Agent Browser

Pass

Audited by Gen Agent Trust Hub on Mar 9, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs users to install the agent-browser package from the NPM registry and provides instructions for cloning the official source repository from Vercel's GitHub account.
  • [REMOTE_CODE_EXECUTION]: The agent-browser eval command allows the agent to execute arbitrary JavaScript code within the context of the active browser page. This is an intended feature for complex web interactions and data extraction.
  • [COMMAND_EXECUTION]: The skill operates by executing commands via the system shell using the agent-browser binary. This requires the agent to have Bash execution permissions.
  • [DATA_EXFILTRATION]: The skill includes commands to retrieve cookies (agent-browser cookies) and save/load session states (agent-browser state save/load auth.json). While these handle sensitive authentication data, the functionality is transparently documented for the purpose of maintaining persistent browser sessions.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 9, 2026, 08:18 PM