ClawHub
Back to skill

Security audit

卖家之家跨境电商物流海外仓搜索

Security checks across malware telemetry and agentic risk

Overview

This search-only skill is not malicious, but it asks for an API key and documents authenticated private messaging that goes beyond simple public logistics lookup.

Install only if you are comfortable with a logistics search skill that may also guide an agent toward authenticated provider outreach. For search-only use, avoid configuring MJZJ_API_KEY; before any private message is sent, require the agent to show the recipient and exact message content and get explicit approval.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
90% confidence
Finding
The trigger guidance is broad enough to activate this skill for generic logistics-related conversations, which can steer the agent into external API use without a clearly bounded user request. That increases the chance of unnecessary third-party data transmission and unintended workflow selection, especially because the skill also links search results to downstream messaging behavior.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
This section enables taking a returned userSlug and sending a private message through an authenticated endpoint without requiring explicit user confirmation or warning that a third party will be contacted using the configured API key. That creates a user-impacting action path that could cause unauthorized outreach, spam, or accidental disclosure of user-provided content to external recipients.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The reusable prompt text instructs the agent to directly send private messages via an authenticated API after search, but omits safeguards for consent, review, or notice about external side effects. Because this is phrased as reusable operational guidance, it increases the likelihood that downstream agents will perform outbound messaging automatically.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal