MCP Browser

No separate API keys or environment variables are required, but “using the user's Chromium session” and “auth” indicate it can leverage existing logged-in browser state to access protected sites. The main concern is indirect use of session cookies/auth state rather than explicit credential collection.

No fixed remote endpoint is declared, and there is no evidence of exfiltration to unrelated services. However, browser automation inherently makes network requests to user-targeted websites, so page content, form data, or session-backed requests may be sent to the sites being automated.

The system flags executes-code, and the description explicitly includes “JS execution.” This means the tool can drive a local browser, execute scripts in the web page context, and manipulate pages. That is standard capability for browser automation, and the material does not show system-level overreach beyond its stated function.

Because it relies on the user's Chromium session, the tool can typically access page content, DOM state, and data associated with the current browsing session, and may interact with authenticated pages. The material does not state arbitrary local file read/write or excessive system permissions, but access to browser-session data is inherently sensitive.

Positive indicators include an official registry source, an open-source repository available for audit, and updates within the last year, all of which reduce supply-chain risk. Points to watch are the missing README, undeclared license, and very low community adoption (0 stars), so while it is auditable, maturity and external validation are limited.