ClawHub
Back to skill

Security audit

Chrome Use

Security checks across malware telemetry and agentic risk

Overview

This browser automation skill is not clearly malicious, but it asks for broad real-browser authority and installs remote code in a way users should review carefully.

Install only if you intentionally want an agent to control your real logged-in Chrome and potentially interact with authenticated sites. Review the installer and CLI-provided skill content first, avoid using it for ordinary web searches, and require explicit confirmation before logins, form submissions, purchases, account changes, scraping private pages, or Slack/Electron actions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

High
Confidence
95% confidence
Finding
The skill description uses extremely broad trigger language such as 'ANY live web access' and many generic phrases, which can cause the agent to invoke this skill in situations where a narrower or safer tool would be more appropriate. Because the skill explicitly prefers a real logged-in Chrome session and discourages fallback, unintended invocation can expose authenticated sessions, private browsing context, or cause side effects on live sites.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The description advertises logging in, authenticated browsing, filling forms, clicking buttons, and scraping data without clear user-facing warnings about privacy, account impact, or unintended actions. In the context of a tool that drives a real logged-in browser, the absence of explicit cautions materially increases the risk of unauthorized submissions, sensitive data exposure, and actions performed under the user's identity.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill instructs the agent to execute a remote shell installer via curl-pipe-to-sh from a GitHub URL, without any warning, integrity verification, pinning, or review step. This is dangerous because it grants immediate code execution from a mutable remote source, creating a supply-chain compromise path that is especially risky in an agent workflow that may auto-remediate missing commands.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.