ClawHub
Back to skill

Security audit

ABTI: AI-Based Type Indicator

Security checks across malware telemetry and agentic risk

Overview

This is a coherent entertainment skill, but it asks agents to analyze sensitive chat history while relying on missing or changeable instructions and allowing derived results to be sent to YouMind.

Install only if you are comfortable with the agent reviewing the chat context available to it. Use a fresh or minimized conversation, inspect any remote prompt before sending it to an AI, and require confirmation before any result is posted to YouMind for a shareable card.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The README encourages users to let an AI analyze their chat history, which can contain sensitive personal, financial, health, or credential-like information. Even if the project claims the developer never sees the chats, the data may still be processed by the user's AI platform, so the missing up-front warning can mislead users about real privacy exposure.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The README tells users to fetch a remote URL, copy the entire contents, and paste it into a chatbot without reviewing it first. That creates a prompt-injection and supply-chain style risk: the remote document can change at any time and may contain adversarial instructions that cause unsafe actions, data disclosure, or policy bypass attempts inside the target AI environment.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to POST personality results to an external service, but it does not require explicit, informed user consent immediately before network transmission. Because the result is derived from the user's chat history and may encode sensitive behavioral or psychological inferences, sending it off-platform creates a real privacy and data disclosure risk even if raw chat text is not included.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The package description explicitly says the skill analyzes how a user talks to AI and that their chat history is the quiz, but it does not define scope, consent, or trigger boundaries. In a skill ecosystem, overly broad language around chat-history analysis can lead to unnecessary access to sensitive prior conversations and make privacy-invasive behavior appear intended by default.

VirusTotal

62/62 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.