ClawHub
Back to skill

Security audit

Agent Touch Layer

Security checks across malware telemetry and agentic risk

Overview

This skill appears purpose-built for iOS Simulator automation, but it uses mutable external code and exposes sensitive browser/app automation powers that need careful review before installation.

Install only if you trust the JordanCoin/Atl repository and are comfortable letting the skill build and install an iOS Simulator app from a mutable GitHub branch. Use it with test accounts or isolated simulator profiles, avoid sensitive apps and logged-in sites unless necessary, and treat screenshots, accessibility snapshots, JavaScript evaluation, and cookie operations as sensitive session-level access.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
95% confidence
Finding
The skill metadata declares shell-based install steps (`git clone` and a setup script) but does not declare corresponding permissions/capabilities. This creates a trust gap where users or orchestration systems may not realize the skill executes host shell commands, including fetching and running remote code during installation.

Missing User Warnings

Medium
Confidence
90% confidence
Finding
The skill prominently supports native app automation and screenshot capture, including apps that may contain personal data, but it does not warn about collecting sensitive on-screen content. In practice, screenshots, accessibility snapshots, and text extraction from apps like Contacts, Photos, Notes, or Settings can expose PII, credentials, device details, or private communications.

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The documented cookie methods allow reading, setting, and deleting browser cookies without warning about session theft, impersonation, or destructive session loss. Because cookies often carry authentication state, these capabilities can expose account access or disrupt active sessions if misused.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.