ClawHub
Back to skill

Security audit

Ozon-蓝海赛道挖掘专家 | 青虎AI

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Ozon market-research workflow that uses a Qinghu API token and may export result tables, but its behavior fits its stated purpose.

Before installing, be prepared to provide or configure a Qinghu API token and approve paid API calls. Large result sets may be exported to local table files, so use it in a workspace where storing Ozon market-research data locally is acceptable.

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
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

Medium
Confidence
93% confidence
Finding
The skill explicitly says it 'must trigger' for broad Ozon category-selection intents, which can cause the agent to invoke it in many ordinary commerce conversations without clear user intent. In this skill’s context, that broad auto-activation is more dangerous because the skill also pushes external API use and paid/tool-driven workflows, increasing the chance of unnecessary data access, charges, and workflow hijacking.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill instructs the agent to export table files by default whenever records are >=10 and explicitly says not to ask the user first. That creates an unsafe write behavior: the agent may create local/output files and persist potentially sensitive business data without explicit consent, which is especially risky because the skill also mentions cached local file paths and downstream reuse of those files.

Static analysis

No suspicious patterns detected.