 87% confidence
• Finding: This markdown file includes a prompt explicitly specifying a Chinese wellness e-commerce app with Chinese-language UI text, but it does not indicate that language selection is user-driven or optional. Under the policy for natural-language violations, forcing a specific language or locale without opt-in should be flagged unless the regional constraint is clearly justified.

Natural-Language Policy Violations

Medium

• Confidence: 84% confidence
• Finding: The layout pattern is labeled as a Chinese-style wellness app and hardcodes Chinese UI labels throughout the example. While regional examples can be acceptable, this section does not explicitly frame the locale as optional, selected by the user, or restricted to a China-specific deployment context.

Natural-Language Policy Violations

Medium

• Confidence: 90% confidence
• Finding: This example repeatedly prescribes Chinese-language labels and a culturally specific product framing, but does not say that such localization should be chosen based on the user's language preference or the intended market. That creates a natural-language policy issue because the skill content appears to mandate a locale rather than offer it as an option.

Natural-Language Policy Violations

Medium

• Confidence: 91% confidence
• Finding: The prompt specifies that Japanese text must be dominant and larger than the English translation, and the accompanying explanation frames this hierarchy as a required design commitment. Because this markdown does not offer a language choice or explain a region-specific need, it imposes a specific language/locale preference in a way that can conflict with organizational language-choice policy.

Natural-Language Policy Violations

Low

• Confidence: 75% confidence
• Finding: The variation actions are defined only in English terms such as 'Vary strong,' 'Polish,' and 'Critique,' which implies a preferred command language. While the README is bilingual, the operational command vocabulary is not presented as a user choice, which can conflict with language-choice policy expectations.

Natural-Language Policy Violations

Low

• Confidence: 82% confidence
• Finding: The scenario-matching table includes Chinese trigger terms like "后台," "官网," "应用," and "界面," but the skill does not explain its language-detection or locale behavior to the user. This creates an implicit language policy choice without an explicit opt-in or documented justification.

Static analysis

No suspicious patterns detected.