ClawHub
Back to skill

Security audit

酷家乐设计-专业版

Security checks across malware telemetry and agentic risk

Overview

This skill mostly matches its interior-design purpose, but it asks for broad command/file authority and enterprise token use with unclear activation rules and unsafe credential examples.

Review this before installing if you will use real Kujiale enterprise credentials. Keep tokens out of repo-tracked config files, require explicit confirmation before running referenced scripts or approving/submitting team changes, and use restricted output directories for generated renders and reports.

SkillSpector

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

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The skill first instructs users to place a long-lived `access_token` directly into `.kjlconfig.json`, then later advises using environment variables or a secret manager instead. This contradiction can lead users to persist sensitive credentials in local files that may be committed, copied, or exposed through workspace access, logs, backups, or shared project archives.

Vague Triggers

Medium
Confidence
82% confidence
Finding
The trigger condition section is malformed and vague, so an agent may not know when the skill should or should not activate. Ambiguous activation criteria increase the chance of inappropriate execution of `exec`/`write`-capable workflows in unrelated contexts, which can cause unintended file changes, external calls, or handling of sensitive inputs without clear user intent.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill documents export, save, and configuration-writing behavior while exposing `read`, `write`, and `exec` tools, but it does not place a clear warning or consent checkpoint next to those operations. In an agent setting, this can normalize state-changing actions and lead to accidental overwrites, unsafe file creation, or persistence of sensitive data without the user fully understanding the consequences.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.