ClawHub

mongolian-llm

Security checks across malware telemetry and agentic risk

Overview

This is a documented Mongolian-language API connector that uses the user's own key and sends requested content to one disclosed remote service.

Install only if you are comfortable sending requested Mongolian-language text, images, audio, and documents to mongol.open-idea.net under your own paid API key. Keep the key in OpenClaw configuration, not chat, and require confirmation before large, sensitive, batch, OCR, ASR, or document-translation jobs.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The trigger list contains many broad, everyday phrases such as general language names and common actions, which can cause the skill to activate in situations where the user did not specifically intend to send content to this external service. Because this skill can process chats, files, OCR, ASR, and translation through a remote API, accidental invocation increases the risk of unintended data transmission and user confusion.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill description states that it sends content to https://mongol.open-idea.net/api/v1 for translation, OCR, ASR, TTS, and document processing, but it does not prominently warn that user text, audio, images, and files may be transmitted to a third-party service. In a skill that handles potentially sensitive communications and documents, the absence of an upfront privacy/data-handling disclosure materially increases the chance of users sharing confidential data without informed consent.

Natural-Language Policy Violations

Medium
Confidence
87% confidence
Finding
The routing rule hard-codes Chinese output for requests involving Mongolian text plus explanation, without stating that this should follow the user's language preference or explicit consent. This can override user intent, cause unintended disclosure in a different language context, and create prompt-level behavior that is inconsistent with least-surprise and user-controlled output policies.

VirusTotal

VirusTotal engine telemetry is currently stale for this artifact.

View on VirusTotal