ClawHub
Back to skill

Security audit

AI检测专家 | 青虎AI

Security checks for vulnerabilities and agentic risk

Overview

This is a mostly coherent Qinghu AI-content detector, but it deserves review because it can send user text/images to a paid third-party API and may automatically write larger results to local files without asking.

Before installing, be comfortable with sending selected text or images to Qinghu and using a Qinghu API key that may consume credits. Do not submit confidential, legal, academic, or unpublished commercial material unless that third-party processing is acceptable. Ask the agent to confirm before exporting result files and to disclose where any local files or caches are stored.

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
90% confidence
Finding

The manifest description says this skill 'must trigger' for broad, common requests like checking authenticity or self-review before publishing. That can cause over-eager routing of ordinary content-review requests into a paid external detection workflow, increasing the chance of unnecessary data disclosure to a third-party service and unwanted charges without clear user intent. In this context, the risk is elevated because the skill sends user text/images to an external API and the trigger language is unusually broad and mandatory.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The skill instructs automatic export of datasets to a file and reuse of cached local file paths without first warning the user or obtaining explicit consent for file creation. This creates a privacy and data-handling risk: potentially sensitive text, images, or analysis results may be written to local storage unexpectedly, persisted longer than needed, and exposed via file links or later reuse. The skill context makes this more dangerous because it handles user-supplied content that may include unpublished commercial materials or authenticity-review artifacts.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

The entire skill is written to operate in Chinese and references fixed Chinese phrasing and presentation, but it does not state that users may choose another language for interaction or results. This can be a language/locale policy issue when a skill implicitly mandates one language without explicit user opt-in or documented justification.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.