ClawHub

Browser Harness

Security checks across malware telemetry and agentic risk

Overview

This skill is not clearly malicious, but it gives an agent broad control over a real browser session and includes an under-explained remote browser option.

Install only if you intentionally want the agent to control a browser. Use a dedicated temporary browser profile, avoid sensitive or logged-in sites unless you explicitly approve that access, close the debug-enabled browser when finished, review the external browser-harness project before installing, and do not enable remote/cloud mode without understanding where page content and credentials may be sent.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (2)

Context-Inappropriate Capability

Medium
Confidence
92% confidence
Finding
The skill is described as controlling a user's local browser via CDP, but the documentation also introduces a remote/cloud browser mode through `BROWSER_USE_API_KEY` and `start_remote_daemon()`. That materially expands the trust boundary and data exposure surface, because browsing actions, page contents, session context, and possibly sensitive user data may be sent to a remote service not disclosed in the core skill purpose.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The trigger description is broad enough to activate for many generic web-related tasks such as screenshots, content extraction, or browser automation, without clear limits around sensitive sites, authenticated sessions, or destructive actions. In this context, the skill can drive the user's real local browser, making overbroad activation more dangerous because it may access live cookies, account data, and in-session content unexpectedly.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal