ClawHub
Back to skill

Security audit

浙江供应商资质业绩核验

Security checks for vulnerabilities and agentic risk

Overview

This skill coherently runs a supplier-check script and uses a disclosed local credential file to call its service.

Before installing, understand that company names you query are sent to the skill's HTTPS service, and the skill keeps a product-specific credential/session file in your home directory for API key and purchase/session continuity. This is normal for this artifact, but avoid using it for confidential supplier research unless you trust the service operator.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

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 (1)

Intent-Code Divergence

Medium
Confidence
89% confidence
Finding
The comment at L015 says the skill does not read device information, specifically mentioning not reading username/hostname. However, the implementation immediately calls homedir() and stores data under a user-specific path in the home directory at L016-L019. While not reading hostname or username explicitly, it does use local system identity context despite the documentation's privacy-oriented claim.

Static analysis

No suspicious patterns detected.