opencode browser Markdown PT

Browser tools

7. What the agent can do in the browser

The agent gets 18 tools. Some are read-only and never ask for permission; the others change something and go through your approval (depending on the mode).

ToolWhat it doesPermission
browser_tabs_contextLists the tabs in the conversation's groupRead
browser_tabs_createOpens a URL in the group (in the background)Read
browser_tabs_attachBrings an already open tab into the groupAsks
browser_tabs_close_tabCloses a tab in the groupAsks
browser_navigateNavigates, goes back or forward in the historyAsks
browser_read_pageReads the page and returns the elements with referencesRead
browser_findFinds elements by text, role or nameRead
browser_get_page_textExtracts the main text, without HTMLRead
browser_form_inputFills in a fieldAsks
browser_computerClicks, types, presses keys, scrolls, moves the mouse, takes a screenshot, zooms, waitsAsks
browser_javascript_toolRuns JavaScript on the pageAsks
browser_read_console_messagesReads the tab's console and errorsRead
browser_read_network_requestsReads the tab's network requestsRead
browser_resize_windowResizes the window to test layoutsNo request
browser_update_planShows domains and steps for you to approve beforehandPlan card
browser_upload_imageSends files to a file field or drop areaAsks
browser_gif_creatorRecords action frames and exports a GIFExport asks
browser_batchRuns several tools in sequence (up to 20)Each action asks for its own

Tool cards

Each call appears in the chat with a label — “Navigate”, “Read page”, “Click on the page”, “Fill field”, “Console”, “Network” — and a status: ok, error or running.

You choose how much detail you want in Settings → Tools in chat:

ModeHow it looks
Grouped (open in a modal) (default)One summary row per run (“4 tools · view”); the details open in the “Tools of this run” modal
Compact rows, one per callOne row per tool, with the error summarized on the row itself
Full cards, with parameters and resultEverything open in the chat, including the input JSON and the output
Hide tools and reasoning (only “working”)The chat shows only the activity (“working: reading the page”)

In grouped mode, the modal has an “Errors only” filter and closes with Esc or by clicking outside.

Screenshots appear as an image inside the card (up to 3 at a time) and are also saved to disk: the file path comes in the agent's reply.