ClawHub
Back to skill

Security audit

mobile-agent

Security checks across malware telemetry and agentic risk

Overview

This skill is transparent about controlling a real phone, but it grants broad device-control and screen-reading capabilities without enough safety, privacy, or consent guidance.

Install only if you are comfortable giving an agent broad control over a connected phone. Review the MCP server and website you are connecting to, avoid using it on phones with sensitive personal or work data, and require explicit approval before messages, purchases, settings changes, or account actions.

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 (2)

Missing User Warnings

Medium
Confidence
94% confidence
Finding
This skill controls a real phone and exposes tools that can capture screenshots, query device details, and return task results, but it does not warn users that these outputs may contain sensitive personal or enterprise data. In practice, screenshots, app contents, notifications, contacts, messages, and device metadata can be exfiltrated or mishandled if users are not explicitly informed before using the skill.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill can perform state-changing actions on a real device, including opening apps, tapping, inputting text, and running multi-step automation that may send messages or change settings, yet the documentation does not warn that these actions can modify device state or act on the user's behalf. Without a clear warning and consent model, users may authorize actions without understanding that the agent could send communications, trigger purchases, alter settings, or otherwise cause unintended real-world effects.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.