Explainer: three ways AI agents already control employee browsers
Original title: 3 Ways AI Agents Already Operate in Your Browser
Debugging protocols, in-browser agents and WebMCP can act inside signed-in sessions that logs may attribute to the user.
BleepingComputer describes three paths for AI agents in employee browsers: external control over Chrome DevTools Protocol, in-browser extensions and agentic browsers, and WebMCP tools exposed by sites. CDP automation and extensions such as Claude in Chrome can use existing authenticated sessions, while Comet and ChatGPT Desktop may sit outside extension controls. Agents act as the user and can follow instructions in page content. WebMCP, previewed in Chrome in 2026 with Google and Microsoft input via a W3C group, lets sites define callable tools. The piece ends as a pitch for Keep Aware, a Jamf company.
- Playwright and Puppeteer have long used CDP for browser automation.
- Claude in Chrome needs scripting, tab and debugger permissions.
- WebMCP tool names and schemas come from the website, not a trusted party.
- Cross-site agents can pass data from one site’s tools into another.
Impact: Agents can act inside authenticated browser sessions as the user
- 1.Inventory debugging-enabled browsers, AI extensions and agentic browsers on employee devices.
- 2.Check whether logs can separate agent actions from the signed-in employee.
- Organizations
- W3CJamfKeep Aware
From the original article
4 min read at bleepingcomputer.com
AI agents can control authenticated browsers, interact with corporate applications, and access sensitive data while appearing to act as legitimate employees. Keep Aware examines three ways agents operate inside browsers and why traditional security tools may struggle to distinguish their activity from human users. [...]
By Sponsored by Keep Aware · bleepingcomputer.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.