agent-browser

Pass

Audited by Gen Agent Trust Hub on Mar 24, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [Browser Automation CLI]: The skill uses a tool called agent-browser to perform web automation tasks. This involves executing shell commands to interact with a browser instance, which is a standard pattern for end-to-end testing and visual verification of web applications.
  • [Local Development Interaction]: The skill is designed to automatically trigger or provide guidance when local development servers (like Next.js or Vite) are detected. This is a common developer workflow meant to facilitate immediate testing and feedback during the development process.
  • [JavaScript Execution]: Includes functionality to evaluate JavaScript within the browser context (agent-browser eval). While this allows for arbitrary code execution inside the browser, it is a standard feature for advanced web scraping and testing where simple selectors are insufficient. It does not appear to execute code on the host system beyond the browser's sandbox.
  • [State Management]: Supports saving and loading browser state (e.g., auth.json). This is a legitimate feature for maintaining authentication sessions during testing, though users should ensure that sensitive state files are managed according to their own security policies.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 24, 2026, 03:25 PM