ClawHub
Back to skill

Security audit

mimo-omni

Security checks across malware telemetry and agentic risk

Overview

This skill appears purpose-built for MiMo media analysis, but users should review it because it uploads local media and prompts to a remote API and can automatically use stored API credentials.

Install only if you trust the publisher and Xiaomi MiMo with the images, videos, audio, URLs, and prompts you submit. Avoid sensitive or regulated media unless you have approval. Prefer setting MIMO_API_KEY explicitly for this tool, and do not set MIMO_API_ENDPOINT unless you intentionally trust that destination to receive both media and the API key.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (4)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explicitly instructs sending local image, video, and audio files to the MiMo API, but does not clearly warn that user media is uploaded to an external third-party service for processing. This is a privacy and data-handling issue because users may provide sensitive screenshots, recordings, or documents without informed consent about external transmission.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The instruction to return stdout directly to the user normalizes blind pass-through of a third-party model's output while omitting disclosure that prompts and media were processed externally. In context, this makes the privacy issue more dangerous because it encourages seamless proxying of sensitive user content to an outside provider without an explicit consent checkpoint or handling notice.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
The tool converts local image, video, and audio files into data URIs and sends them, along with the user's question, to a remote API. Because there is no explicit consent prompt, disclosure notice, or guardrail around sensitive local content, users may unintentionally exfiltrate private media and embedded text/audio to a third party.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The script automatically loads credentials and sends user-supplied media and prompts to a third-party API, but there is no explicit consent prompt or disclosure at the point of transmission. In a multimodal skill, users may provide sensitive images, videos, or audio without realizing they are being uploaded off-device, creating a real privacy and data-handling risk.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.