ClawHub
Back to skill

Security audit

Agent Browser

Security checks across malware telemetry and agentic risk

Overview

This is a coherent browser automation skill, but it gives an agent broad control over authenticated browser sessions without enough safety boundaries.

Install only if you need agent-driven browser automation. Use a disposable browser profile or test account, avoid saving auth.json in shared or committed locations, delete saved state, screenshots, PDFs, and recordings when done, and require explicit confirmation before the agent submits forms, uploads files, changes account data, posts content, or reads cookies/localStorage.

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)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill explicitly documents capabilities to submit forms, upload files, capture screenshots/PDFs, and save/load authenticated browser state, but it does not warn about handling sensitive data, privacy exposure, or credential/session-token risks. In an agent context, these features can easily lead to unintended exfiltration or reuse of sensitive information if operators are not clearly instructed about consent, storage hygiene, and data minimization.

Vague Triggers

Medium
Confidence
92% confidence
Finding
The trigger phrases are broad enough to match many generic browsing or scraping requests, which can cause this skill to activate in contexts the user did not specifically intend. Because this skill enables browser automation, accidental invocation could lead to unintended navigation, data entry, scraping, or interaction with sensitive web content.

VirusTotal

63/63 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.