browser-automation

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS] (MEDIUM): The skill directs users to install a plugin from an unverified GitHub repository (sawyerhood/dev-browser). This source is not included in the list of trusted organizations or repositories, posing a risk of installing malicious or unvetted code.
  • [COMMAND_EXECUTION] (MEDIUM): The skill provides explicit instructions to execute plugin installation commands (/plugin marketplace add...) for an unverified external source.
  • [PROMPT_INJECTION] (LOW): The skill is inherently vulnerable to Indirect Prompt Injection (Category 8) due to its core functionality.
  • Ingestion points: Untrusted data enters the agent context via browser_snapshot and browser_take_screenshot (SKILL.md).
  • Boundary markers: Absent. There are no instructions to the agent to ignore or delimit embedded instructions within the page content.
  • Capability inventory: The agent has a wide range of interactive capabilities including browser_click, browser_type, browser_fill_form, and browser_press_key across all sections of the skill.
  • Sanitization: Absent. The skill provides no guidance on escaping or validating external web content before the agent interacts with it.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 06:40 PM