ClawHub
Back to skill

Security audit

Anthropic Frontend Design

Security checks across malware telemetry and agentic risk

Overview

This skill is a coherent frontend design helper that reads bundled design reference data and can optionally write generated design-system markdown files.

Install this if you want a frontend design assistant that uses bundled local reference data. Be careful when using --persist or --output-dir: keep writes inside your project or a dedicated folder, and review generated MASTER.md/page files before relying on them.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
80% confidence
Finding
The skill invokes a local Python script and static analysis indicates file read/write capabilities, but the skill declares no permissions. Undeclared filesystem access is dangerous because it can surprise operators, bypass expected trust boundaries, and allow persistent modification of local files if the surrounding agent executes the referenced tooling.

Tp4

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding
The skill is presented as design guidance, but its referenced behavior includes local dataset search and persistent file generation such as MASTER.md and page override files. This mismatch is dangerous because users may approve or run the skill expecting advisory output only, while it can actually read local resources and alter the filesystem, increasing the chance of unauthorized persistence or unintended overwrites.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The persistence path writes files to disk whenever `persist=True` without any built-in confirmation, dry-run, or explicit safety guard in the write path. In an agent/tooling context, this can lead to unintended filesystem modification, cluttering project directories or overwriting expected design-system artifacts if upstream inputs trigger persistence implicitly.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.