ClawHub

MLX STT

Security checks across malware telemetry and agentic risk

Overview

This skill coherently performs local speech-to-text on Apple Silicon, with the main caveat that installation pulls external command-line dependencies.

Install only if you are comfortable letting Homebrew and uv add or update local tools, including mlx-audio. Prefer invoking it explicitly with `/mlx-stt <audio>` and remember that transcripts will be printed into the agent session.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • 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
Findings (2)

Vague Triggers

Medium
Confidence
92% confidence
Finding
The trigger phrases are broad natural-language patterns such as "STT ...", "ASR ...", and "Transcribe ...", which can match ordinary user requests rather than an explicit invocation. This increases the chance the skill runs unexpectedly on user-provided audio paths or transcription requests, creating unintended execution and expanding the attack surface for any installation or runtime behavior in the referenced scripts.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The install script automatically runs system-modifying package installation commands via Homebrew and uv without any prompt, warning, or prior disclosure to the user. This is dangerous because executing the script changes the local environment, pulls remote packages, and expands the supply-chain trust boundary in a way users may not expect from simply installing or invoking a skill.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal