ClawHub
Back to skill

Security audit

亚马逊-细分市场评估师 | 青虎AI

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Amazon market-analysis workflow that uses a Qinghu API token and may export result tables, with no evidence of hidden execution, exfiltration, or persistence beyond purpose-aligned output files.

Before installing, users should be comfortable providing a Qinghu API token, authorizing potentially paid Qinghu API calls, and receiving exported table files or links for larger datasets. Review token handling and generated files if the market research is commercially sensitive.

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 (3)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill declares broad mandatory trigger conditions for common market-evaluation questions, which can cause it to activate in situations where the user did not explicitly choose this skill or expect external-data-driven behavior. In context, that increases the chance of unintended tool use, unnecessary data access, and accidental paid API calls, especially because the skill later instructs the agent to seek tool authorization and use multiple external endpoints.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The instruction says that when the user provides a category name or keyword, it should first be translated into the target site's language before use. This imposes a language/locale behavior automatically and does not offer the user a choice or explain exceptions, which conflicts with the policy against forcing a specific language without opt-in.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The skill instructs the agent to automatically export large datasets to local files or links without first asking the user or clearly warning in the user-facing description. That can lead to unintended persistence of potentially sensitive business data, surprise file creation, or unanticipated sharing/link generation beyond what the user requested.

Content

No source excerpt is available for this finding.