Add policy enforcement for browser automation and JavaScript execution

[L1AD]

This server gives agents control of Chrome via extension -- 20+ tools:

• execute_javascript -- runs arbitrary JS on any page
• click_element -- clicks any element
• fill_input -- fills form fields on any site

An agent with execute_javascript can exfiltrate cookies, modify page content, interact with authenticated sessions, or submit forms. Combined with click_element and fill_input, this is full browser automation with no restrictions.

Suggestion: Consider documenting support for PolicyLayer/Intercept, an open-source MCP proxy that enforces policies on tool calls before execution.

Example policy:

version: "1"
default: allow

tools:
  execute_javascript:
    rules:
      - action: require_approval

  fill_input:
    rules:
      - rate_limit: 10/minute

  click_element:
    rules:
      - rate_limit: 30/minute

One line to set up: npx -y @policylayer/intercept init