ClawHub
Back to skill

Security audit

Page Agent Claw Controller

Security checks across malware telemetry and agentic risk

Overview

This is a browser-control skill whose core purpose is legitimate, but it asks agents to control a user’s real browser session and install an external CLI without enough built-in safety boundaries.

Install only if you intentionally want an agent to control your browser. Prefer an isolated browser profile or test account, verify the GitHub Release before installing the CLI globally, and require explicit confirmation before logins, payments, crypto/account actions, uploads, downloads, form submissions, or reading authenticated pages.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
91% confidence
Finding
The skill’s activation criteria are broad enough that an agent may invoke browser automation whenever a prompt merely mentions page-agent or vaguely needs local browsing. That increases the chance of unnecessary browser control, navigation to sensitive pages, or execution of side-effecting actions without clear user confirmation or scope limits.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill instructs the agent to submit arbitrary browser-automation tasks but does not warn about privacy risks, authentication context, destructive actions, or data exfiltration from pages already open in the user’s browser environment. In this context, browser automation can click, search, navigate, and extract content, so missing safety guidance materially increases the risk of unintended side effects and exposure of sensitive information.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.