Get browser automation engine and extension connection status

List Chrome windows

List open Chrome tabs (existing session)

Get the active Chrome tab

Focus a tab by id

Read page title, text, forms (untrusted content)

Read accessibility tree for the page

Find elements by text, selector, or role

Get details for an element selector

Capture a screenshot of a tab

Extract hyperlinks from the page

Extract a table from the page

Read recent console messages

Read recent network errors

List recent downloads

Navigate to a URL in a tab

Create a new tab

Close a tab

Reload the current or given tab

History back

History forward

Click an element (DOM-first). Requires confirmation in ask mode.

Type text into a field (does not submit)

Clear an input field

Select a dropdown option

Check a checkbox

Uncheck a checkbox

Scroll the page or an element

Hover an element

Drag from one selector to another

Upload a file from an approved path (requires confirmation)

Download a file to an approved folder (requires confirmation)

Wait for a selector or timeout

Accept or dismiss a browser dialog

Execute a saved workflow by id (requires confirmation)

List OS windows (computer-use)

Focus a window by title substring

Capture full screen (prefer browser screenshot)

OCR locate text on screen (fallback)

Click screen coordinates (high caution)

Type via OS keyboard (high caution)

Press a key via OS

Scroll via OS

Wait milliseconds (max 30000)

Pause and request the user to complete a step

Run system health check

Check whether Emergency Stop is active

Detect forms on the page (no secrets returned)

Preview autofill mapping; passwords/payment never returned to LLM

Fill form fields after approval (does NOT submit)