Agent Browser

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: Downloads the agent-browser utility from Vercel Labs' official GitHub repository and the NPM registry.
  • [COMMAND_EXECUTION]: Provides an interface for executing complex browser automation tasks through the Bash tool, including element interaction, session management, and state persistence.
  • [INDIRECT_PROMPT_INJECTION]: As the skill is designed to navigate and extract data from the open web, it creates an attack surface where malicious instructions embedded in a website could influence the agent's behavior.
  • Ingestion points: Web content retrieved via snapshot, get text, get html, and browser console logs.
  • Boundary markers: No explicit delimiters or warnings are present in the instructions to prevent the agent from obeying instructions found within web content.
  • Capability inventory: The skill has access to the Bash tool, can perform network navigation, write local files (screenshot, pdf, state save), and execute browser-side JavaScript (eval).
  • Sanitization: No sanitization or filtering of the extracted web content is implemented within the skill definition.
  • [DYNAMIC_EXECUTION]: Supports the execution of arbitrary JavaScript within the browser context via the eval command and dynamic wait functions (wait --fn).
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 04:23 PM