browser-manager

Use this agent when the user wants to interact with, inspect, or test something in a live Chrome browser. Handles navigation, form filling, clicking buttons, scrolling, reading console logs, inspecting network requests, and reporting what's visible on screen. <example> Context: User is developing a web app and wants to verify it works correctly. user: "Check in browser if the login form submits correctly" assistant: "I'll use the browser-agent to navigate to the login form, fill it out, and verify the submission behavior." <commentary> The phrase "check in browser" is a direct trigger for this agent. The agent will use chrome-devtools MCP to interact with the live browser session. </commentary> </example> <example> Context: User wants to debug a network issue. user: "Check the network requests when I click submit on this page" assistant: "I'll use the browser-agent to monitor network traffic while triggering the submit action." <commentary> Network inspection is a core capability of this agent — it should monitor DevTools network panel data. </commentary> </example> <example> Context: User wants to check for JavaScript errors. user: "Are there any console errors on the dashboard page?" assistant: "I'll use the browser-agent to navigate to the dashboard and read the console output." <commentary> Console log inspection is a primary use case. The agent reads errors, warnings, and info messages from DevTools. </commentary> </example> <example> Context: User wants to verify UI behavior. user: "Open https://example.com in browser and tell me what's on the page" assistant: "I'll use the browser-agent to navigate to that URL and describe the page content." <commentary> Opening a URL and reporting page content is a basic navigation task this agent handles. </commentary> </example> <example> Context: User wants to test a form. user: "Fill in the registration form with test data and submit it" assistant: "I'll use the browser-agent to locate the registration form fields, fill them with test data, and submit." <commentary> Form filling and submission is a direct automation task for this agent. </commentary> </example> <example> Context: User wants to test in browser after making code changes. user: "Test this in the browser and see if it works" assistant: "I'll use the browser-agent to test the current state of the application in the browser." <commentary> Generic "test in browser" requests should trigger this agent to perform interactive verification. </commentary> </example>