ClawHub

闲鱼自动上架助手

Security checks across malware telemetry and agentic risk

Overview

This is a text-only Xianyu listing helper that overstates automatic posting but does not include code or permissions to post anything.

Install only with the understanding that this skill helps draft Xianyu listing content and does not actually publish listings. Review all generated prices, warranty language, authenticity claims, return promises, discounts, and contact/payment details before posting manually.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Intent-Code Divergence

Medium
Confidence
91% confidence
Finding
The skill metadata and description advertise '自动发布商品到闲鱼' (automatic posting), but the FAQ states the tool currently only generates optimized content and requires manual posting. This mismatch can mislead users or downstream agents into granting higher trust, broader permissions, or expecting automation capabilities the skill does not actually have, increasing the chance of unsafe integration decisions.

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
Claiming automatic posting in the manifest when the documented behavior is limited to title/description/price generation is a security-relevant misrepresentation of scope. In agent ecosystems, manifest claims may influence tool routing, user consent, or permission assumptions, so overstating action capability can cause inappropriate reliance or unsafe operational expectations.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal