ClawHub
Back to skill

Security audit

Offload Tasks to LM Studio Models

Security checks across malware telemetry and agentic risk

Overview

This skill does what it claims by routing selected work to a local LM Studio server, with manageable privacy considerations around broad triggers, local storage, and optional logging.

Install this if you run and trust LM Studio. For sensitive documents or code, confirm the API URL is localhost or another trusted endpoint, and avoid --log or --stateful unless you are comfortable retaining prompt, response, or conversation identifiers locally.

SkillSpector

By NVIDIA
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)

Vague Triggers

Medium
Confidence
84% confidence
Finding
The README states the skill is automatically triggered for a wide set of common tasks such as summarization, extraction, rewriting, brainstorming, and first-pass code review. Broad trigger conditions can cause unintended invocation, which may route sensitive prompts or data to the local LM Studio service without sufficiently explicit user intent, increasing the chance of unexpected data handling or tool misuse.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
When `--log` is used, the script appends the full request payload and full API response to disk, which can include sensitive prompts, model outputs, identifiers, and potentially private user data. In a skill explicitly positioned for privacy-sensitive/local use, silent persistence of this data materially increases exposure through local file access, backups, or later accidental disclosure.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

Detected: suspicious.env_credential_access

Environment variable access combined with network send.

Critical
Code
suspicious.env_credential_access
Location
scripts/lmstudio-api.mjs:12

Environment variable access combined with network send.

Critical
Code
suspicious.env_credential_access
Location
scripts/load.mjs:8

Environment variable access combined with network send.

Critical
Code
suspicious.env_credential_access
Location
scripts/test.mjs:9

Environment variable access combined with network send.

Critical
Code
suspicious.env_credential_access
Location
scripts/unload.mjs:9