ClawHub

LM Studio

Security checks across malware telemetry and agentic risk

Overview

This is a markdown-only LM Studio helper whose local server, model, API, MCP, and troubleshooting guidance is coherent with its stated purpose.

Install this if you want structured help operating LM Studio locally. Before installing, be aware it may keep local operational notes under ~/lm-studio/ and may guide commands that change which local models are loaded. Only enable MCP servers or remote endpoints from sources you trust, because those can expand file, tool, or network access beyond the local LM Studio server.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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)

Vague Triggers

Medium
Confidence
82% confidence
Finding
The setup text allows the skill to decide later whether to activate proactively for broad categories like local models, localhost APIs, embeddings, or MCP work, but it does not define firm boundaries or require clear user consent. In an agent setting, ambiguous activation rules can cause the skill to trigger in conversations where the user did not intend LM Studio assistance, increasing the risk of overreach, unintended guidance, or inappropriate handling of related local-service workflows.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill instructs the agent to persist user-specific operational details in ~/lm-studio/memory.md, including preferences, hardware-adjacent constraints, failure patterns, and trust boundaries, while also saying to never mention internal setup files or memory mechanics in user-facing replies. That combination creates a transparency and privacy problem: users may not realize their environment details and behavioral preferences are being retained, which can lead to unintended persistence of sensitive local configuration information.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal