ClawHub
Back to skill

Security audit

Frontend Design Extractor

Security checks across malware telemetry and agentic risk

Overview

This skill coherently helps an agent document and optionally apply UI-only design-system refactors, with no evidence of hidden credential access, persistence, exfiltration, or destructive behavior.

Install this when you want an agent to inspect frontend code and create UI/UX documentation, or to make explicitly approved UI-only refactors. Before refactor use, review the proposed plan and diffs carefully. Because the package references helper scripts that are not included, do not let an agent run similarly named project-local scripts unless you have inspected them or explicitly intended that behavior.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Description-Behavior Mismatch

High
Confidence
95% confidence
Finding
The skill’s declared purpose is extraction and documentation of reusable UI/UX patterns, but this section expands scope into planning and executing refactors and code edits on a target project. That creates a capability mismatch where an analysis-oriented skill can be used to modify code, increasing the risk of unauthorized changes, hidden side effects, or abuse of the agent to perform broader repository modifications than the user expects.

Description-Behavior Mismatch

Medium
Confidence
88% confidence
Finding
The quick-start workflow directs the agent to create output folders and write generated artifacts, which exceeds a strictly analysis-only role and introduces filesystem write behavior. While lower risk than code refactoring, this still enables unrequested repository changes, clutter, or overwriting of files if paths are mis-scoped or a user assumes the skill is read-only.

Context-Inappropriate Capability

High
Confidence
96% confidence
Finding
These instructions explicitly tell the agent to proceed from planning into phased edits against the target project, despite the skill being described as a UI/UX extractor. Granting code-change authority without a justified need violates least privilege and can be exploited to alter application behavior, introduce defects, or mask broader changes under the label of 'UI/UX alignment.'

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.