ClawHub
Back to skill

Security audit

AI KEY·重讲一个概念

Security checks for vulnerabilities and agentic risk

Overview

This is a Chinese-language content workflow skill that reads and writes project notes for drafting concept videos, with no executable code, hidden persistence, credential use, or exfiltration behavior found.

Before installing, expect this skill to operate mainly in Chinese and to read or update your AI KEY/vault content files when invoked. Use explicit commands if you want to avoid accidental activation from broad phrases like asking to explain a concept.

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)

Natural-Language Policy Violations

Medium
Confidence
78% confidence
Finding
Natural-language policy violations apply to all file types, including manifest and markdown content. This skill prominently specifies behavior and triggers in Chinese while also embedding Chinese-only operational instructions, but it does not state that the user can choose language or that the locale restriction is intentional and justified.

Vague Triggers

Medium
Confidence
93% confidence
Finding
The trigger list includes very broad natural-language phrases such as asking to 'explain X properly' or 'what is X really,' which can overlap with ordinary conversation and unintentionally invoke the skill outside clear user intent. In an agent environment, this can cause prompt routing confusion and unauthorized execution of the skill's workflow, especially because the skill contains strong procedural instructions and file-reading behavior.

Static analysis

No suspicious patterns detected.