ClawHub
Back to skill

Security audit

Dlazy Logo Branding System

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed dLazy logo and branding workflow that uses a third-party CLI and cloud API, with notable but purpose-aligned credential and upload behavior.

Install only if you are comfortable using dLazy's cloud service. Prompts and any referenced local media files may leave your machine, and the CLI can save an API key under ~/.dlazy/config.json; prefer npx or an environment variable if you do not want a global install or persistent key.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The document minimizes risk by saying the skill itself will not access the network or file system, but it also instructs the agent to use a CLI that transmits prompts to remote APIs and uploads local media files. That mismatch can mislead users and downstream agents about data handling and trust boundaries, increasing the chance that sensitive local assets are sent off-host without fully informed consent.

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill mandates terminal command execution for a design workflow, including authenticated CLI use and remote generation, which grants the agent a stronger execution capability than the high-level branding purpose alone implies. In an agent environment, normalizing command execution expands the attack surface: prompts, local asset paths, credentials, and generated outputs can all be routed through a toolchain that the user may not have expected to authorize.

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger phrase is broad and can cause this skill to activate in loosely related conversations about logos or branding without clear user intent to use an external CLI-backed workflow. Because the skill instructs an agent to install and use a third-party tool and send prompts/assets to remote API endpoints, over-broad triggering increases the chance of unintended tool use, data disclosure, or workflow hijacking.

Natural-Language Policy Violations

Medium
Confidence
95% confidence
Finding
These instructions force a specific interaction and execution model, including mandatory terminal command use and Chinese-language operational guidance, without user opt-in. That is dangerous because it can override normal agent safety and consent expectations, pushing the agent toward executing third-party CLI commands and remote API operations even when the user did not request that mode or language.

VirusTotal

1/64 vendors flagged this skill as malicious, and 63/64 flagged it as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.